Research On The Technology Of Detecting And Warning Abnormal Network Behavior Of Students In School

In recent years,the number of netizens in China has increased dramatically with the rapid development of the Internet,among which students account for the largest number of netizens.As of June 2019,students accounted for 26.0% of the total number of netizens.Students the suicide,depression,posted on the Internet users in the network bullying behavior,anxiety and other information referred to as network information,they in weibo,BBS,post bar and other platform using the information such as text,audio,video,images,to express their emotions,which release the greatest influence on society,the suicide of information is the key point of this study.The traditional methods of student group identification and suicide tendency detection cannot meet the current needs due to the low accuracy rate.With the advent of big data and artificial intelligence,it is particularly important to use more intelligent methods for detection.Aiming at the recognition of students in school,this paper proposed to use the TextCNN algorithm of deep learning,and carried out experiments on the TextCNN,FastText and TextRNN models,respectively.It was found that the TextCNN model had the best effect on the recognition of students in school.As for the detection of network abnormal behaviors,this paper puts forward the detection of network abnormal behaviors based on emotional classification.Based on emotional classification,it analyzes the abnormal behaviors of users with negative emotions,focusing on the detection of network abnormal behaviors with suicidal tendency.The method of suicide tendency detection is to extract keywords by tf-idf algorithm and represent text information by vector space model.Then,RF,KNN,LR,NB,SVM and DT are tested.The RF model has the highest accuracy.For the RF model,by adding artificially constructed indicators,the experimental comparison again found that the accuracy of suicide tendency detection increased by 1.7%.The experimental results show that the method used in this paper can basically realize the detection and early warning of network abnormal behaviors of school students.This method can quickly obtain the students with suicidal tendencies in school,provide these information to relevant departments for early warning,and intervene in the occurrence of suicidal behaviors.