Research On The Identification Algorithm Of College Students’ Mental Health Problems Based On Campus Big Data

In recent years,a series of mental health problems such as anxiety,depression,inferiority complex,interpersonal sensitivity,etc.have occurred frequently among the college student population,and even more have also produced suicidal thoughts.This has a very serious negative impact on individuals,families and society.If the mental health problems of college students can be discovered as early as possible,the relevant departments and counselors of the school can provide targeted assistance to such students in a timely manner.At the same time,these high-risk students can be treated early to reduce harm.Therefore,it is very valuable to find an effective method to find students with mental health problems early.Traditionally,researchers usually use a questionnaire survey,which has the disadvantages of easy concealment and low efficiency.In recent years,researchers have begun to use web logs to identify students with mental health problems,but this method still has deficiencies.First,they still use questionnaires to obtain tags.Secondly,students’ psychological activities are not only reflected in online behaviors,other daily behaviors may also express their psychological activities.In order to obtain a better recognition effect of college students’ mental health problems,this thesis attempts to use multi-source data,the data set includes:consumption data,entry and exit dormitory data,network logs,historical score data,and proposes a mental health problem recognition algorithm based on multi-source data.The main research focus is feature extraction.Attempt to use 1D Convolutional Neural Network(1D-CNN)to mine students’ online trajectory patterns from online behavior sequences;according to the consumption data of students in the cafeteria,abnormal scores are calculated to characterize the differences in diet between students.At the same time,this thesis uses the students’ psychological state data provided by the psychological center as a label to improve the deficiencies caused by the questionnaire.This thesis uses the training set to train five common classification algorithms,and evaluates them through the validation set,selects the classifier with the best effect as our algorithm,and uses it to identify students with mental health problems in the test set.Experimental results show that Precision reaches 0.68,Recall reaches 0.56,and F1-Measure reaches 0.67.In order to further optimize the model recognition effect,this thesis proposes a mental health problem recognition algorithm based on the Deep Psy model.Constructed an online trajectory matrix,used 2D Convolutional Neural Network(2D-CNN)to extract one-day online patterns,and used Long Short-Term Memory Network(LSTM)to capture the time dependence between each day.Combined with basic features and online trajectory patterns,we design a deep learning network.Experiments show that Precision reaches 0.71,Recall reaches 0.75,and F1-Measure reaches 0.72,which can identify 75% of students with mental health problems.