Analysis On Depression Tendency And Causes Of Weibo Users

Depression is a serious psychological disorder,which harms health.Although depression can be diagnosed and treated effectively,there are still large numbers of people suffering from depression who are not effectively cured due to a lack of professional diagnosis.The burden of medical expenses and the concealment of depression had hindered patients from getting rid of psychological diseases.With the continuous development of social networks,the microblogged platform attracts more and more users to express and share their emotions in life.As one of the social software for people to actively express emotions,the content of microblogs is more reliable,and it also provides an effective way for people to understand the mentation and mental activities of patients with depression.Therefore,how to accurately identify users with depressive tendencies and conduct psychological counseling is a valuable topic.At present,domestic and foreign scholars mainly focus on the analysis of the characteristics of emotional tendency.In this paper,we explore the expression of depression of microblog users using text mining based on the release time of microblog and study psychological characteristics and behavioral characteristics.Based on the constructed depression characteristic dictionary,a depression tendency scoring model was established.The method of identifying depression tendency from single text to continuous text completes the accurate identification of microblog users.The proposed method can be applied to the identification of depression-prone users and the counseling of depression on social platforms.The study data are obtained from Weibo comments and blog posts by crawler technology.The structure of the paper is as follows.The second chapter describes the expression characteristics of the micro-blog text of patients with depression obtained through data cleaning,word segmentation,and other preprocessing.It is found that depressed users prefer to publish micro-blog at night.The third chapter establishes a dictionary of depression characteristics by generating similar words,extracts depression characteristics in microblog texts,and conducts vector quantization.This paper extracts depression information about microblog text based on depression tendency.Based on the characteristic score vector of depression,the differences in emotion and behavior between depressed users and non-depressed users were analyzed.The fourth chapter establishes a depression tendency score model.The logistic regression model was established to obtain the score of depression tendency of each blog,and the ROC curve was used to optimize the threshold of depression tendency of the model.The test set verifies that the prediction accuracy of the model is 0.88,indicating that the depression tendency score model has a good effect on depression recognition of microblog texts.By continuously monitoring Weibo users’ blog posts,this paper identifies their depression tendency based on the percentage of depressed blog posts.The fifth chapter establishes a dictionary of depression causes.By extracting the words of depression causes of microblog texts,this chapter summarizes the important factors leading to users’ depression tendency and summarizes them into four aspects: family,workplace work,love status,and interpersonal relationship.Finally,it is proposed to apply reading therapy to the psychological counseling field of depressed users,which is of great value to the treatment of depressed patients.