Click-through rate (CTR) prediction occupies a very important position in recommendation systems and computational advertising. Accurate prediction of the click-through rate can help advertisers make more detailed advertising and marketing planning and can also help the media in defining their service audience and recommend more popular content to users more clearly. In this paper, we propose a model to make full use of the explicit feature interactions to improve the click-through rate of advertising. In the second-order feature interactions, our model not only considers the fact that the same feature uses different latent vectors when interacting with other features but also adds the latent vectors of each feature to form the embedding vector of the whole feature and then carries out explicit high-order interaction, to learn more nonlinear relations. We, inspired by the Squeeze-and-Excitation Networks (SENet), give different weights to different feature interactions. Extensive experimentation has been done on Avazu and Criteo to show that our model is better than the existing advanced models.