hanshengjiang / Reference_Effects

Software package for intertemporal pricing optimization under reference effects and consumer heterogeneity estimation. Please see REAMDE.md for more details.

IPRE: Intertemporal Pricing under Reference Effects

Overview

This repository contains numerical implementation for the paper Intertemporal Pricing via Nonparametric Estimation: Integrating Reference Effects and Consumer Heterogeneity. Along with the reproduction code, this repository code also contains general functions for implementing nonparametric estimation of consumer heterogeneity.

Folders

• scripts/: Python and R files
• illustrations/: two .png pictures illustrating logit demand, and how it depends on reference price and price
• simulation_results: estimated coefficients in simulation stored in .csv files
• pricing_output/: .png pictures showing pricing policy and cumulative revenue, for both simulation and MSOM (real) data
• MSOM_data_cleaned/: extracted feature data in .csv files, ready as inputs of the estimation algorithm
• MSOM_data_estimated/: estimated coefficients of MSOM data stored in .csv files
• MSOM_data_optimized/: revenue comparison for real data study
• MSOM_Data/: MSOM-JD.com dataset

Scripts and Modules

Each Python script in scripts/ starting with run_ is used for one run of a certain numerical experiment, while each python script ending with _py defines some functions to be imported by other files.

Based on the purposes of all the scripts, we further categorize them into the following modules.

• Data preprocessing and feature extraction
  • run_data_cleaning.py, py_MSOM_cleaning.py, run_extract_features.py, run_freq_user.py, run_freq_estimate.py,
• Heterogeneous Reference Effects Estimation
  • Functions: py_estimation.py, cross_validation.py, mmnl_simualtion.py
  • For simulated data: run_mmnl_estimation_simulation.py
  • For MSOM data: run_mmnl_estimation.py, run_mmnl_estimation_compare.py
• Pricing Optimization
  • Functions: optimal_pricing_policy_exp_update.py
  • For simulated data: run_pricing_optimization.py
  • For MSOM data: run_mmnl_pricing_optimization.py, run_mmnl_revenue_compare.py

Real Data and Access

The MSOM-JD.com dataset can be downloaded from this link given membership access, and a general introduction to the dataset is available in this paper. To be compatible with the codes, the uncompressed .csv data files should be stored in the folder ./MSOM_Data/.