Polyhedral Conic Classifier for CTR Prediction
This work addresses CTR prediction problems for industrial recommender systems, representing an incremental improvement with a novel method for a known bottleneck.
The paper tackles the challenges of numerical imbalance and geometric asymmetry in click-through rate (CTR) prediction for industrial recommender systems by introducing a deep neural network classifier using polyhedral conic functions, achieving superior performance over Binary Cross Entropy Loss on four public datasets.
This paper introduces a novel approach for click-through rate (CTR) prediction within industrial recommender systems, addressing the inherent challenges of numerical imbalance and geometric asymmetry. These challenges stem from imbalanced datasets, where positive (click) instances occur less frequently than negatives (non-clicks), and geometrically asymmetric distributions, where positive samples exhibit visually coherent patterns while negatives demonstrate greater diversity. To address these challenges, we have used a deep neural network classifier that uses the polyhedral conic functions. This classifier is similar to the one-class classifiers in spirit and it returns compact polyhedral acceptance regions to separate the positive class samples from the negative samples that have diverse distributions. Extensive experiments have been conducted to test the proposed approach using state-of-the-art (SOTA) CTR prediction models on four public datasets, namely Criteo, Avazu, MovieLens and Frappe. The experimental evaluations highlight the superiority of our proposed approach over Binary Cross Entropy (BCE) Loss, which is widely used in CTR prediction tasks.