I build statistical models to predict and explain task outcomes, particularly tasks that involve subjective opinions. The models that I build concern three topics: (1) how to aggregate conflicting opinions, (2) decide if more opinions are necessary about the same or different items, and (3) who should we ask for further opinions.
While pursuing these research problems, I rely on a set of models and technology:
- Theory: Decision Theory, Multi-Armed Bandits, Voting, Utility theory, and Prospect theory.
- Probabilistic Models: Causal Inference, Bayesian Inference, Markov models, and Bayesian Optimization.
- Empirical research: Quasi-experimental and Observational studies, Survival analysis, Analyzes of covariance, Time series, and Sensitivity analysis.
- Application domains: Crowdsourcing, Software Debugging, Self-Adaptive Systems, and, more recently, Comments from Q&A sites or News sites
- Tools/Languages: R, Python, Java, Scala, Spark