DexDeepFM: Ensemble Diversity Enhanced Extreme Deep Factorization Machine Model

Predicting user positive response (e.g., purchases and clicks) probability is a critical task in Web applications. To identify predictive features from raw data, the state-of-the-art extreme deep factorization machine model (xDeepFM) introduces new interaction network to leverage feature interactions at vector-wise level explicitly. However, since each hidden layer collection of maps, it can be viewed essentially as an ensemble different maps. In this case, only using single objective minimize prediction loss may lead overfitting generate correlated errors. article, diversity enhanced (DexDeepFM) proposed, which designs measure considers both accuracy function. addition, attention mechanism introduced discriminate importance measures with orders. Extensive experiments on three public real-world datasets are conducted show effectiveness proposed model.