A text classification framework for depression tendency detection over online social platform
Yunchun Liao, Jian Shu, Sicheng Liu · 2021
As social networks become more and more prevalent, people like to share their feelings on the Internet, which provides a new research idea for the detection of depression patients. In this paper, we propose to detect the depression tendency of microblog texts based on LSTM model. First, we embed the preprocessed data set into word vectors, then put them into the model, LSTM taking the word vector layer as the first hidden layer, building several hidden layers, the output layer outputting the labels of prediction data and the detection of depression tendency on the batch micro blog data. Through many experiments, we constantly adjust the super parameters of the model, and in the end, we achieved an accuracy of 93.68%, which is superior to the general machine learning model, and makes the research on depression in online social platform have practical value.