Automatic Historical Feature Generation through Tree-based Method in Ads Prediction

TL;DR

This paper introduces a tree-based method for automatically generating historical features in ad CTR prediction by selecting counting keys through decision trees trained per user, improving over manual features.

Contribution

The paper presents a novel tree-based approach for automatic historical feature generation tailored to user-specific data in ad CTR prediction.

Findings

Automatically identified features outperform manual features in experiments.

The method is effective in both online and offline settings.

Large-scale experiments validate the approach on Twitter video ads.

Abstract

Historical features are important in ads click-through rate (CTR) prediction, because they account for past engagements between users and ads. In this paper, we study how to efficiently construct historical features through counting features. The key challenge of such problem lies in how to automatically identify counting keys. We propose a tree-based method for counting key selection. The intuition is that a decision tree naturally provides various combinations of features, which could be used as counting key candidate. In order to select personalized counting features, we train one decision tree model per user, and the counting keys are selected across different users with a frequency-based importance measure. To validate the effectiveness of proposed solution, we conduct large scale experiments on Twitter video advertising data. In both online learning and offline training settings, the automatically identified counting features outperform the manually curated counting features.