- models in the age of data

- why we need many models

- wisdom hierarchy

- book summary/outline

- why we need many models

- wisdom hierarchy

- book summary/outline

- model uses (7)

`REDCAPE (reason, explain, design, communicate, act, predict & explore)`

- many models as independent lies

- condorcet jury theorem

- diversity prediction theorem

- categorization models

- one big model & granularity

- R^2: pct of variance

- model error decomposition theorem

- one-to-many

- one-to-many: higher powers (X^N)

- bagging

- condorcet jury theorem

- diversity prediction theorem

- categorization models

- one big model & granularity

- R^2: pct of variance

- model error decomposition theorem

- one-to-many

- one-to-many: higher powers (X^N)

- bagging

- the challenge

- rational-actor models

- rational-actor consumption model

- arguments for rational choice

- psych biases

- prospect theory example

- rule-based models

- el farol model: adaptive rules

- cognitive closure, a big question, & many models

- the lucas critique

- rational-actor models

- rational-actor consumption model

- arguments for rational choice

- psych biases

- prospect theory example

- rule-based models

- el farol model: adaptive rules

- cognitive closure, a big question, & many models

- the lucas critique

- structure

- central limit theorem

- square root rules

- testing significance

- six sigma

- lognormal distributions (multiplying shocks)

- summary

- central limit theorem

- square root rules

- testing significance

- six sigma

- lognormal distributions (multiplying shocks)

- summary

- structure

- distributions

- power law

- zipf's law

- models & power laws

- preferential attachments

- forest fire model (self-organized criticality)

- implications

- equity

- catastrophies

- volatility

- a long-tail world

- search & opportunity

- distributions

- power law

- zipf's law

- models & power laws

- preferential attachments

- forest fire model (self-organized criticality)

- implications

- equity

- catastrophies

- volatility

- a long-tail world

- search & opportunity

- definition

- sign, significance, magnitude

- correlation vs causation

- multivariable LMs

- success equation

- multivariable regression

- big coefficients & new realities

- binary classifications

- linear classification

- nonlinear classification

- forests of decision trees

- sign, significance, magnitude

- correlation vs causation

- multivariable LMs

- success equation

- multivariable regression

- big coefficients & new realities

- binary classifications

- linear classification

- nonlinear classification

- forests of decision trees

- convexity

- exponential growth

- rule of 72

- half-life

- concave functions

- economic growth

- cobb-douglas model

- simple growth

- solow (simplified) growth

- why nations succeed & fail

- japanese chinese economic dominance

- it's a nonlinear world

- exponential growth

- rule of 72

- half-life

- concave functions

- economic growth

- cobb-douglas model

- simple growth

- solow (simplified) growth

- why nations succeed & fail

- japanese chinese economic dominance

- it's a nonlinear world

- cooperative games

- shapley values (SVs)

- axiomatic basis for SVs

- SVs & alternate uses test

- shapley-shubik index

- summary

- shapley values (SVs)

- axiomatic basis for SVs

- SVs & alternate uses test

- shapley-shubik index

- summary

- structure

- statistics (degree, length, betweenness, clustering)

- common structures

- monte carlo method - random nets

- network formation: logic

- why networks matter

- six degrees of separation

- robustness

- summary

- statistics (degree, length, betweenness, clustering)

- common structures

- monte carlo method - random nets

- network formation: logic

- why networks matter

- six degrees of separation

- robustness

- summary

- broadcast model

- broadcast model : data fitting

- diffusion model

- bass model

- SIR model

- R0 (basic reproduction number)

- R0, superspreaders & degree squaring

- one-to-many

- broadcast model : data fitting

- diffusion model

- bass model

- SIR model

- R0 (basic reproduction number)

- R0, superspreaders & degree squaring

- one-to-many

- information entropy

- axiomatic foundations

- using entropy to distinguish outcome classes

- max entropy & distributional assumptions

- max entropy distributions
- positive & normative implications

- axiomatic foundations

- using entropy to distinguish outcome classes

- max entropy & distributional assumptions

- max entropy distributions

`uniform, exponential, normal`

- bernoulli urn model

- simple RW

- using RWs to estimate network size

- RWs & efficient markets

- simple RW

- using RWs to estimate network size

- RWs & efficient markets

- polya process

- balancing process

- path dependence or tipping point

- further applications

- value-at-risk & volatility

- balancing process

- path dependence or tipping point

- further applications

- value-at-risk & volatility

- local majority model

- pure coordination game

- paradox of coordination

- game of life

- summary

- pure coordination game

- paradox of coordination

- game of life

- summary

- lyapunov functions

- race to the bottom game

- equilibrium in local majority model

- self-organization: new york & disney world

- pure exchange commodities

- models without lyapunov functions

- summary

- race to the bottom game

- equilibrium in local majority model

- self-organization: new york & disney world

- pure exchange commodities

- models without lyapunov functions

- summary

- two examples

- perron-frobenius theorem

- sales-durability paradox

- markov: one-to-many

- summary

- perron-frobenius theorem

- sales-durability paradox

- markov: one-to-many

- summary

- system dynamics model components

- predator-prey model

- lotka-volterra model

- using SDMs to guide action

- WORLD3 model

- summary

`sources, sinks, stocks, flows`

- lotka-volterra model

- using SDMs to guide action

- WORLD3 model

- summary

- granovetter's riot model

- riot model

- market creation & double riots

- two models of segregation

- schelling's party model

- threshold models | negative feedbacks

- ping-pong model

- summary: model granularity

- algorithmic riots

- riot model

- market creation & double riots

- two models of segregation

- schelling's party model

- threshold models | negative feedbacks

- ping-pong model

- summary: model granularity

- algorithmic riots

- spatial competition model

- increasing #attributes

- downsian model of spatial competition

- status quo effects, agenda control, veto players

- hedonic competition

- hybrid model of product competition

- summary

- increasing #attributes

- downsian model of spatial competition

- status quo effects, agenda control, veto players

- hedonic competition

- hybrid model of product competition

- summary

- normal-form zero-sum games

- sequential games

- continuous action games

- effort game

- summary

- sequential games

- continuous action games

- effort game

- summary

- prisoner's dilemma

- cooperation thru repetition & reputation

- connectedness & reputation

- cooperation among rule-playing behaviors

- cooperative action model

- clustering bootstraps cooperation

- group selection

- summary

- cooperation thru repetition & reputation

- connectedness & reputation

- cooperation among rule-playing behaviors

- cooperative action model

- clustering bootstraps cooperation

- group selection

- summary

- intro

- public goods

- altruists

- congestion

- multiple congestible goods

- renewable resource extraction

- solved & unsolved CA problems

- public goods

- altruists

- congestion

- multiple congestible goods

- renewable resource extraction

- solved & unsolved CA problems

- mount-reiter diagram

- pareto efficiency

- majority rule & kingmaker mechanism

- three auctions

- revenue equivalence

- public project decision mechanisms

- majority-vote equal sharing

- pivot mechanism

- summary

`environment, outcomes, actions, behavioral rules, outcome functions, social choice correspondence`

- majority rule & kingmaker mechanism

- three auctions

`ascending-bid, second-price, first-price`

- public project decision mechanisms

- majority-vote equal sharing

- pivot mechanism

- summary

- discrete signals

- continuous signals

- continuous signals: separation

- uses & values

- summary

- continuous signals

- continuous signals: separation

- uses & values

- summary

- individual learning: reinforcement learning (RL)

- social learning: replicator dynamics

- learning in games

- generous | spiteful game

- spiteful man | magic lamp

- combining models

- culture trumps strategy?

- social learning: replicator dynamics

- learning in games

- generous | spiteful game

- spiteful man | magic lamp

- combining models

- culture trumps strategy?

- bernoulli bandit problems

- bernoulli bandit problems (multiarmed)

- gittins index

- summary

- bernoulli bandit problems (multiarmed)

- gittins index

- summary

- fitness landscape

- rugged landscapes

- NK model

- ruggedness & dancing landscapes

- do we patent knowledge?

- rugged landscapes

- NK model

- ruggedness & dancing landscapes

- do we patent knowledge?

- many models: opioid epidemic

- a model of opioid approval

- transition-to-addiction model

- paths to heroin addiction

- many models: covid pandemic

- fatality rate model

- IHME model

- flattening|unflattening the curve

- SIR with latency & severity (eisenberg)

- imperial college microsimulation model

- many models of inequality

- technology & human capital model

- positive feedbacks to talent

- ceo political capture

- rent-from-capital model (piketty)

- assortive mating

- intergenerational income (wealth) dynamics

- persistent inequality (durlauf)

- into the world

- a model of opioid approval

- transition-to-addiction model

- paths to heroin addiction

- many models: covid pandemic

- fatality rate model

- IHME model

- flattening|unflattening the curve

- SIR with latency & severity (eisenberg)

- imperial college microsimulation model

- many models of inequality

- technology & human capital model

- positive feedbacks to talent

- ceo political capture

- rent-from-capital model (piketty)

- assortive mating

- intergenerational income (wealth) dynamics

- persistent inequality (durlauf)

- into the world

notes