Advanced Discrete Choice Modeling: Selected Topics

This module introduces Multivariate Extreme Value models, which address some limitations of the logit model. It explores how these models can better capture behavioral reality and avoid erroneous forecasts, using examples like the "red bus-blue bus" paradox.

This module focuses on the critical impact of sampling procedures on choice data collection and model estimation. It introduces classical sampling procedures and analyzes their implications for model estimation in detail.

This module introduces a new family of models based on "mixtures" to address limitations of the logit model. It covers the definition of mixtures, calculation methods, and important modeling assumptions they can accommodate.

This module explores hybrid choice models that can capture subjective dimensions of the choice process using latent variables. It discusses how these models can account for apparent irrationality in decision-making that contradicts traditional utility theory.

This module focuses on incorporating the time dimension into choice models using panel data. It covers the collection of longitudinal data and the econometric implications of analyzing choices that evolve over time, capturing learning and habit formation.

This module compares and contrasts machine learning algorithms with discrete choice models. It discusses the similarities, differences, and potential limitations of machine learning in the context of analyzing choice data.