Research on the Method of Electronic Literature Recommendation Based on TextCNN Model
This paper puts forward the method of electronic journal literature recommendation based on TextCNN model,aiming to better accurately grasp the deep relationship between the essential characteristics of literature content and the demands of users,and realize the personalization and precision of electronic journal literature recommendation service.The word2vec is used to vector quantify the literature title information,the TextCNN model is used to train the literature recommendation model,and finally the literature that meets user needs is actively pushed to scientific research users.Practice has proved that the recommendation model designed in this paper can recommend electronic journal literature for users with good results.
TextCNNText classificationElectronic literature recommendationBehavior data
NETL
NSTL
万方数据
