Personality classification via Weibo based on deep learning model

With the coming of the information age and the popularity of Weibo, people will publish their daily behavior records on Weibo. These Weibo data are linked to individual personality, which can be accurately predicted through Weibo text. But this work only remains a theoretical target in the field of medical psychology, psychiatry, etc. and has not been implemented. The project firstly marks the personality of Weibo users, and builds over ten thousands of Chinese Weibo personality classification data base. We uses word2vec to veetorize the Weibo text, extracts the features of the Transformer encoder, builds a model by lstm to train the personality classification, and establishes Personality classification model combined the transformer with lstm. Finally, by comparing with the conventional model, the results show that the model has better personality detection results in terms of precision, recall and F1 score.