Predicting Choice with Set-Dependent Aggregation

TL;DR

This paper introduces a versatile, scalable, and theoretically grounded choice prediction framework that captures set-related invariances in human decision-making, outperforming existing models on large datasets.

Contribution

We propose a new learning framework for choice prediction that models set-dependent invariances, enabling expressive, scalable, and theoretically sound predictions.

Findings

Accurately predicts choices across large datasets

Captures set-related invariances in human decision-making

Enjoys favorable sample complexity guarantees

Abstract

Providing users with alternatives to choose from is an essential component in many online platforms, making the accurate prediction of choice vital to their success. A renewed interest in learning choice models has led to significant progress in modeling power, but most current methods are either limited in the types of choice behavior they capture, cannot be applied to large-scale data, or both. Here we propose a learning framework for predicting choice that is accurate, versatile, theoretically grounded, and scales well. Our key modeling point is that to account for how humans choose, predictive models must capture certain set-related invariances. Building on recent results in economics, we derive a class of models that can express any behavioral choice pattern, enjoy favorable sample complexity guarantees, and can be efficiently trained end-to-end. Experiments on three large choice datasets demonstrate the utility of our approach.