User individual behavior prediction method based on DPI time sequence word embedding vector

The invention discloses a user individual behavior prediction method based on a DPI time sequence word embedding vector. The method comprises a data preprocessing step S1, a word vector embedding step S2 of constructing a user access DPI access sequence through deep learning, a model building step S3 and a click rate prediction step S4. Therefore, according to the method, internal association between user access DPI is learned through a word vector embedding technology based on a deep learning framework, and the internal association and original user features are used as model input to obtain a final dichotomous learner for predicting user click, namely, the dichotomous learner is inspired by the word vector technology in natural language processing; the DPI sequence accessed by the user is learned through the word vector technology, association in the DPI sequence is mined, and the method has important theoretical and application values for improving the accuracy of user click rate prediction. 一种基于DPI时间序列词嵌入