際際滷

際際滷Share a Scribd company logo

Topic0 e

Download as PPTX, PDF
D
David Keck

This document discusses model thinking and different types of models. It states that models are approximations of reality and while all models are wrong, some can be useful. It discusses using models to think more clearly, understand data, and make decisions. Model thinking involves using data to create data models to gain insights and make predictions. A single model can apply to multiple problems and multiple models can be used to study one problem. The document lists different types of models used across individual decision making, organizations, biological systems, physical systems, and more.

1 of 11
Download to read offline
Model Thinking including Data Science Simulation Optimization 1
Topic0 e
Models "The most that can be expected from any model is that it can supply a useful approximation to reality. All models are wrong; some models are useful." George Box
Topic0 e
Theories Hypotheses Laws Models Mathematical Physical Virtual Data
Model Thinking ? 6 Think and reason more clearly Intelligent citizen the world Understand and use data Decide, strategize, and design Scott E. Page
traditional science data theoretical models insight & prediction Models data science data data models insight & prediction
8 The Many Model Thinker One model many applications Many models one application Scott E. Page
Model Thinking 9 The wisdom or madness of crowds is due to independence and diversity of models for solving a problem Scott E. Page A model can apply to multiple disciplines This leverages insight across disciplines Model Proble m Proble m Proble m Multiple models can be used to study a problem Model 2 Proble m Model 3 Model 1
Range of Models Individu al Decision Theory Two People Game Theory Organizatio n Agent models for culture and cooperation emergence Societies Economie s Cities Agent based, game theory models Biologic al Systems Agent and evolutionar y models Physical Systems Random, deterministic , probabilistic
11 Social models: Segregation and peer effects Aggregation of numbers, preferences, ? Randomness, random processes ,random walks Cellular automata for social and biological systems Convergence and optimality Probability Social models: Segregation and peer effects Decision making: rational actor, behavioral, rule based Numerical and categorical data Linear and nonlinear Regression and classification models Emergence, tipping points, contagion, SIS, diffusion, percolation Growth models, exponential growth, Solow growth model Problem solving, perspectives and innovation Heuristics Problem solving and teams Markov processes Decision models Decision models Coordination and culture Emergence of culture Path dependence Chaos Increasing returns to scale Network, structure, logic, formation Skills and luck Game theory and Colonel Blotto game Decision models Competition Cooperation and collective action Prediction and diversity The Many Model Thinker

More Related Content

Topic0 e

  • 3. Models "The most that can be expected from any model is that it can supply a useful approximation to reality. All models are wrong; some models are useful." George Box
  • 6. Model Thinking ? 6 Think and reason more clearly Intelligent citizen the world Understand and use data Decide, strategize, and design Scott E. Page
  • 7. traditional science data theoretical models insight & prediction Models data science data data models insight & prediction
  • 8. 8 The Many Model Thinker One model many applications Many models one application Scott E. Page
  • 9. Model Thinking 9 The wisdom or madness of crowds is due to independence and diversity of models for solving a problem Scott E. Page A model can apply to multiple disciplines This leverages insight across disciplines Model Proble m Proble m Proble m Multiple models can be used to study a problem Model 2 Proble m Model 3 Model 1
  • 10. Range of Models Individu al Decision Theory Two People Game Theory Organizatio n Agent models for culture and cooperation emergence Societies Economie s Cities Agent based, game theory models Biologic al Systems Agent and evolutionar y models Physical Systems Random, deterministic , probabilistic
  • 11. 11 Social models: Segregation and peer effects Aggregation of numbers, preferences, ? Randomness, random processes ,random walks Cellular automata for social and biological systems Convergence and optimality Probability Social models: Segregation and peer effects Decision making: rational actor, behavioral, rule based Numerical and categorical data Linear and nonlinear Regression and classification models Emergence, tipping points, contagion, SIS, diffusion, percolation Growth models, exponential growth, Solow growth model Problem solving, perspectives and innovation Heuristics Problem solving and teams Markov processes Decision models Decision models Coordination and culture Emergence of culture Path dependence Chaos Increasing returns to scale Network, structure, logic, formation Skills and luck Game theory and Colonel Blotto game Decision models Competition Cooperation and collective action Prediction and diversity The Many Model Thinker

Editor's Notes

  • #4: Back to Scott Page who again coined the term and has huge influence worldwide now from his MOOC that weve all taken. I think we all agree it was most influential MOOC From the introductory lecture of his course Brain is a real time 3-D modeler, takes in sensory perception and makes predictions about movement Models for prediction and insight ----------- Wisdom of crowds, value of diversity How do numbers aggregate ? Add, multiply, set logic, How do infinitesimal volumes and properties aggregate? Volume integral How do random numbers aggregate? Law of large numbers, variance, How do independent random errors add? Central limit theorem (normal distribution) How do independent random errors multiply? Central limit theorem (log normal distribution) How do preferences aggregate ? How do rules aggregate ? How do individual behaviors aggregate to the organization level ? Rational ? Behavioral ? Rule based ?
  • #8: 1841 2005
  • #11: From the intro lecture of his MOOC. Apply one model to many problems. Thats not the way universities work. Simplex story Apply multiple models to one problem. Scott proved that better results insight and predictions result