Causal Simulations for Uplift Modeling

TL;DR

This paper introduces a simulation method for creating environments with causal relationships, enabling the testing of uplift modeling techniques without relying solely on experimental data, thus addressing practical challenges in model deployment and adaptation.

Contribution

The paper proposes a novel simulation approach that models causal relationships, facilitating the evaluation of uplift models in non-experimental, drifting environments.

Findings

Simulated environments with causal parameters enable testing of uplift models.

The method supports evaluation in environments with concept drift.

Provides a tool for assessing models without extensive experimental data.

Abstract

Uplift modeling requires experimental data, preferably collected in random fashion. This places a logistical and financial burden upon any organisation aspiring such models. Once deployed, uplift models are subject to effects from concept drift. Hence, methods are being developed that are able to learn from newly gained experience, as well as handle drifting environments. As these new methods attempt to eliminate the need for experimental data, another approach to test such methods must be formulated. Therefore, we propose a method to simulate environments that offer causal relationships in their parameters.