Prediction Of Intervention Timing For Social Media Users’ Psychological Crisis

In recent years,psychological crisis has gradually become a hot topic of discussion.Many crimes and suicide cases caused by psychological crisis have attracted a lot of attention in various media.Crisis monitoring and prevention and control has become an urgent problem.At present,social media is an important way for many people to express their emotions,and it is also a very important window for psychological crisis monitoring.At present,mental health research based on social media focuses on the identification of psychological crisis.There is no work to study the timing of psychological crisis intervention,and the willingness of users to accept psychological crisis assistance is ignored,which may have a negative impact on the communication between users and donors.The intervention willingness of psychological crisis is essentially the user’s willingness to help the outside world.Many users will post for help on the forum when they feel sad.The main goal of this paper is to analyze and predict the behavior of users’ posting for help.Firstly,according to the knowledge of psychostatistics and natural language processing technology,we extract the features of users’ posts on reddit forum.According to the method of time window partition,the data set is processed,and the content characteristics of users’ social texts are analyzed vertically to analyze users’ psychological changes.Through the analysis of linguistic features,we can observe the differences of frequency(static)of different parts of speech in this period of time;with the analysis of emotional dynamics features,we can find the differences of frequency fluctuation(dynamic)of different parts of speech.In addition,by analyzing the distribution of posts with different suicide risk levels in time window samples,we can draw a conclusion: users with a particularly high suicide risk level are less willing to accept intervention,which is consistent with the psychological conclusion,and also shows the importance of time-class identification and active intervention of psychological crisis.In the process of building the prediction model,we define whether the user has intervention intention according to whether the user has post behavior on the suicidewatch,and build positive and negative samples of data set including intervention intention according to the time window division strategy.Finally,theintervention timing is predicted based on SVM.The experimental results show that the method proposed in this paper can effectively identify the intervention time of users with psychological crisis.By analyzing the characteristics,the prediction model is evaluated according to precision,recall and F-score.The most effective prediction model,whose features are divided into adaptive time windows,and the LGF,NGF,and EDF features are fused.Finally,the prediction effect of the prediction model on the data set is: F1 value is 0.711,accuracy rate is 0.786,recall rate is 0.649.