GCF: Generalized Causal Forest for Heterogeneous Treatment Effect Estimation in Online Marketplace

Abstract

Uplift modeling is a rapidly growing approach that utilizes machine learning and causal inference methods to estimate the heterogeneous treatment effects. It has been widely adopted and applied to online marketplaces to assist large-scale decision-making in recent years. The existing popular methods, like forest-based modeling, either work only for discrete treatments or make partially linear or parametric assumptions that may suffer from model misspecification. To alleviate these problems, we extend causal forest (CF) with non-parametric dose-response functions (DRFs) that can be estimated locally using a kernel-based doubly robust estimator. Moreover, we propose a distance-based splitting criterion in the functional space of conditional DRFs to capture the heterogeneity for the continuous treatments. We call the proposed algorithm generalized causal forest (GCF) as it generalizes the use case of CF to a much broader setup. We show the effectiveness of GCF by comparing it to popular uplift modeling models on both synthetic and real-world datasets. We implement GCF in Spark and successfully deploy it into DiDi's real-time pricing system. Online A/B testing results further validate the superiority of GCF. CCS CONCEPTS• Computing methodologies → Kernel methods; Classification and regression trees; • Applied computing → Decision analysis; • Mathematics of computing → Continuous functions.

Shenzhen, China

“…Given the importance and unique challenges of uplift modeling, researchers in academia and industry have done extensive research in recent years. Uplift prediction models have evolved from metalearners-based [4,19,23], tree and forest-based [1,3,9,24,29,30], Knapsack problem-based [2,14] to deep neural networks-based architecture. Notably, many recent works have focused on developing new neural network architectures to better adapt to uplift modeling in industrial scenarios and shown remarkable performance improvements, such as EUEN [18], DESCN [34] and EFIN [21] and so on.…”