# 🐅 Computational Bayesian Analysis - 계산적 베이지안 분석 > *"What are the right things to do?"* ## Field Overview CompBayes field는 **베이지안 추론**, **불확실성 정량화**, **인과추론**을 통해 전략적 의사결정을 연구합니다. 한산대첩의 학익진처럼, 데이터 기반 전략과 증거 기반 선택이 핵심입니다. ### Core Questions - How do we quantify uncertainty in decisions? - How can we learn from limited data? - What is the causal effect of interventions? - How do we update beliefs with new evidence? ### Key Journals - **Journal of the American Statistical Association** - **Bayesian Analysis** - **Statistical Science** - **Journal of Computational and Graphical Statistics** ## Research Themes ### 1. 🎲 Bayesian Inference **Learning and updating beliefs** Related papers: - [[📜Gelman21_bayesholes]] - Avoiding Bayesian pitfalls - [[📜gelman06_boxer🥊]] - Boxing metaphor for Bayesian thinking - [[📜mcelreath25_rethinking]] - Statistical rethinking - [[Space/Sources/Papers/📜Richters21_incredible_utility]] - Incredible utility of Bayes **Core idea**: Coherent belief updating --- ### 2. 🔧 Computational Methods **Scaling Bayesian inference** Related papers: - [[Space/Sources/Papers/📜Stan manual on auto param tuning in warmup]] - MCMC tuning - [[📜Burkner23_Bayesian model taxonomy]] - Model taxonomy - [[Space/Sources/Papers/📜Margossian24_nested_rhat]] - Convergence diagnostics - [[📜👾_mansinghka25_automate(formalization, programming)]] - Automated inference **Core idea**: Efficient computation enables complex models --- ### 3. 📊 Model Building **Structuring uncertainty** Related papers: - [[Space/Sources/Papers/📜Cronin21_synthesize(theory)]] - Theory synthesis - [[📜🐢_meehl90_appraise(theory, amendments)]] - Theory appraisal - [[📜👾_meehl67_test(theory, method)]] - Theory testing **Core idea**: Models as scientific tools --- ### 4. 🎯 Decision Theory **Optimal choices under uncertainty** Related papers: - [[Space/Sources/Papers/📜Walters23_invest_beh_epis_alea]] - Epistemic and aleatory uncertainty - [[📜Hullman_How far can exchangeability get us toward agreeing on individual probability?]] - Exchangeability - [[Space/Sources/Papers/📜Phillips19_How We Know What Not To Think]] - Negative knowledge **Core idea**: Decisions require probability --- ### 5. 🔬 Experimental Design **Learning efficiently** Related papers: - [[📜🐅_kerr14_systematize(experimentation, entrepreneurship)]] - Systematic experimentation - [[📜🐅_camuffo25_experiment(beliefs, entrepreneurs)]] - Belief experiments - [[📜👾_camuffo19_structure(experiments, learning)]] - Structured learning **Core idea**: Experiment to learn --- ### 6. ⚖️ Moral Hazard and Incentives **Aligning interests** Related papers: - [[📜🐅_bolton24_moral_hazard]] - Moral hazard theory - [[📜Bolton24]] - Bolton's latest work **Core idea**: Incentives shape behavior --- ### 7. 📈 Portfolio Optimization **Balancing exploration and exploitation** Related papers: - [[📜🐅_loch02_optimize(portfolio, selection)]] - Portfolio selection - [[📜🐅_kavadias03_sequence(projects, optimization)]] - Project sequencing - [[📜🐅_dada07_diversify(sourcing, suppliers)]] - Supplier diversification **Core idea**: Optimize the portfolio, not individual projects --- ### 8. 🎭 Collective Behavior **Modeling social dynamics** Related papers: - [[📜🐅_granovetter78_model(collective-behavior, thresholds)]] - Threshold models **Core idea**: Individual to collective --- ## Methodology Focus ### Bayesian Methods - MCMC (HMC, NUTS) - Variational inference - Hierarchical models ### Causal Inference - DAGs - Potential outcomes - Instrumental variables ### Model Comparison - Cross-validation - Information criteria - Bayes factors ## Writing Guide for 🐅 CompBayes Papers ### Structure 1. **Question**: What decision/inference problem? 2. **Model**: What probabilistic model captures uncertainty? 3. **Inference**: How do we compute the posterior? 4. **Validation**: How do we check the model? 5. **Decision**: What action follows from inference? ### Language - Use **probabilistic constructs**: prior, likelihood, posterior, uncertainty - Emphasize **quantification**: "how much" over "whether" - Bridge **inference and action**: belief → decision ### Common Mistakes ❌ Model without checking ❌ Point estimates without uncertainty ❌ Ignoring prior sensitivity ❌ No connection to decision ## Related Fields - [[4👾Cognition]] - Bayesian cognition - [[1🐢Innovation]] - Uncertainty in innovation - [[3🐙Operations]] - Probabilistic operations ## Key Papers by Theme ![[papers.base#papers-cba]] --- > *"智혜로운 머리 - 데이터 기반 의사결정으로 적을 포위하다"*