Combining feature interaction and channel attention for Click-Through Rate Prediction

Click-through rate prediction is in considerable demand and has been extensively studied because it allows advertisers and advertising platforms to target consumers for particular marketing contents. In this paper, we propose a novel model that can learn the interaction between various dimensions. Besides, our model can also strengthen the weight of important interactions and weaken the weight of unimportant relationships. Furthermore, our model can learn not only the low-order interaction between features but also the high-order interaction between features, which improves the interaction ability of features. We performed extensive experiments using two open datasets of Avazu and Criteo, which proved that our model is better than other state-of-the-art models.