際際滷

際際滷Share a Scribd company logo

Vinay bamane

Download as PPT, PDF
Vinay Bamane
Vinay Bamane

This document outlines a proposed system for securely mining association rules from horizontally distributed databases while preserving data privacy. It discusses how existing algorithms have limitations regarding communication overhead and privacy as the number of database sites increases. The proposed system aims to monitor multiple sites to mine association rules securely without sites revealing private database contents. It designs modules for user transactions, admin analytics, privacy-preserving data mining, and distributed computation. The system would optimize pattern detection in distributed databases and generate association rules securely.

1 of 31
Download to read offline
SECURE MINING OF RULES IN HORIZONTLLY DISTRIBUTED DATABASE BY Mr. BAMANE VINAY VISHNU UNDER THE GUIDANCE OF Prof. VIPUL BAG
Outline Introduction Literature Survey Problem Statement Necessity of Proposed System Proposed System Detail Design Implementation Conclusion and Future Scope 2
Introduction Secure mining of association rules in Horizontally Distributed databases Fast Distributed Mining (FDM) algorithm unsecured distributed version of the Apriori algorithm. Two novel secure multi-party algorithms 1. The union of private subsets that each of the interacting players hold. 2. The inclusion of an element held by one player in a subset held by another. Simpler and is significantly more efficient in terms of communication rounds, communication cost and computational costs. 3
Literature Survey Secure Mining of Association Rules in Horizontally Distribute Databases. Privacy-Preserving Distributed Mining of Association Rules on Horizontally Partitioned Data 4
5 1. Databases that share the same schema but hold information on different entities. 2. Distributed association rule mining techniques can discover association rules among multiple sites. 3. Each site faced problem of communication and computation overhead. 4. When computing global frequent item sets as number of sites increases overhead also increases. Secure Mining of Association Rules in Horizontally Distribute Databases
Privacy-Preserving Distributed Mining of Association Rules on Horizontally Partitioned Data Encryption technology to minimize the sharing of information. Privacy concerns may prevent the parties from directly sharing the data. Secure multi party computation protocol, holds frequent item sets globally. 6
Problem Statement It is a difficult task to securely mining association rules from horizontal distributed database. In case need of efficient algorithms for mining frequent item sets are crucial for mining association rules. The challenge found in frequent item sets mining is a large number of result patterns are generated. 7
Necessity of Proposed System There is a need to optimize the process of finding patterns. Which should detect the important patterns and generate association rules in distributed database. There is also need of a preserving privacy of transaction database Because data processed among databases in some business environments. Distributed among several sites, but none of the sites is allowed to expose its database to another. 8
Proposed System The application is designed in such a way which secure mining of association rules in distributed database and also protecting the data records. The application monitors during the time number of site should be able to learn contents of transactions at any other site and also maintain security and data efficiency. 9
Modules Design User Module. Admin Module. Privacy Preserving Data Mining. Distributed Computation. 10
User Module Privacy preserving data mining has considered two related settings The data owner and the data miner are two different entities. Data is distributed among several parties who aim to jointly perform data mining. In the first setting, the goal is to protect the data records from the data miner. General trends in the data, without revealing original record information. In the second setting the goal is to perform data mining While protecting the data records of each of the data owners from the other data owners. 11
User Module 12 Home Fund Transfer Add Third Party Transitions
User Module 13 Home Fund Transfer Add Third Party Transitions
User Module 14 Home Fund Transfer Add Third Party Transitions
User Module 15 Home Fund Transfer Add Third Party Transitions
Admin Module View user details The item set based on the user processing details using association role with Apriori algorithm. 16
Admin Module 17 Home Register Amount Verify Third party Transitions
Admin Module 18 Home Register Amount Verify Third party Transitions
Admin SBI 19 Home Register Amount Verify Third party Transitions
Admin IDBI 20 Home Register Amount Verify Third party Transitions
Admin HDFC 21 Home Register Amount Verify Third party Transitions
Privacy Preserving Data Mining The data owner and the data miner are two different entities To protect the data records from the data miner. The data owner aims at anonym zing the data prior to its release. Perform data mining while protecting the data records of each of the data owners from the other data owners. 22
Privacy Preserving Data Mining 23
Distributed Computation 24 The two implementations with respect to three measures: 1) Total computation time of the complete protocols . The Apriori computation time, and the time to identify the globally s-frequent item sets. 2) Total computation time of the unification protocols all players. 3) Total message size. N the number of transactions in the unified database.
Detail Design Activity Diagram Sequence Diagram 25
Activity Diagram 26
Sequence Diagram 27 Fig. Sequence Diagram Admin
Sequence Diagram 28 Fig. Sequence Diagram User
Conclusion and Future Work As the algorithm generate strong association rules from different data sets spread over various sites and also preserving privacy of data. This application can be used for Business Analysis in Banking Sector. The research can be also extended association rules from different area like stock market, college, school and medical datasets. 29
Publications Secure Extraction Of Association Rules In Parallel Disseminated Data Paper ID IJCSN-2016-5-5-111 30
31 THANK YOUTHANK YOU

More Related Content

Vinay bamane

  • 1. SECURE MINING OF RULES IN HORIZONTLLY DISTRIBUTED DATABASE BY Mr. BAMANE VINAY VISHNU UNDER THE GUIDANCE OF Prof. VIPUL BAG
  • 2. Outline Introduction Literature Survey Problem Statement Necessity of Proposed System Proposed System Detail Design Implementation Conclusion and Future Scope 2
  • 3. Introduction Secure mining of association rules in Horizontally Distributed databases Fast Distributed Mining (FDM) algorithm unsecured distributed version of the Apriori algorithm. Two novel secure multi-party algorithms 1. The union of private subsets that each of the interacting players hold. 2. The inclusion of an element held by one player in a subset held by another. Simpler and is significantly more efficient in terms of communication rounds, communication cost and computational costs. 3
  • 4. Literature Survey Secure Mining of Association Rules in Horizontally Distribute Databases. Privacy-Preserving Distributed Mining of Association Rules on Horizontally Partitioned Data 4
  • 5. 5 1. Databases that share the same schema but hold information on different entities. 2. Distributed association rule mining techniques can discover association rules among multiple sites. 3. Each site faced problem of communication and computation overhead. 4. When computing global frequent item sets as number of sites increases overhead also increases. Secure Mining of Association Rules in Horizontally Distribute Databases
  • 6. Privacy-Preserving Distributed Mining of Association Rules on Horizontally Partitioned Data Encryption technology to minimize the sharing of information. Privacy concerns may prevent the parties from directly sharing the data. Secure multi party computation protocol, holds frequent item sets globally. 6
  • 7. Problem Statement It is a difficult task to securely mining association rules from horizontal distributed database. In case need of efficient algorithms for mining frequent item sets are crucial for mining association rules. The challenge found in frequent item sets mining is a large number of result patterns are generated. 7
  • 8. Necessity of Proposed System There is a need to optimize the process of finding patterns. Which should detect the important patterns and generate association rules in distributed database. There is also need of a preserving privacy of transaction database Because data processed among databases in some business environments. Distributed among several sites, but none of the sites is allowed to expose its database to another. 8
  • 9. Proposed System The application is designed in such a way which secure mining of association rules in distributed database and also protecting the data records. The application monitors during the time number of site should be able to learn contents of transactions at any other site and also maintain security and data efficiency. 9
  • 10. Modules Design User Module. Admin Module. Privacy Preserving Data Mining. Distributed Computation. 10
  • 11. User Module Privacy preserving data mining has considered two related settings The data owner and the data miner are two different entities. Data is distributed among several parties who aim to jointly perform data mining. In the first setting, the goal is to protect the data records from the data miner. General trends in the data, without revealing original record information. In the second setting the goal is to perform data mining While protecting the data records of each of the data owners from the other data owners. 11
  • 12. User Module 12 Home Fund Transfer Add Third Party Transitions
  • 13. User Module 13 Home Fund Transfer Add Third Party Transitions
  • 14. User Module 14 Home Fund Transfer Add Third Party Transitions
  • 15. User Module 15 Home Fund Transfer Add Third Party Transitions
  • 16. Admin Module View user details The item set based on the user processing details using association role with Apriori algorithm. 16
  • 17. Admin Module 17 Home Register Amount Verify Third party Transitions
  • 18. Admin Module 18 Home Register Amount Verify Third party Transitions
  • 19. Admin SBI 19 Home Register Amount Verify Third party Transitions
  • 20. Admin IDBI 20 Home Register Amount Verify Third party Transitions
  • 21. Admin HDFC 21 Home Register Amount Verify Third party Transitions
  • 22. Privacy Preserving Data Mining The data owner and the data miner are two different entities To protect the data records from the data miner. The data owner aims at anonym zing the data prior to its release. Perform data mining while protecting the data records of each of the data owners from the other data owners. 22
  • 24. Distributed Computation 24 The two implementations with respect to three measures: 1) Total computation time of the complete protocols . The Apriori computation time, and the time to identify the globally s-frequent item sets. 2) Total computation time of the unification protocols all players. 3) Total message size. N the number of transactions in the unified database.
  • 25. Detail Design Activity Diagram Sequence Diagram 25
  • 29. Conclusion and Future Work As the algorithm generate strong association rules from different data sets spread over various sites and also preserving privacy of data. This application can be used for Business Analysis in Banking Sector. The research can be also extended association rules from different area like stock market, college, school and medical datasets. 29
  • 30. Publications Secure Extraction Of Association Rules In Parallel Disseminated Data Paper ID IJCSN-2016-5-5-111 30