�ݺ�ߣshows by User: albahnsen

�ݺ�ߣshows by User: albahnsen / http://www.slideshare.net/images/logo.gif �ݺ�ߣshows by User: albahnsen / Tue, 04 Dec 2018 22:26:05 GMT �ݺ�ߣShare feed for �ݺ�ߣshows by User: albahnsen black hat deephish /slideshow/black-hat-deephish/124982274 201812blackhatdeepphish-181204222605
DeepPhish - Simulating malicious AI. ]]>
DeepPhish - Simulating malicious AI. ]]> Tue, 04 Dec 2018 22:26:05 GMT /slideshow/black-hat-deephish/124982274 albahnsen@slideshare.net(albahnsen) black hat deephish albahnsen DeepPhish - Simulating malicious AI. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/201812blackhatdeepphish-181204222605-thumbnail.jpg?width=120&height=120&fit=bounds" /> DeepPhish - Simulating malicious AI.

black hat deephish from Alejandro Correa Bahnsen, PhD

]]> 837 4 https://cdn.slidesharecdn.com/ss_thumbnails/201812blackhatdeepphish-181204222605-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 DeepPhish: Simulating malicious AI /slideshow/deepphish-simulating-malicious-ai/98311818 201805deepphishapwgecrime-180523182835
In this work we describe how threat actors may use AI algorithms to bypass AI phishing detection systems. We analyzed more than a million phishing URLs to understand the different strategies that threat actors use to create phishing URLs. Assuming the role of an attacker, we simulate how different threat actors may leverage Deep Neural Networks to enhance their effectiveness rate. Using Long Short-Term Memory Networks, we created DeepPhish, an algorithm that learns to create better phishing attacks. By training the DeepPhish algorithm for two different threat actors, they were able to increase their effectiveness from 0.69% to 20.9%, and 4.91% to 36.28%, respectively.]]>
In this work we describe how threat actors may use AI algorithms to bypass AI phishing detection systems. We analyzed more than a million phishing URLs to understand the different strategies that threat actors use to create phishing URLs. Assuming the role of an attacker, we simulate how different threat actors may leverage Deep Neural Networks to enhance their effectiveness rate. Using Long Short-Term Memory Networks, we created DeepPhish, an algorithm that learns to create better phishing attacks. By training the DeepPhish algorithm for two different threat actors, they were able to increase their effectiveness from 0.69% to 20.9%, and 4.91% to 36.28%, respectively.]]> Wed, 23 May 2018 18:28:35 GMT /slideshow/deepphish-simulating-malicious-ai/98311818 albahnsen@slideshare.net(albahnsen) DeepPhish: Simulating malicious AI albahnsen In this work we describe how threat actors may use AI algorithms to bypass AI phishing detection systems. We analyzed more than a million phishing URLs to understand the different strategies that threat actors use to create phishing URLs. Assuming the role of an attacker, we simulate how different threat actors may leverage Deep Neural Networks to enhance their effectiveness rate. Using Long Short-Term Memory Networks, we created DeepPhish, an algorithm that learns to create better phishing attacks. By training the DeepPhish algorithm for two different threat actors, they were able to increase their effectiveness from 0.69% to 20.9%, and 4.91% to 36.28%, respectively. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/201805deepphishapwgecrime-180523182835-thumbnail.jpg?width=120&height=120&fit=bounds" /> In this work we describe how threat actors may use AI algorithms to bypass AI phishing detection systems. We analyzed more than a million phishing URLs to understand the different strategies that threat actors use to create phishing URLs. Assuming the role of an attacker, we simulate how different threat actors may leverage Deep Neural Networks to enhance their effectiveness rate. Using Long Short-Term Memory Networks, we created DeepPhish, an algorithm that learns to create better phishing attacks. By training the DeepPhish algorithm for two different threat actors, they were able to increase their effectiveness from 0.69% to 20.9%, and 4.91% to 36.28%, respectively.

DeepPhish: Simulating malicious AI from Alejandro Correa Bahnsen, PhD

]]> 2821 4 https://cdn.slidesharecdn.com/ss_thumbnails/201805deepphishapwgecrime-180523182835-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 AI vs. AI: Can Predictive Models Stop the Tide of Hacker AI? /slideshow/ai-vs-ai-can-predictive-models-stop-the-tide-of-hacker-ai/95771190 enpresentationcds-180503020944
Machine learning and artificial intelligence (AI) have become essential components of any effective cyber security plan. They allow us to spot and take down fraud more quickly and accurately than any traditional method, and provide a much higher level of efficiency. When we are fighting against humans with AI on our side, we know that we will always have the advantage. We also know that every defensive action we take will undoubtedly create a reaction from fraudsters. So, what would happen if fraudsters were to begin using AI technology? Would they be able to defeat existing defenses? Our dedicated research team of data scientists decided to find out.]]>
Machine learning and artificial intelligence (AI) have become essential components of any effective cyber security plan. They allow us to spot and take down fraud more quickly and accurately than any traditional method, and provide a much higher level of efficiency. When we are fighting against humans with AI on our side, we know that we will always have the advantage. We also know that every defensive action we take will undoubtedly create a reaction from fraudsters. So, what would happen if fraudsters were to begin using AI technology? Would they be able to defeat existing defenses? Our dedicated research team of data scientists decided to find out.]]> Thu, 03 May 2018 02:09:44 GMT /slideshow/ai-vs-ai-can-predictive-models-stop-the-tide-of-hacker-ai/95771190 albahnsen@slideshare.net(albahnsen) AI vs. AI: Can Predictive Models Stop the Tide of Hacker AI? albahnsen Machine learning and artificial intelligence (AI) have become essential components of any effective cyber security plan. They allow us to spot and take down fraud more quickly and accurately than any traditional method, and provide a much higher level of efficiency. When we are fighting against humans with AI on our side, we know that we will always have the advantage. We also know that every defensive action we take will undoubtedly create a reaction from fraudsters. So, what would happen if fraudsters were to begin using AI technology? Would they be able to defeat existing defenses? Our dedicated research team of data scientists decided to find out. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/enpresentationcds-180503020944-thumbnail.jpg?width=120&height=120&fit=bounds" /> Machine learning and artificial intelligence (AI) have become essential components of any effective cyber security plan. They allow us to spot and take down fraud more quickly and accurately than any traditional method, and provide a much higher level of efficiency. When we are fighting against humans with AI on our side, we know that we will always have the advantage. We also know that every defensive action we take will undoubtedly create a reaction from fraudsters. So, what would happen if fraudsters were to begin using AI technology? Would they be able to defeat existing defenses? Our dedicated research team of data scientists decided to find out.

AI vs. AI: Can Predictive Models Stop the Tide of Hacker AI? from Alejandro Correa Bahnsen, PhD

]]> 316 3 https://cdn.slidesharecdn.com/ss_thumbnails/enpresentationcds-180503020944-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 How I Learned to Stop Worrying and Love Building Data Products /slideshow/how-i-learned-to-stop-worrying-and-love-building-data-products/81247510 behindthescenesinbuildingdataproducts-171026152811
Most people think a successful data product requires just three things: data, the right algorithm, and good execution. But as anyone who’s tried to create one knows, an effective product requires much more. In this talk, Dr. Correa Bahnsen will share his successes—and failures—in building data products for information security, and why an isolated data science team is a recipe for failure.]]>
Most people think a successful data product requires just three things: data, the right algorithm, and good execution. But as anyone who’s tried to create one knows, an effective product requires much more. In this talk, Dr. Correa Bahnsen will share his successes—and failures—in building data products for information security, and why an isolated data science team is a recipe for failure.]]> Thu, 26 Oct 2017 15:28:11 GMT /slideshow/how-i-learned-to-stop-worrying-and-love-building-data-products/81247510 albahnsen@slideshare.net(albahnsen) How I Learned to Stop Worrying and Love Building Data Products albahnsen Most people think a successful data product requires just three things: data, the right algorithm, and good execution. But as anyone who’s tried to create one knows, an effective product requires much more. In this talk, Dr. Correa Bahnsen will share his successes—and failures—in building data products for information security, and why an isolated data science team is a recipe for failure. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/behindthescenesinbuildingdataproducts-171026152811-thumbnail.jpg?width=120&height=120&fit=bounds" /> Most people think a successful data product requires just three things: data, the right algorithm, and good execution. But as anyone who’s tried to create one knows, an effective product requires much more. In this talk, Dr. Correa Bahnsen will share his successes—and failures—in building data products for information security, and why an isolated data science team is a recipe for failure.

How I Learned to Stop Worrying and Love Building Data Products from Alejandro Correa Bahnsen, PhD

]]> 651 1 https://cdn.slidesharecdn.com/ss_thumbnails/behindthescenesinbuildingdataproducts-171026152811-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Fraud Detection by Stacking Cost-Sensitive Decision Trees /albahnsen/2017-stacking-csdtacorrea 2017stackingcsdtacorrea-170926224621
Worldwide, billions of euros are lost every year due to credit card fraud. Increasingly, fraud has diversified to different digital channels, including mobile and online payments, creating new challenges as innovative new fraud patterns emerge. Hence, it remains challenging to find effective methods of mitigating fraud. Existing solutions include simple if-then rules and classical machine learning algorithms. Credit card fraud is by definition an example-dependent and cost-sensitive classification problem, in which the costs due to is classification vary between examples and not only within classes, i.e., misclassifying a fraudulent transaction may have a financial impact ranging from a few to thousands of euros. In this paper, we propose an extension to the cost-sensitive decision trees algorithm, by creating an ensemble of such trees, and combining them using a stacking approach with a cost-sensitive logistic regression. We compare our method with standard machine learning algorithms and state-of-the-art cost-sensitive classification methods using a real credit card fraud dataset provided by a large European card processing company. The results show that our method achieves savings of up to 73.3%, more than 2 percentage points more than a single cost-sensitive decision tree.]]>
Worldwide, billions of euros are lost every year due to credit card fraud. Increasingly, fraud has diversified to different digital channels, including mobile and online payments, creating new challenges as innovative new fraud patterns emerge. Hence, it remains challenging to find effective methods of mitigating fraud. Existing solutions include simple if-then rules and classical machine learning algorithms. Credit card fraud is by definition an example-dependent and cost-sensitive classification problem, in which the costs due to is classification vary between examples and not only within classes, i.e., misclassifying a fraudulent transaction may have a financial impact ranging from a few to thousands of euros. In this paper, we propose an extension to the cost-sensitive decision trees algorithm, by creating an ensemble of such trees, and combining them using a stacking approach with a cost-sensitive logistic regression. We compare our method with standard machine learning algorithms and state-of-the-art cost-sensitive classification methods using a real credit card fraud dataset provided by a large European card processing company. The results show that our method achieves savings of up to 73.3%, more than 2 percentage points more than a single cost-sensitive decision tree.]]> Tue, 26 Sep 2017 22:46:21 GMT /albahnsen/2017-stacking-csdtacorrea albahnsen@slideshare.net(albahnsen) Fraud Detection by Stacking Cost-Sensitive Decision Trees albahnsen Worldwide, billions of euros are lost every year due to credit card fraud. Increasingly, fraud has diversified to different digital channels, including mobile and online payments, creating new challenges as innovative new fraud patterns emerge. Hence, it remains challenging to find effective methods of mitigating fraud. Existing solutions include simple if-then rules and classical machine learning algorithms. Credit card fraud is by definition an example-dependent and cost-sensitive classification problem, in which the costs due to is classification vary between examples and not only within classes, i.e., misclassifying a fraudulent transaction may have a financial impact ranging from a few to thousands of euros. In this paper, we propose an extension to the cost-sensitive decision trees algorithm, by creating an ensemble of such trees, and combining them using a stacking approach with a cost-sensitive logistic regression. We compare our method with standard machine learning algorithms and state-of-the-art cost-sensitive classification methods using a real credit card fraud dataset provided by a large European card processing company. The results show that our method achieves savings of up to 73.3%, more than 2 percentage points more than a single cost-sensitive decision tree. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/2017stackingcsdtacorrea-170926224621-thumbnail.jpg?width=120&height=120&fit=bounds" /> Worldwide, billions of euros are lost every year due to credit card fraud. Increasingly, fraud has diversified to different digital channels, including mobile and online payments, creating new challenges as innovative new fraud patterns emerge. Hence, it remains challenging to find effective methods of mitigating fraud. Existing solutions include simple if-then rules and classical machine learning algorithms. Credit card fraud is by definition an example-dependent and cost-sensitive classification problem, in which the costs due to is classification vary between examples and not only within classes, i.e., misclassifying a fraudulent transaction may have a financial impact ranging from a few to thousands of euros. In this paper, we propose an extension to the cost-sensitive decision trees algorithm, by creating an ensemble of such trees, and combining them using a stacking approach with a cost-sensitive logistic regression. We compare our method with standard machine learning algorithms and state-of-the-art cost-sensitive classification methods using a real credit card fraud dataset provided by a large European card processing company. The results show that our method achieves savings of up to 73.3%, more than 2 percentage points more than a single cost-sensitive decision tree.

Fraud Detection by Stacking Cost-Sensitive Decision Trees from Alejandro Correa Bahnsen, PhD

]]> 4565 7 https://cdn.slidesharecdn.com/ss_thumbnails/2017stackingcsdtacorrea-170926224621-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Maximizing a churn campaigns profitability with cost sensitive machine learning /slideshow/maximizing-a-churn-campaigns-profitability-with-cost-sensitive-machine-learning/79221477 2017maximizingachurncampaignsprofitabilitywithcostsensitivemachinelearning-170828163100
Customer churn predictive modeling deals with predicting the probability of a customer defecting using historical, behavioral and socio-economical information. This tool is of great benefit to subscription based companies allowing them to maximize the results of retention campaigns. The problem of churn predictive modeling has been widely studied by the data mining and machine learning communities. It is usually tackled by using classification algorithms in order to learn the different patterns of both the churners and non-churners. Nevertheless, current state-of-the-art classification algorithms are not well aligned with commercial goals, in the sense that, the models miss to include the real financial costs and benefits during the training and evaluation phases. In the case of churn, evaluating a model based on a traditional measure such as accuracy or predictive power, does not yield to the best results when measured by the actual financial cost, ie. investment per subscriber on a loyalty campaign and the financial impact of failing to detect a real churner versus wrongly predicting a non-churner as a churner. In this presentacion, we present a new cost-sensitive framework for customer churn predictive modeling. First we propose a new financial based measure for evaluating the effectiveness of a churn campaign taking into account the available portfolio of offers, their individual financial cost and probability of offer acceptance depending on the customer profile. Then, using a real-world churn dataset we compare different cost-insensitive and cost-sensitive classification algorithms and measure their effectiveness based on their predictive power and also the cost optimization. The results show that using a cost-sensitive approach yields to an increase in cost savings of up to 26.4 %.]]>
Customer churn predictive modeling deals with predicting the probability of a customer defecting using historical, behavioral and socio-economical information. This tool is of great benefit to subscription based companies allowing them to maximize the results of retention campaigns. The problem of churn predictive modeling has been widely studied by the data mining and machine learning communities. It is usually tackled by using classification algorithms in order to learn the different patterns of both the churners and non-churners. Nevertheless, current state-of-the-art classification algorithms are not well aligned with commercial goals, in the sense that, the models miss to include the real financial costs and benefits during the training and evaluation phases. In the case of churn, evaluating a model based on a traditional measure such as accuracy or predictive power, does not yield to the best results when measured by the actual financial cost, ie. investment per subscriber on a loyalty campaign and the financial impact of failing to detect a real churner versus wrongly predicting a non-churner as a churner. In this presentacion, we present a new cost-sensitive framework for customer churn predictive modeling. First we propose a new financial based measure for evaluating the effectiveness of a churn campaign taking into account the available portfolio of offers, their individual financial cost and probability of offer acceptance depending on the customer profile. Then, using a real-world churn dataset we compare different cost-insensitive and cost-sensitive classification algorithms and measure their effectiveness based on their predictive power and also the cost optimization. The results show that using a cost-sensitive approach yields to an increase in cost savings of up to 26.4 %.]]> Mon, 28 Aug 2017 16:31:00 GMT /slideshow/maximizing-a-churn-campaigns-profitability-with-cost-sensitive-machine-learning/79221477 albahnsen@slideshare.net(albahnsen) Maximizing a churn campaigns profitability with cost sensitive machine learning albahnsen Customer churn predictive modeling deals with predicting the probability of a customer defecting using historical, behavioral and socio-economical information. This tool is of great benefit to subscription based companies allowing them to maximize the results of retention campaigns. The problem of churn predictive modeling has been widely studied by the data mining and machine learning communities. It is usually tackled by using classification algorithms in order to learn the different patterns of both the churners and non-churners. Nevertheless, current state-of-the-art classification algorithms are not well aligned with commercial goals, in the sense that, the models miss to include the real financial costs and benefits during the training and evaluation phases. In the case of churn, evaluating a model based on a traditional measure such as accuracy or predictive power, does not yield to the best results when measured by the actual financial cost, ie. investment per subscriber on a loyalty campaign and the financial impact of failing to detect a real churner versus wrongly predicting a non-churner as a churner. In this presentacion, we present a new cost-sensitive framework for customer churn predictive modeling. First we propose a new financial based measure for evaluating the effectiveness of a churn campaign taking into account the available portfolio of offers, their individual financial cost and probability of offer acceptance depending on the customer profile. Then, using a real-world churn dataset we compare different cost-insensitive and cost-sensitive classification algorithms and measure their effectiveness based on their predictive power and also the cost optimization. The results show that using a cost-sensitive approach yields to an increase in cost savings of up to 26.4 %. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/2017maximizingachurncampaignsprofitabilitywithcostsensitivemachinelearning-170828163100-thumbnail.jpg?width=120&height=120&fit=bounds" /> Customer churn predictive modeling deals with predicting the probability of a customer defecting using historical, behavioral and socio-economical information. This tool is of great benefit to subscription based companies allowing them to maximize the results of retention campaigns. The problem of churn predictive modeling has been widely studied by the data mining and machine learning communities. It is usually tackled by using classification algorithms in order to learn the different patterns of both the churners and non-churners. Nevertheless, current state-of-the-art classification algorithms are not well aligned with commercial goals, in the sense that, the models miss to include the real financial costs and benefits during the training and evaluation phases. In the case of churn, evaluating a model based on a traditional measure such as accuracy or predictive power, does not yield to the best results when measured by the actual financial cost, ie. investment per subscriber on a loyalty campaign and the financial impact of failing to detect a real churner versus wrongly predicting a non-churner as a churner. In this presentacion, we present a new cost-sensitive framework for customer churn predictive modeling. First we propose a new financial based measure for evaluating the effectiveness of a churn campaign taking into account the available portfolio of offers, their individual financial cost and probability of offer acceptance depending on the customer profile. Then, using a real-world churn dataset we compare different cost-insensitive and cost-sensitive classification algorithms and measure their effectiveness based on their predictive power and also the cost optimization. The results show that using a cost-sensitive approach yields to an increase in cost savings of up to 26.4 %.

Maximizing a churn campaigns profitability with cost sensitive machine learning from Alejandro Correa Bahnsen, PhD

]]> 1741 9 https://cdn.slidesharecdn.com/ss_thumbnails/2017maximizingachurncampaignsprofitabilitywithcostsensitivemachinelearning-170828163100-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Classifying Phishing URLs Using Recurrent Neural Networks /slideshow/classifying-phishing-urls-using-recurrent-neural-networks-75314663/75314663 classifyingphishingurlsusingrecurrentneuralnetworks-170423023022
As the technical skills and costs associated with the deployment of phishing attacks decrease, we are witnessing an unprecedented level of scams that push the need for better methods to proactively detect phishing threats. In this work, we explored the use of URLs as input for machine learning models applied for phishing site prediction. In this way, we compared a feature-engineering approach followed by a random forest classifier against a novel method based on recurrent neural networks. We determined that the recurrent neural network approach provides an accuracy rate of 98.7% even without the need of manual feature creation, beating by 5% the random forest method. This means it is a scalable and fast-acting proactive detection system that does not require full content analysis.]]>
As the technical skills and costs associated with the deployment of phishing attacks decrease, we are witnessing an unprecedented level of scams that push the need for better methods to proactively detect phishing threats. In this work, we explored the use of URLs as input for machine learning models applied for phishing site prediction. In this way, we compared a feature-engineering approach followed by a random forest classifier against a novel method based on recurrent neural networks. We determined that the recurrent neural network approach provides an accuracy rate of 98.7% even without the need of manual feature creation, beating by 5% the random forest method. This means it is a scalable and fast-acting proactive detection system that does not require full content analysis.]]> Sun, 23 Apr 2017 02:30:22 GMT /slideshow/classifying-phishing-urls-using-recurrent-neural-networks-75314663/75314663 albahnsen@slideshare.net(albahnsen) Classifying Phishing URLs Using Recurrent Neural Networks albahnsen As the technical skills and costs associated with the deployment of phishing attacks decrease, we are witnessing an unprecedented level of scams that push the need for better methods to proactively detect phishing threats. In this work, we explored the use of URLs as input for machine learning models applied for phishing site prediction. In this way, we compared a feature-engineering approach followed by a random forest classifier against a novel method based on recurrent neural networks. We determined that the recurrent neural network approach provides an accuracy rate of 98.7% even without the need of manual feature creation, beating by 5% the random forest method. This means it is a scalable and fast-acting proactive detection system that does not require full content analysis. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/classifyingphishingurlsusingrecurrentneuralnetworks-170423023022-thumbnail.jpg?width=120&height=120&fit=bounds" /> As the technical skills and costs associated with the deployment of phishing attacks decrease, we are witnessing an unprecedented level of scams that push the need for better methods to proactively detect phishing threats. In this work, we explored the use of URLs as input for machine learning models applied for phishing site prediction. In this way, we compared a feature-engineering approach followed by a random forest classifier against a novel method based on recurrent neural networks. We determined that the recurrent neural network approach provides an accuracy rate of 98.7% even without the need of manual feature creation, beating by 5% the random forest method. This means it is a scalable and fast-acting proactive detection system that does not require full content analysis.

Classifying Phishing URLs Using Recurrent Neural Networks from Alejandro Correa Bahnsen, PhD

]]> 3327 4 https://cdn.slidesharecdn.com/ss_thumbnails/classifyingphishingurlsusingrecurrentneuralnetworks-170423023022-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Demystifying machine learning using lime /slideshow/demystifying-machine-learning-using-lime/71738625 demystifyingmachinelearningusinglime-170203200127
�ݺ�ߣs of my Pycon 2017 short talk "Demystifying machine learning using lime" Jupyter Notebook with code in https://github.com/albahnsen/Talk_Demystifying_Machine_Learning]]>
�ݺ�ߣs of my Pycon 2017 short talk "Demystifying machine learning using lime" Jupyter Notebook with code in https://github.com/albahnsen/Talk_Demystifying_Machine_Learning]]> Fri, 03 Feb 2017 20:01:27 GMT /slideshow/demystifying-machine-learning-using-lime/71738625 albahnsen@slideshare.net(albahnsen) Demystifying machine learning using lime albahnsen �ݺ�ߣs of my Pycon 2017 short talk "Demystifying machine learning using lime" Jupyter Notebook with code in https://github.com/albahnsen/Talk_Demystifying_Machine_Learning <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/demystifyingmachinelearningusinglime-170203200127-thumbnail.jpg?width=120&height=120&fit=bounds" /> �ݺ�ߣs of my Pycon 2017 short talk "Demystifying machine learning using lime" Jupyter Notebook with code in https://github.com/albahnsen/Talk_Demystifying_Machine_Learning

Demystifying machine learning using lime from Alejandro Correa Bahnsen, PhD

]]> 2059 5 https://cdn.slidesharecdn.com/ss_thumbnails/demystifyingmachinelearningusinglime-170203200127-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 1609 Fraud Data Science /slideshow/1609-fraud-data-science/65699522 zq91zc3ltewq5jgmt94y-signature-4eabb1c7e815073ebad4999250e2502a073c9a4e830bb46681ba0f4be909ce8f-poli-160905123724
Online fraud costs the global economy more than $400 billion, with more than 800 million personal records stolen in 2013 alone. Increasingly, fraud has diversified to different digital channels, including mobile and online payments, creating new challenges as innovative fraud patterns emerge. Hence it is still a challenge to find effective methods to mitigate fraud. Existing solutions include simple if-then rules and classical machine learning algorithms. From an academic perspective, credit card fraud detection is a standard classification problem, in which historical transaction data is used to predict future frauds. However, practical aspects make the problem more complex. Indeed, existent comparison measures lack a realistic representation of monetary gains and losses, which is necessary for effective fraud detection. Moreover, there is an enormous amount of transactions from which only a tiny part are frauds, which implies a huge class imbalance. Additionally, a real fraud detection system is required to give a response in milliseconds. This criterion needs to be taken into account in the modeling process in order for the system to be successfully implemented. To solve these problems, in this presentation two recently proposed algorithms are compared: Bayes minimum risk and example-dependent cost-sensitive decision tree. These methods are compared with state of the art algorithms and shows significant improvements measured by financial savings.]]>
Online fraud costs the global economy more than $400 billion, with more than 800 million personal records stolen in 2013 alone. Increasingly, fraud has diversified to different digital channels, including mobile and online payments, creating new challenges as innovative fraud patterns emerge. Hence it is still a challenge to find effective methods to mitigate fraud. Existing solutions include simple if-then rules and classical machine learning algorithms. From an academic perspective, credit card fraud detection is a standard classification problem, in which historical transaction data is used to predict future frauds. However, practical aspects make the problem more complex. Indeed, existent comparison measures lack a realistic representation of monetary gains and losses, which is necessary for effective fraud detection. Moreover, there is an enormous amount of transactions from which only a tiny part are frauds, which implies a huge class imbalance. Additionally, a real fraud detection system is required to give a response in milliseconds. This criterion needs to be taken into account in the modeling process in order for the system to be successfully implemented. To solve these problems, in this presentation two recently proposed algorithms are compared: Bayes minimum risk and example-dependent cost-sensitive decision tree. These methods are compared with state of the art algorithms and shows significant improvements measured by financial savings.]]> Mon, 05 Sep 2016 12:37:23 GMT /slideshow/1609-fraud-data-science/65699522 albahnsen@slideshare.net(albahnsen) 1609 Fraud Data Science albahnsen Online fraud costs the global economy more than $400 billion, with more than 800 million personal records stolen in 2013 alone. Increasingly, fraud has diversified to different digital channels, including mobile and online payments, creating new challenges as innovative fraud patterns emerge. Hence it is still a challenge to find effective methods to mitigate fraud. Existing solutions include simple if-then rules and classical machine learning algorithms. From an academic perspective, credit card fraud detection is a standard classification problem, in which historical transaction data is used to predict future frauds. However, practical aspects make the problem more complex. Indeed, existent comparison measures lack a realistic representation of monetary gains and losses, which is necessary for effective fraud detection. Moreover, there is an enormous amount of transactions from which only a tiny part are frauds, which implies a huge class imbalance. Additionally, a real fraud detection system is required to give a response in milliseconds. This criterion needs to be taken into account in the modeling process in order for the system to be successfully implemented. To solve these problems, in this presentation two recently proposed algorithms are compared: Bayes minimum risk and example-dependent cost-sensitive decision tree. These methods are compared with state of the art algorithms and shows significant improvements measured by financial savings. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/zq91zc3ltewq5jgmt94y-signature-4eabb1c7e815073ebad4999250e2502a073c9a4e830bb46681ba0f4be909ce8f-poli-160905123724-thumbnail.jpg?width=120&height=120&fit=bounds" /> Online fraud costs the global economy more than $400 billion, with more than 800 million personal records stolen in 2013 alone. Increasingly, fraud has diversified to different digital channels, including mobile and online payments, creating new challenges as innovative fraud patterns emerge. Hence it is still a challenge to find effective methods to mitigate fraud. Existing solutions include simple if-then rules and classical machine learning algorithms. From an academic perspective, credit card fraud detection is a standard classification problem, in which historical transaction data is used to predict future frauds. However, practical aspects make the problem more complex. Indeed, existent comparison measures lack a realistic representation of monetary gains and losses, which is necessary for effective fraud detection. Moreover, there is an enormous amount of transactions from which only a tiny part are frauds, which implies a huge class imbalance. Additionally, a real fraud detection system is required to give a response in milliseconds. This criterion needs to be taken into account in the modeling process in order for the system to be successfully implemented. To solve these problems, in this presentation two recently proposed algorithms are compared: Bayes minimum risk and example-dependent cost-sensitive decision tree. These methods are compared with state of the art algorithms and shows significant improvements measured by financial savings.

1609 Fraud Data Science from Alejandro Correa Bahnsen, PhD

]]> 2517 4 https://cdn.slidesharecdn.com/ss_thumbnails/zq91zc3ltewq5jgmt94y-signature-4eabb1c7e815073ebad4999250e2502a073c9a4e830bb46681ba0f4be909ce8f-poli-160905123724-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Modern Data Science /slideshow/modern-data-science/61871966 kglc7hlszupqpasjf6hq-signature-e682d3f3780a3e6abd7eca308f557c736f573e06774b4cf8c0eda9b8cf976823-poli-160510171522
Presentation on Modern Data Science Data scientists are in high demand. There is simply not enough talent to fill the jobs. Why? Because the sexiest job of 21th century requires a mixture of broad, multidisciplinary skills ranging from an intersection of mathematics, statistics, computer science, communication and business. Finding a data scientist is hard. Finding people who understand who a data scientist is, is equally hard. Check the video in spanish here :https://www.youtube.com/watch?v=R3jeBHLLiiM]]>
Presentation on Modern Data Science Data scientists are in high demand. There is simply not enough talent to fill the jobs. Why? Because the sexiest job of 21th century requires a mixture of broad, multidisciplinary skills ranging from an intersection of mathematics, statistics, computer science, communication and business. Finding a data scientist is hard. Finding people who understand who a data scientist is, is equally hard. Check the video in spanish here :https://www.youtube.com/watch?v=R3jeBHLLiiM]]> Tue, 10 May 2016 17:15:22 GMT /slideshow/modern-data-science/61871966 albahnsen@slideshare.net(albahnsen) Modern Data Science albahnsen Presentation on Modern Data Science Data scientists are in high demand. There is simply not enough talent to fill the jobs. Why? Because the sexiest job of 21th century requires a mixture of broad, multidisciplinary skills ranging from an intersection of mathematics, statistics, computer science, communication and business. Finding a data scientist is hard. Finding people who understand who a data scientist is, is equally hard. Check the video in spanish here :https://www.youtube.com/watch?v=R3jeBHLLiiM <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/kglc7hlszupqpasjf6hq-signature-e682d3f3780a3e6abd7eca308f557c736f573e06774b4cf8c0eda9b8cf976823-poli-160510171522-thumbnail.jpg?width=120&height=120&fit=bounds" /> Presentation on Modern Data Science Data scientists are in high demand. There is simply not enough talent to fill the jobs. Why? Because the sexiest job of 21th century requires a mixture of broad, multidisciplinary skills ranging from an intersection of mathematics, statistics, computer science, communication and business. Finding a data scientist is hard. Finding people who understand who a data scientist is, is equally hard. Check the video in spanish here :https://www.youtube.com/watch?v=R3jeBHLLiiM

Modern Data Science from Alejandro Correa Bahnsen, PhD

]]> 7156 12 https://cdn.slidesharecdn.com/ss_thumbnails/kglc7hlszupqpasjf6hq-signature-e682d3f3780a3e6abd7eca308f557c736f573e06774b4cf8c0eda9b8cf976823-poli-160510171522-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Fraud Detection with Cost-Sensitive Predictive Analytics /albahnsen/1510-fraud-analyticslong qnt7dkbbrv6zap0ouflw-signature-b28e4b4b660cd47b819b06da91d3e7ede31289a505adaf79c828de1b2bef6891-poli-160118132522
Online fraud costs the global economy more than $400 billion, with more than 800 million personal records stolen in 2013 alone. Increasingly, fraud has diversified to different digital channels, including mobile and online payments, creating new challenges as innovative fraud patterns emerge. Hence it is still a challenge to find effective methods to mitigate fraud. Existing solutions include simple if-then rules and classical machine learning algorithms. From an academic perspective, credit card fraud detection is a standard classification problem, in which historical transaction data is used to predict future frauds. However, practical aspects make the problem more complex. Indeed, existent comparison measures lack a realistic representation of monetary gains and losses, which is necessary for effective fraud detection. Moreover, there is an enormous amount of transactions from which only a tiny part are frauds, which implies a huge class imbalance. Additionally, a real fraud detection system is required to give a response in milliseconds. This criterion needs to be taken into account in the modeling process in order for the system to be successfully implemented. To solve these problems, in this presentation two recently proposed algorithms are compared: Bayes minimum risk and example-dependent cost-sensitive decision tree. These methods are compared with state of the art algorithms and shows significant improvements measured by financial savings.]]>
Online fraud costs the global economy more than $400 billion, with more than 800 million personal records stolen in 2013 alone. Increasingly, fraud has diversified to different digital channels, including mobile and online payments, creating new challenges as innovative fraud patterns emerge. Hence it is still a challenge to find effective methods to mitigate fraud. Existing solutions include simple if-then rules and classical machine learning algorithms. From an academic perspective, credit card fraud detection is a standard classification problem, in which historical transaction data is used to predict future frauds. However, practical aspects make the problem more complex. Indeed, existent comparison measures lack a realistic representation of monetary gains and losses, which is necessary for effective fraud detection. Moreover, there is an enormous amount of transactions from which only a tiny part are frauds, which implies a huge class imbalance. Additionally, a real fraud detection system is required to give a response in milliseconds. This criterion needs to be taken into account in the modeling process in order for the system to be successfully implemented. To solve these problems, in this presentation two recently proposed algorithms are compared: Bayes minimum risk and example-dependent cost-sensitive decision tree. These methods are compared with state of the art algorithms and shows significant improvements measured by financial savings.]]> Mon, 18 Jan 2016 13:25:22 GMT /albahnsen/1510-fraud-analyticslong albahnsen@slideshare.net(albahnsen) Fraud Detection with Cost-Sensitive Predictive Analytics albahnsen Online fraud costs the global economy more than $400 billion, with more than 800 million personal records stolen in 2013 alone. Increasingly, fraud has diversified to different digital channels, including mobile and online payments, creating new challenges as innovative fraud patterns emerge. Hence it is still a challenge to find effective methods to mitigate fraud. Existing solutions include simple if-then rules and classical machine learning algorithms. From an academic perspective, credit card fraud detection is a standard classification problem, in which historical transaction data is used to predict future frauds. However, practical aspects make the problem more complex. Indeed, existent comparison measures lack a realistic representation of monetary gains and losses, which is necessary for effective fraud detection. Moreover, there is an enormous amount of transactions from which only a tiny part are frauds, which implies a huge class imbalance. Additionally, a real fraud detection system is required to give a response in milliseconds. This criterion needs to be taken into account in the modeling process in order for the system to be successfully implemented. To solve these problems, in this presentation two recently proposed algorithms are compared: Bayes minimum risk and example-dependent cost-sensitive decision tree. These methods are compared with state of the art algorithms and shows significant improvements measured by financial savings. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/qnt7dkbbrv6zap0ouflw-signature-b28e4b4b660cd47b819b06da91d3e7ede31289a505adaf79c828de1b2bef6891-poli-160118132522-thumbnail.jpg?width=120&height=120&fit=bounds" /> Online fraud costs the global economy more than $400 billion, with more than 800 million personal records stolen in 2013 alone. Increasingly, fraud has diversified to different digital channels, including mobile and online payments, creating new challenges as innovative fraud patterns emerge. Hence it is still a challenge to find effective methods to mitigate fraud. Existing solutions include simple if-then rules and classical machine learning algorithms. From an academic perspective, credit card fraud detection is a standard classification problem, in which historical transaction data is used to predict future frauds. However, practical aspects make the problem more complex. Indeed, existent comparison measures lack a realistic representation of monetary gains and losses, which is necessary for effective fraud detection. Moreover, there is an enormous amount of transactions from which only a tiny part are frauds, which implies a huge class imbalance. Additionally, a real fraud detection system is required to give a response in milliseconds. This criterion needs to be taken into account in the modeling process in order for the system to be successfully implemented. To solve these problems, in this presentation two recently proposed algorithms are compared: Bayes minimum risk and example-dependent cost-sensitive decision tree. These methods are compared with state of the art algorithms and shows significant improvements measured by financial savings.

Fraud Detection with Cost-Sensitive Predictive Analytics from Alejandro Correa Bahnsen, PhD

]]> 3526 6 https://cdn.slidesharecdn.com/ss_thumbnails/qnt7dkbbrv6zap0ouflw-signature-b28e4b4b660cd47b819b06da91d3e7ede31289a505adaf79c828de1b2bef6891-poli-160118132522-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 PhD Defense - Example-Dependent Cost-Sensitive Classification /slideshow/phd-defense-exampledependent-costsensitive-classification/53092921 thesisexampledependentcostsensitiveclassificationslides-150923041551-lva1-app6892
�ݺ�ߣs from my PhD defense Example-Dependent Cost-Sensitive Classification Applications in Financial Risk Modeling and Marketing Analytics https://github.com/albahnsen/phd-thesis]]>
�ݺ�ߣs from my PhD defense Example-Dependent Cost-Sensitive Classification Applications in Financial Risk Modeling and Marketing Analytics https://github.com/albahnsen/phd-thesis]]> Wed, 23 Sep 2015 04:15:51 GMT /slideshow/phd-defense-exampledependent-costsensitive-classification/53092921 albahnsen@slideshare.net(albahnsen) PhD Defense - Example-Dependent Cost-Sensitive Classification albahnsen �ݺ�ߣs from my PhD defense Example-Dependent Cost-Sensitive Classification Applications in Financial Risk Modeling and Marketing Analytics https://github.com/albahnsen/phd-thesis <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/thesisexampledependentcostsensitiveclassificationslides-150923041551-lva1-app6892-thumbnail.jpg?width=120&height=120&fit=bounds" /> �ݺ�ߣs from my PhD defense Example-Dependent Cost-Sensitive Classification Applications in Financial Risk Modeling and Marketing Analytics https://github.com/albahnsen/phd-thesis

PhD Defense - Example-Dependent Cost-Sensitive Classification from Alejandro Correa Bahnsen, PhD

]]> 3234 13 https://cdn.slidesharecdn.com/ss_thumbnails/thesisexampledependentcostsensitiveclassificationslides-150923041551-lva1-app6892-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Ensembles of example dependent cost-sensitive decision trees slides /slideshow/ensembles-of-example-dependent-costsensitive-decision-trees-slides/48601218 ensemblesofexample-dependentcost-sensitivedecisiontreesslides-150526100803-lva1-app6892
�ݺ�ߣs of the paper http://arxiv.org/abs/1505.04637 source code is available at https://github.com/albahnsen/CostSensitiveClassification/blob/master/costcla/models/cost_tree.py#L15 Abstract: Several real-world classification problems are example-dependent cost-sensitive in nature, where the costs due to misclassification vary between examples and not only within classes. However, standard classification methods do not take these costs into account, and assume a constant cost of misclassification errors. In previous works, some methods that take into account the financial costs into the training of different algorithms have been proposed, with the example-dependent cost-sensitive decision tree algorithm being the one that gives the highest savings. In this paper we propose a new framework of ensembles of example-dependent cost-sensitive decision-trees. The framework consists in creating different example-dependent cost-sensitive decision trees on random subsamples of the training set, and then combining them using three different combination approaches. Moreover, we propose two new cost-sensitive combination approaches; cost-sensitive weighted voting and cost-sensitive stacking, the latter being based on the cost-sensitive logistic regression method. Finally, using five different databases, from four real-world applications: credit card fraud detection, churn modeling, credit scoring and direct marketing, we evaluate the proposed method against state-of-the-art example-dependent cost-sensitive techniques, namely, cost-proportionate sampling, Bayes minimum risk and cost-sensitive decision trees. The results show that the proposed algorithms have better results for all databases, in the sense of higher savings.]]>
�ݺ�ߣs of the paper http://arxiv.org/abs/1505.04637 source code is available at https://github.com/albahnsen/CostSensitiveClassification/blob/master/costcla/models/cost_tree.py#L15 Abstract: Several real-world classification problems are example-dependent cost-sensitive in nature, where the costs due to misclassification vary between examples and not only within classes. However, standard classification methods do not take these costs into account, and assume a constant cost of misclassification errors. In previous works, some methods that take into account the financial costs into the training of different algorithms have been proposed, with the example-dependent cost-sensitive decision tree algorithm being the one that gives the highest savings. In this paper we propose a new framework of ensembles of example-dependent cost-sensitive decision-trees. The framework consists in creating different example-dependent cost-sensitive decision trees on random subsamples of the training set, and then combining them using three different combination approaches. Moreover, we propose two new cost-sensitive combination approaches; cost-sensitive weighted voting and cost-sensitive stacking, the latter being based on the cost-sensitive logistic regression method. Finally, using five different databases, from four real-world applications: credit card fraud detection, churn modeling, credit scoring and direct marketing, we evaluate the proposed method against state-of-the-art example-dependent cost-sensitive techniques, namely, cost-proportionate sampling, Bayes minimum risk and cost-sensitive decision trees. The results show that the proposed algorithms have better results for all databases, in the sense of higher savings.]]> Tue, 26 May 2015 10:08:03 GMT /slideshow/ensembles-of-example-dependent-costsensitive-decision-trees-slides/48601218 albahnsen@slideshare.net(albahnsen) Ensembles of example dependent cost-sensitive decision trees slides albahnsen �ݺ�ߣs of the paper http://arxiv.org/abs/1505.04637 source code is available at https://github.com/albahnsen/CostSensitiveClassification/blob/master/costcla/models/cost_tree.py#L15 Abstract: Several real-world classification problems are example-dependent cost-sensitive in nature, where the costs due to misclassification vary between examples and not only within classes. However, standard classification methods do not take these costs into account, and assume a constant cost of misclassification errors. In previous works, some methods that take into account the financial costs into the training of different algorithms have been proposed, with the example-dependent cost-sensitive decision tree algorithm being the one that gives the highest savings. In this paper we propose a new framework of ensembles of example-dependent cost-sensitive decision-trees. The framework consists in creating different example-dependent cost-sensitive decision trees on random subsamples of the training set, and then combining them using three different combination approaches. Moreover, we propose two new cost-sensitive combination approaches; cost-sensitive weighted voting and cost-sensitive stacking, the latter being based on the cost-sensitive logistic regression method. Finally, using five different databases, from four real-world applications: credit card fraud detection, churn modeling, credit scoring and direct marketing, we evaluate the proposed method against state-of-the-art example-dependent cost-sensitive techniques, namely, cost-proportionate sampling, Bayes minimum risk and cost-sensitive decision trees. The results show that the proposed algorithms have better results for all databases, in the sense of higher savings. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/ensemblesofexample-dependentcost-sensitivedecisiontreesslides-150526100803-lva1-app6892-thumbnail.jpg?width=120&height=120&fit=bounds" /> �ݺ�ߣs of the paper http://arxiv.org/abs/1505.04637 source code is available at https://github.com/albahnsen/CostSensitiveClassification/blob/master/costcla/models/cost_tree.py#L15 Abstract: Several real-world classification problems are example-dependent cost-sensitive in nature, where the costs due to misclassification vary between examples and not only within classes. However, standard classification methods do not take these costs into account, and assume a constant cost of misclassification errors. In previous works, some methods that take into account the financial costs into the training of different algorithms have been proposed, with the example-dependent cost-sensitive decision tree algorithm being the one that gives the highest savings. In this paper we propose a new framework of ensembles of example-dependent cost-sensitive decision-trees. The framework consists in creating different example-dependent cost-sensitive decision trees on random subsamples of the training set, and then combining them using three different combination approaches. Moreover, we propose two new cost-sensitive combination approaches; cost-sensitive weighted voting and cost-sensitive stacking, the latter being based on the cost-sensitive logistic regression method. Finally, using five different databases, from four real-world applications: credit card fraud detection, churn modeling, credit scoring and direct marketing, we evaluate the proposed method against state-of-the-art example-dependent cost-sensitive techniques, namely, cost-proportionate sampling, Bayes minimum risk and cost-sensitive decision trees. The results show that the proposed algorithms have better results for all databases, in the sense of higher savings.

Ensembles of example dependent cost-sensitive decision trees slides from Alejandro Correa Bahnsen, PhD

]]> 5311 4 https://cdn.slidesharecdn.com/ss_thumbnails/ensemblesofexample-dependentcost-sensitivedecisiontreesslides-150526100803-lva1-app6892-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Fraud analytics detección y prevención de fraudes en la era del big data slideshare https://es.slideshare.net/slideshow/fraud-analytics-deteccin-y-prevencin-de-fraudes-en-la-era-del-big-data-slideshare/39976262 fraudanalyticsdeteccinyprevencindefraudesenlaeradelbigdataslideshare-141007102519-conversion-gate01
Fraud Analytics: Detección y prevención de fraudes en la era del BigData Durante el 2012 el nivel de fraude en tarjeta de crédito llego a 11.3 billones de dólares, un aumento de casi un 15% comparado con el 2011, esto demuestra el problema que el fraude representa no solo a las instituciones financieras sino también para la sociedad. Tradicionalmente la prevención del fraude consistía en proteger físicamente la infraestructura, sin embargo con cada vez más medios y canales de pago, la información financiera se ha vuelto cada vez más susceptible a ser hurtada. La siguiente opción para prevenir y controlar el fraude consiste en determinar si una transacción está siendo realizada por el cliente de acuerdo con sus patrones históricos de comportamiento. Este es el enfoque de Fraud Analytics. En esta presentación se mostrara cómo es posible por medio de Fraud Analytics, determinar la probabilidad que una transacción sea o no realizada por el cliente, utilizando la información de compra de los clientes, sus interacciones con la entidad financiera, y por medio de análisis de redes sociales. Adicionalmente, se discutirán y compararan los resultados de las comúnmente utilizadas reglas de decisión y modelos avanzados de inteligencia artificial. ------------------------------------- Alejandro Correa Bahnsen ------------------------------------- Ingeniero Industrial con Maestría en Ingeniería Industrial de la Universidad de los Andes. Candidato a Doctorado en inteligencia artificial de la Universidad de Luxemburgo. Actualmente se encuentra trabajando en SIX, uno de los operadores de tarjeta de crédito más grande de Europa, desarrollando un sistema inteligente para la prevención de fraude. Experiencia como profesor de analytics y econometría en las universidades de Luxemburgo y de los Andes, respectivamente. Conferencista de analytics en SAS Analytics (Orlando, Las Vegas, Londres, Frankfurt), SAS Global Forum (Orlando, San Francisco), IEEE International Conference on Data Mining (Vancouver, Bruselas, Dallas), IEEE International Conference on Machine Learning and Applications (Miami, Detroit) y European Conference on Data Analysis (Luxemburgo). Fundador de la comunidad Data Science Luxembourg y organizador del workshop IEEE Data Mining Case Studies (Dallas).]]>
Fraud Analytics: Detección y prevención de fraudes en la era del BigData Durante el 2012 el nivel de fraude en tarjeta de crédito llego a 11.3 billones de dólares, un aumento de casi un 15% comparado con el 2011, esto demuestra el problema que el fraude representa no solo a las instituciones financieras sino también para la sociedad. Tradicionalmente la prevención del fraude consistía en proteger físicamente la infraestructura, sin embargo con cada vez más medios y canales de pago, la información financiera se ha vuelto cada vez más susceptible a ser hurtada. La siguiente opción para prevenir y controlar el fraude consiste en determinar si una transacción está siendo realizada por el cliente de acuerdo con sus patrones históricos de comportamiento. Este es el enfoque de Fraud Analytics. En esta presentación se mostrara cómo es posible por medio de Fraud Analytics, determinar la probabilidad que una transacción sea o no realizada por el cliente, utilizando la información de compra de los clientes, sus interacciones con la entidad financiera, y por medio de análisis de redes sociales. Adicionalmente, se discutirán y compararan los resultados de las comúnmente utilizadas reglas de decisión y modelos avanzados de inteligencia artificial. ------------------------------------- Alejandro Correa Bahnsen ------------------------------------- Ingeniero Industrial con Maestría en Ingeniería Industrial de la Universidad de los Andes. Candidato a Doctorado en inteligencia artificial de la Universidad de Luxemburgo. Actualmente se encuentra trabajando en SIX, uno de los operadores de tarjeta de crédito más grande de Europa, desarrollando un sistema inteligente para la prevención de fraude. Experiencia como profesor de analytics y econometría en las universidades de Luxemburgo y de los Andes, respectivamente. Conferencista de analytics en SAS Analytics (Orlando, Las Vegas, Londres, Frankfurt), SAS Global Forum (Orlando, San Francisco), IEEE International Conference on Data Mining (Vancouver, Bruselas, Dallas), IEEE International Conference on Machine Learning and Applications (Miami, Detroit) y European Conference on Data Analysis (Luxemburgo). Fundador de la comunidad Data Science Luxembourg y organizador del workshop IEEE Data Mining Case Studies (Dallas).]]> Tue, 07 Oct 2014 10:25:19 GMT https://es.slideshare.net/slideshow/fraud-analytics-deteccin-y-prevencin-de-fraudes-en-la-era-del-big-data-slideshare/39976262 albahnsen@slideshare.net(albahnsen) Fraud analytics detección y prevención de fraudes en la era del big data slideshare albahnsen Fraud Analytics: Detección y prevención de fraudes en la era del BigData Durante el 2012 el nivel de fraude en tarjeta de crédito llego a 11.3 billones de dólares, un aumento de casi un 15% comparado con el 2011, esto demuestra el problema que el fraude representa no solo a las instituciones financieras sino también para la sociedad. Tradicionalmente la prevención del fraude consistía en proteger físicamente la infraestructura, sin embargo con cada vez más medios y canales de pago, la información financiera se ha vuelto cada vez más susceptible a ser hurtada. La siguiente opción para prevenir y controlar el fraude consiste en determinar si una transacción está siendo realizada por el cliente de acuerdo con sus patrones históricos de comportamiento. Este es el enfoque de Fraud Analytics. En esta presentación se mostrara cómo es posible por medio de Fraud Analytics, determinar la probabilidad que una transacción sea o no realizada por el cliente, utilizando la información de compra de los clientes, sus interacciones con la entidad financiera, y por medio de análisis de redes sociales. Adicionalmente, se discutirán y compararan los resultados de las comúnmente utilizadas reglas de decisión y modelos avanzados de inteligencia artificial. ------------------------------------- Alejandro Correa Bahnsen ------------------------------------- Ingeniero Industrial con Maestría en Ingeniería Industrial de la Universidad de los Andes. Candidato a Doctorado en inteligencia artificial de la Universidad de Luxemburgo. Actualmente se encuentra trabajando en SIX, uno de los operadores de tarjeta de crédito más grande de Europa, desarrollando un sistema inteligente para la prevención de fraude. Experiencia como profesor de analytics y econometría en las universidades de Luxemburgo y de los Andes, respectivamente. Conferencista de analytics en SAS Analytics (Orlando, Las Vegas, Londres, Frankfurt), SAS Global Forum (Orlando, San Francisco), IEEE International Conference on Data Mining (Vancouver, Bruselas, Dallas), IEEE International Conference on Machine Learning and Applications (Miami, Detroit) y European Conference on Data Analysis (Luxemburgo). Fundador de la comunidad Data Science Luxembourg y organizador del workshop IEEE Data Mining Case Studies (Dallas). <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/fraudanalyticsdeteccinyprevencindefraudesenlaeradelbigdataslideshare-141007102519-conversion-gate01-thumbnail.jpg?width=120&height=120&fit=bounds" /> Fraud Analytics: Detección y prevención de fraudes en la era del BigData Durante el 2012 el nivel de fraude en tarjeta de crédito llego a 11.3 billones de dólares, un aumento de casi un 15% comparado con el 2011, esto demuestra el problema que el fraude representa no solo a las instituciones financieras sino también para la sociedad. Tradicionalmente la prevención del fraude consistía en proteger físicamente la infraestructura, sin embargo con cada vez más medios y canales de pago, la información financiera se ha vuelto cada vez más susceptible a ser hurtada. La siguiente opción para prevenir y controlar el fraude consiste en determinar si una transacción está siendo realizada por el cliente de acuerdo con sus patrones históricos de comportamiento. Este es el enfoque de Fraud Analytics. En esta presentación se mostrara cómo es posible por medio de Fraud Analytics, determinar la probabilidad que una transacción sea o no realizada por el cliente, utilizando la información de compra de los clientes, sus interacciones con la entidad financiera, y por medio de análisis de redes sociales. Adicionalmente, se discutirán y compararan los resultados de las comúnmente utilizadas reglas de decisión y modelos avanzados de inteligencia artificial. ------------------------------------- Alejandro Correa Bahnsen ------------------------------------- Ingeniero Industrial con Maestría en Ingeniería Industrial de la Universidad de los Andes. Candidato a Doctorado en inteligencia artificial de la Universidad de Luxemburgo. Actualmente se encuentra trabajando en SIX, uno de los operadores de tarjeta de crédito más grande de Europa, desarrollando un sistema inteligente para la prevención de fraude. Experiencia como profesor de analytics y econometría en las universidades de Luxemburgo y de los Andes, respectivamente. Conferencista de analytics en SAS Analytics (Orlando, Las Vegas, Londres, Frankfurt), SAS Global Forum (Orlando, San Francisco), IEEE International Conference on Data Mining (Vancouver, Bruselas, Dallas), IEEE International Conference on Machine Learning and Applications (Miami, Detroit) y European Conference on Data Analysis (Luxemburgo). Fundador de la comunidad Data Science Luxembourg y organizador del workshop IEEE Data Mining Case Studies (Dallas).

from Alejandro Correa Bahnsen, PhD

]]> 3091 75 https://cdn.slidesharecdn.com/ss_thumbnails/fraudanalyticsdeteccinyprevencindefraudesenlaeradelbigdataslideshare-141007102519-conversion-gate01-thumbnail.jpg?width=120&height=120&fit=bounds presentation 000000 http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Analytics - compitiendo en la era de la informacion https://es.slideshare.net/albahnsen/analytics-compitiendo-en-la-era-de-la-informacion analyticscompitiendoenlaeradelainformacionc-140924114102-phpapp01
Analytics: Compitiendo en la era de la información En años recientes el mundo ha entrado en la era de la información, la evolución de la tecnología y el desarrollo de las redes sociales ha permitido a las compañías obtener más información sobre el comportamiento de sus clientes. Adicionalmente, los sistemas se han vuelto mas eficientes y económicos, dando la oportunidad a las empresas de almacenar gran cantidad de datos. Sin embargo, toda esta información solo almacenada no genera valor agregado para las empresas; entonces la pregunta es como obtener beneficios y tomar decisiones mas informadas usando los datos recolectados? La respuesta es analytics. Analytics es el uso de métodos y herramientas para entender la información y tomar decisiones más precisas. Permite a las compañías hacer predicciones de comportamientos, identificar potenciales clientes, crear segmentaciones inteligentes, dirigir de una manera eficiente campañas publicitarias, identificar riesgos y anticipar cambios en el mercado. Así, por medio de analytics las compañías logran diferenciarse más de sus competidores y entender mejor las necesidades de sus clientes. Alejandro Correa Bahnsen Ingeniero Industrial con Maestría en Ingeniería Industrial de la Universidad de los Andes. Candidato a Doctorado en inteligencia artificial de la Universidad de Luxemburgo. Actualmente se encuentra trabajando en Cetrel, el operador de tarjeta de crédito mas grande de Europa, desarrollando un sistema inteligente para la prevención de fraude. Experiencia como profesor de analytics y econometría en las universidades de Luxemburgo y de los Andes, respectivamente. Conferencista de analytics en SAS Analytics (Orlando, Las Vegas, Londres), SAS Global Forum (Orlando, San Francisco), IEEE International Conference on Data Mining (Vancouver, Bruselas), IEEE International Conference on Machine Learning and Applications (Miami) y European Conference on Data Analysis (Luxemburgo). Fundador de la comunidad Data Science Luxembourg y organizador del workshop IEEE Data Mining Case Studies (Dallas).]]>
Analytics: Compitiendo en la era de la información En años recientes el mundo ha entrado en la era de la información, la evolución de la tecnología y el desarrollo de las redes sociales ha permitido a las compañías obtener más información sobre el comportamiento de sus clientes. Adicionalmente, los sistemas se han vuelto mas eficientes y económicos, dando la oportunidad a las empresas de almacenar gran cantidad de datos. Sin embargo, toda esta información solo almacenada no genera valor agregado para las empresas; entonces la pregunta es como obtener beneficios y tomar decisiones mas informadas usando los datos recolectados? La respuesta es analytics. Analytics es el uso de métodos y herramientas para entender la información y tomar decisiones más precisas. Permite a las compañías hacer predicciones de comportamientos, identificar potenciales clientes, crear segmentaciones inteligentes, dirigir de una manera eficiente campañas publicitarias, identificar riesgos y anticipar cambios en el mercado. Así, por medio de analytics las compañías logran diferenciarse más de sus competidores y entender mejor las necesidades de sus clientes. Alejandro Correa Bahnsen Ingeniero Industrial con Maestría en Ingeniería Industrial de la Universidad de los Andes. Candidato a Doctorado en inteligencia artificial de la Universidad de Luxemburgo. Actualmente se encuentra trabajando en Cetrel, el operador de tarjeta de crédito mas grande de Europa, desarrollando un sistema inteligente para la prevención de fraude. Experiencia como profesor de analytics y econometría en las universidades de Luxemburgo y de los Andes, respectivamente. Conferencista de analytics en SAS Analytics (Orlando, Las Vegas, Londres), SAS Global Forum (Orlando, San Francisco), IEEE International Conference on Data Mining (Vancouver, Bruselas), IEEE International Conference on Machine Learning and Applications (Miami) y European Conference on Data Analysis (Luxemburgo). Fundador de la comunidad Data Science Luxembourg y organizador del workshop IEEE Data Mining Case Studies (Dallas).]]> Wed, 24 Sep 2014 11:41:02 GMT https://es.slideshare.net/albahnsen/analytics-compitiendo-en-la-era-de-la-informacion albahnsen@slideshare.net(albahnsen) Analytics - compitiendo en la era de la informacion albahnsen Analytics: Compitiendo en la era de la información En años recientes el mundo ha entrado en la era de la información, la evolución de la tecnología y el desarrollo de las redes sociales ha permitido a las compañías obtener más información sobre el comportamiento de sus clientes. Adicionalmente, los sistemas se han vuelto mas eficientes y económicos, dando la oportunidad a las empresas de almacenar gran cantidad de datos. Sin embargo, toda esta información solo almacenada no genera valor agregado para las empresas; entonces la pregunta es como obtener beneficios y tomar decisiones mas informadas usando los datos recolectados? La respuesta es analytics. Analytics es el uso de métodos y herramientas para entender la información y tomar decisiones más precisas. Permite a las compañías hacer predicciones de comportamientos, identificar potenciales clientes, crear segmentaciones inteligentes, dirigir de una manera eficiente campañas publicitarias, identificar riesgos y anticipar cambios en el mercado. Así, por medio de analytics las compañías logran diferenciarse más de sus competidores y entender mejor las necesidades de sus clientes. Alejandro Correa Bahnsen Ingeniero Industrial con Maestría en Ingeniería Industrial de la Universidad de los Andes. Candidato a Doctorado en inteligencia artificial de la Universidad de Luxemburgo. Actualmente se encuentra trabajando en Cetrel, el operador de tarjeta de crédito mas grande de Europa, desarrollando un sistema inteligente para la prevención de fraude. Experiencia como profesor de analytics y econometría en las universidades de Luxemburgo y de los Andes, respectivamente. Conferencista de analytics en SAS Analytics (Orlando, Las Vegas, Londres), SAS Global Forum (Orlando, San Francisco), IEEE International Conference on Data Mining (Vancouver, Bruselas), IEEE International Conference on Machine Learning and Applications (Miami) y European Conference on Data Analysis (Luxemburgo). Fundador de la comunidad Data Science Luxembourg y organizador del workshop IEEE Data Mining Case Studies (Dallas). <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/analyticscompitiendoenlaeradelainformacionc-140924114102-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds" /> Analytics: Compitiendo en la era de la información En años recientes el mundo ha entrado en la era de la información, la evolución de la tecnología y el desarrollo de las redes sociales ha permitido a las compañías obtener más información sobre el comportamiento de sus clientes. Adicionalmente, los sistemas se han vuelto mas eficientes y económicos, dando la oportunidad a las empresas de almacenar gran cantidad de datos. Sin embargo, toda esta información solo almacenada no genera valor agregado para las empresas; entonces la pregunta es como obtener beneficios y tomar decisiones mas informadas usando los datos recolectados? La respuesta es analytics. Analytics es el uso de métodos y herramientas para entender la información y tomar decisiones más precisas. Permite a las compañías hacer predicciones de comportamientos, identificar potenciales clientes, crear segmentaciones inteligentes, dirigir de una manera eficiente campañas publicitarias, identificar riesgos y anticipar cambios en el mercado. Así, por medio de analytics las compañías logran diferenciarse más de sus competidores y entender mejor las necesidades de sus clientes. Alejandro Correa Bahnsen Ingeniero Industrial con Maestría en Ingeniería Industrial de la Universidad de los Andes. Candidato a Doctorado en inteligencia artificial de la Universidad de Luxemburgo. Actualmente se encuentra trabajando en Cetrel, el operador de tarjeta de crédito mas grande de Europa, desarrollando un sistema inteligente para la prevención de fraude. Experiencia como profesor de analytics y econometría en las universidades de Luxemburgo y de los Andes, respectivamente. Conferencista de analytics en SAS Analytics (Orlando, Las Vegas, Londres), SAS Global Forum (Orlando, San Francisco), IEEE International Conference on Data Mining (Vancouver, Bruselas), IEEE International Conference on Machine Learning and Applications (Miami) y European Conference on Data Analysis (Luxemburgo). Fundador de la comunidad Data Science Luxembourg y organizador del workshop IEEE Data Mining Case Studies (Dallas).

from Alejandro Correa Bahnsen, PhD

]]> 2393 22 https://cdn.slidesharecdn.com/ss_thumbnails/analyticscompitiendoenlaeradelainformacionc-140924114102-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Maximizing a churn campaign’s profitability with cost sensitive predictive analytics /slideshow/maximizing-a-churn-campaigns-profitability-with-cost-sensitive-predictive-analytics/35577390 maximizingachurncampaignsprofitabilitywithcost-sensitivepredictiveanalytics-140606131451-phpapp01
Presentation at SAS Analytics conference 2014 Predictive analytics has been applied to solve a wide range of real-world problems. Nevertheless, current state-of-the-art predictive analytics models are not well aligned with business needs since they don't include the real financial costs and benefits during the training and evaluation phases. Churn modeling does not yield the best results when it's measured by investment per subscriber on a loyalty campaign and the financial impact of failing to detect a churner versus wrongly predicting a non-churner. This presentation will show how using a cost-sensitive modeling approach leads to better results in terms of profitability and predictive power – and is applicable to many other business challenges. ]]>
Presentation at SAS Analytics conference 2014 Predictive analytics has been applied to solve a wide range of real-world problems. Nevertheless, current state-of-the-art predictive analytics models are not well aligned with business needs since they don't include the real financial costs and benefits during the training and evaluation phases. Churn modeling does not yield the best results when it's measured by investment per subscriber on a loyalty campaign and the financial impact of failing to detect a churner versus wrongly predicting a non-churner. This presentation will show how using a cost-sensitive modeling approach leads to better results in terms of profitability and predictive power – and is applicable to many other business challenges. ]]> Fri, 06 Jun 2014 13:14:51 GMT /slideshow/maximizing-a-churn-campaigns-profitability-with-cost-sensitive-predictive-analytics/35577390 albahnsen@slideshare.net(albahnsen) Maximizing a churn campaign’s profitability with cost sensitive predictive analytics albahnsen Presentation at SAS Analytics conference 2014 Predictive analytics has been applied to solve a wide range of real-world problems. Nevertheless, current state-of-the-art predictive analytics models are not well aligned with business needs since they don't include the real financial costs and benefits during the training and evaluation phases. Churn modeling does not yield the best results when it's measured by investment per subscriber on a loyalty campaign and the financial impact of failing to detect a churner versus wrongly predicting a non-churner. This presentation will show how using a cost-sensitive modeling approach leads to better results in terms of profitability and predictive power – and is applicable to many other business challenges. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/maximizingachurncampaignsprofitabilitywithcost-sensitivepredictiveanalytics-140606131451-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds" /> Presentation at SAS Analytics conference 2014 Predictive analytics has been applied to solve a wide range of real-world problems. Nevertheless, current state-of-the-art predictive analytics models are not well aligned with business needs since they don't include the real financial costs and benefits during the training and evaluation phases. Churn modeling does not yield the best results when it's measured by investment per subscriber on a loyalty campaign and the financial impact of failing to detect a churner versus wrongly predicting a non-churner. This presentation will show how using a cost-sensitive modeling approach leads to better results in terms of profitability and predictive power – and is applicable to many other business challenges.

Maximizing a churn campaign’s profitability with cost sensitive predictive analytics from Alejandro Correa Bahnsen, PhD

]]> 4611 4 https://cdn.slidesharecdn.com/ss_thumbnails/maximizingachurncampaignsprofitabilitywithcost-sensitivepredictiveanalytics-140606131451-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 Example-Dependent Cost-Sensitive Credit Card Fraud Detection /slideshow/2014-03-21-acorreabahnsen-edcs-fraud-detection/32628022 20140321acorreabahnsenedcsfrauddetection-140323060236-phpapp02
Credit card fraud is a growing problem that affects card holders around the world. Fraud detection has been an interesting topic in machine learning. Nevertheless, current state of the art credit card fraud detection algorithms miss to include the real costs of credit card fraud as a measure to evaluate algorithms. In this paper a new comparison measure that realistically represents the monetary gains and losses due to fraud detection is proposed. Moreover, using the proposed cost measure a cost sensitive method based on Bayes minimum risk is presented. This method is compared with state of the art algorithms and shows improvements up to 23% measured by cost. The results of this paper are based on real life transactional data provided by a large European card processing company. ]]>
Credit card fraud is a growing problem that affects card holders around the world. Fraud detection has been an interesting topic in machine learning. Nevertheless, current state of the art credit card fraud detection algorithms miss to include the real costs of credit card fraud as a measure to evaluate algorithms. In this paper a new comparison measure that realistically represents the monetary gains and losses due to fraud detection is proposed. Moreover, using the proposed cost measure a cost sensitive method based on Bayes minimum risk is presented. This method is compared with state of the art algorithms and shows improvements up to 23% measured by cost. The results of this paper are based on real life transactional data provided by a large European card processing company. ]]> Sun, 23 Mar 2014 06:02:35 GMT /slideshow/2014-03-21-acorreabahnsen-edcs-fraud-detection/32628022 albahnsen@slideshare.net(albahnsen) Example-Dependent Cost-Sensitive Credit Card Fraud Detection albahnsen Credit card fraud is a growing problem that affects card holders around the world. Fraud detection has been an interesting topic in machine learning. Nevertheless, current state of the art credit card fraud detection algorithms miss to include the real costs of credit card fraud as a measure to evaluate algorithms. In this paper a new comparison measure that realistically represents the monetary gains and losses due to fraud detection is proposed. Moreover, using the proposed cost measure a cost sensitive method based on Bayes minimum risk is presented. This method is compared with state of the art algorithms and shows improvements up to 23% measured by cost. The results of this paper are based on real life transactional data provided by a large European card processing company. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/20140321acorreabahnsenedcsfrauddetection-140323060236-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds" /> Credit card fraud is a growing problem that affects card holders around the world. Fraud detection has been an interesting topic in machine learning. Nevertheless, current state of the art credit card fraud detection algorithms miss to include the real costs of credit card fraud as a measure to evaluate algorithms. In this paper a new comparison measure that realistically represents the monetary gains and losses due to fraud detection is proposed. Moreover, using the proposed cost measure a cost sensitive method based on Bayes minimum risk is presented. This method is compared with state of the art algorithms and shows improvements up to 23% measured by cost. The results of this paper are based on real life transactional data provided by a large European card processing company.

Example-Dependent Cost-Sensitive Credit Card Fraud Detection from Alejandro Correa Bahnsen, PhD

]]> 5814 12 https://cdn.slidesharecdn.com/ss_thumbnails/20140321acorreabahnsenedcsfrauddetection-140323060236-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation White http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 2012 predictive clusters /slideshow/2012-predictive-clusters/23233712 predclustersslideshare-130620040211-phpapp02
Presentation at the SAS Global Forum 2012, Orlando, FL. Presenters: Alejandro Correa Bahnsen Andres Felipe Gonzalez Montoya]]>
Presentation at the SAS Global Forum 2012, Orlando, FL. Presenters: Alejandro Correa Bahnsen Andres Felipe Gonzalez Montoya]]> Thu, 20 Jun 2013 04:02:11 GMT /slideshow/2012-predictive-clusters/23233712 albahnsen@slideshare.net(albahnsen) 2012 predictive clusters albahnsen Presentation at the SAS Global Forum 2012, Orlando, FL. Presenters: Alejandro Correa Bahnsen Andres Felipe Gonzalez Montoya <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/predclustersslideshare-130620040211-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds" /> Presentation at the SAS Global Forum 2012, Orlando, FL. Presenters: Alejandro Correa Bahnsen Andres Felipe Gonzalez Montoya

2012 predictive clusters from Alejandro Correa Bahnsen, PhD

]]> 881 3 https://cdn.slidesharecdn.com/ss_thumbnails/predclustersslideshare-130620040211-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 2013 credit card fraud detection why theory dosent adjust to practice /slideshow/correa-bahnsen-alejandroanalytics2013slideshare/23233163 correabahnsenalejandroanalytics2013slideshare-130620034606-phpapp02
Presentation at the SAS Analytics Conference 2013, London, UK. Presenter: Alejandro Correa Bahnsen ]]>
Presentation at the SAS Analytics Conference 2013, London, UK. Presenter: Alejandro Correa Bahnsen ]]> Thu, 20 Jun 2013 03:46:05 GMT /slideshow/correa-bahnsen-alejandroanalytics2013slideshare/23233163 albahnsen@slideshare.net(albahnsen) 2013 credit card fraud detection why theory dosent adjust to practice albahnsen Presentation at the SAS Analytics Conference 2013, London, UK. Presenter: Alejandro Correa Bahnsen <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/correabahnsenalejandroanalytics2013slideshare-130620034606-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds" /> Presentation at the SAS Analytics Conference 2013, London, UK. Presenter: Alejandro Correa Bahnsen

2013 credit card fraud detection why theory dosent adjust to practice from Alejandro Correa Bahnsen, PhD

]]> 4309 8 https://cdn.slidesharecdn.com/ss_thumbnails/correabahnsenalejandroanalytics2013slideshare-130620034606-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 2011 advanced analytics through the credit cycle /slideshow/2011-advanced-analytics-through-the-credit-cycle/17993130 2011advancedanalyticsthroughthecreditcycle-130401071644-phpapp02
Presentation at the SAS Analytics Conference 2011, Orlando, FL. Presenters: Alejandro Correa Bahnsen Andres Felipe Gonzalez Montoya]]>
Presentation at the SAS Analytics Conference 2011, Orlando, FL. Presenters: Alejandro Correa Bahnsen Andres Felipe Gonzalez Montoya]]> Mon, 01 Apr 2013 07:16:44 GMT /slideshow/2011-advanced-analytics-through-the-credit-cycle/17993130 albahnsen@slideshare.net(albahnsen) 2011 advanced analytics through the credit cycle albahnsen Presentation at the SAS Analytics Conference 2011, Orlando, FL. Presenters: Alejandro Correa Bahnsen Andres Felipe Gonzalez Montoya <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/2011advancedanalyticsthroughthecreditcycle-130401071644-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds" /> Presentation at the SAS Analytics Conference 2011, Orlando, FL. Presenters: Alejandro Correa Bahnsen Andres Felipe Gonzalez Montoya

2011 advanced analytics through the credit cycle from Alejandro Correa Bahnsen, PhD

]]> 2788 4 https://cdn.slidesharecdn.com/ss_thumbnails/2011advancedanalyticsthroughthecreditcycle-130401071644-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation White http://activitystrea.ms/schema/1.0/post

http://activitystrea.ms/schema/1.0/posted

0 https://cdn.slidesharecdn.com/profile-photo-albahnsen-48x48.jpg?cb=1549405264 Passionate about Machine Learning and Data Science. I hold a PhD in Machine Learning and Pattern Recognition from Luxembourg University. Many years of experience in the use and development of predictive models to real-world problems such as: Algorithmic trading, fraud detection, HR analytics, credit scoring, collections, churn and direct marketing. I have written and publish many academic and industrial papers in best per-review publications. Lastly, I also have experience as instructor of econometrics, financial risk management and Machine Learning. I Enjoy giving talks on successful applications of big data science on different applications. albahnsen.com https://cdn.slidesharecdn.com/ss_thumbnails/201812blackhatdeepphish-181204222605-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/black-hat-deephish/124982274 black hat deephish https://cdn.slidesharecdn.com/ss_thumbnails/201805deepphishapwgecrime-180523182835-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/deepphish-simulating-malicious-ai/98311818 DeepPhish: Simulating ... https://cdn.slidesharecdn.com/ss_thumbnails/enpresentationcds-180503020944-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/ai-vs-ai-can-predictive-models-stop-the-tide-of-hacker-ai/95771190 AI vs. AI: Can Predict...