Text Classification based on Model of Feature Fusion and Voting Mechanism for Personality Recognition
Download as PDF
Shoryu Teragawa, Ming Li
Psychology, as a discipline that relies on traditional statistical methods, has its limitations in research methods. Based on this fact, this article attempts to make up for this shortcoming by deep learning method. First, the words in the user text are converted into vectors by frequency through embedding, and the feature information in the user text is extracted through different neural networks. Then feature fusion through voting and other methods enables more accurate user information to be used for classification. According to the experimental results, it is found that the model proposed in this paper is more effective in extracting multi-dimensional features from user texts, and effectively optimizes traditional algorithms. Compared with traditional models, the accuracy is improved.
User personality recognition; multi-dimensional features; neural network