📅 2024-09-05 — Session: Nested Logit Model Refinement for Express Bus Choice
🕒 22:20–23:05
🏷️ Labels: Nested Logit Model, Probability, Transportation, Data Analysis, Model Refinement
📂 Project: Dev
⭐ Priority: MEDIUM
Session Goal
The goal of this session was to refine the nested logit model used for calculating the probability of a traveler choosing between Express and Local bus options, specifically focusing on the Red Express and Blue Express buses.
Key Activities
- Detailed explanation and breakdown of the nested logit model structure and calculations, including inclusive values and probabilities.
- Revised the model structure to treat Blue and Red as separate nests.
- Adjusted the correlation parameter and analyzed its impact on the probability of choosing an Express bus.
- Recalculated logsum values and assessed the scaling factor’s effect on model results.
- Identified discrepancies in the probability outcomes compared to expected values, suggesting further refinements.
Achievements
- Gained insights into the structural setup of the nested logit model and the application of correlation parameters.
- Highlighted the need for further adjustments to align model outputs with expected probabilities.
Pending Tasks
- Further refine the scaling and handling of utilities in the nested logit model.
- Reassess the correlation parameter application to correct probability discrepancies.