Model Distillation for Revenue Optimization: Interpretable Personalized Pricing

TL;DR

This paper introduces a tree-based distillation method that simplifies complex machine learning pricing models into interpretable policies, maximizing revenue while ensuring transparency and fairness.

Contribution

It develops a prescriptive tree algorithm that distills complex models into simple, interpretable pricing policies optimized for revenue.

Findings

The method achieves near-optimal revenue in synthetic datasets.

It demonstrates effective application on real-world pricing data.

The approach maintains interpretability without sacrificing significant revenue.

Abstract

Data-driven pricing strategies are becoming increasingly common, where customers are offered a personalized price based on features that are predictive of their valuation of a product. It is desirable for this pricing policy to be simple and interpretable, so it can be verified, checked for fairness, and easily implemented. However, efforts to incorporate machine learning into a pricing framework often lead to complex pricing policies which are not interpretable, resulting in slow adoption in practice. We present a customized, prescriptive tree-based algorithm that distills knowledge from a complex black-box machine learning algorithm, segments customers with similar valuations and prescribes prices in such a way that maximizes revenue while maintaining interpretability. We quantify the regret of a resulting policy and demonstrate its efficacy in applications with both synthetic and real-world datasets.