Analysis Of User Portrait Of Mental Health Cognitive Attitude

User portrait is a data analysis method that abstracts the labeled user model based on the relevant information of the research object and forms a representative user role for the target group.With the development of data science and big data technology,user portrait as a method of researching the characteristics of target objects,its application fields continue to expand,from the commercial market to various fields of social life,such as the field of mental health research.Under the fast pace of modern society and the pressure of competition,mental health issues have gradually attracted different levels of attention from all walks of life.Studies have shown that the higher the individual’s attention to their own psychological problems,the better the effect of their control of mental illness.Therefore,the use of scientific methods to study individuals’ cognitive attitudes towards mental health is of great practical significance for the prevention and management of mental illness.Using the basic principles and basic process framework of user portraits,and using questionnaire data on workers’ mental health cognition,construct a mental health cognitive attitude portrait model,and discuss effective methods for establishing mental health cognitive risk early warning models.Recognizing the mental health cognitive attitude of the target object through the model can provide a reference for related management and research.On the basis of literature study,combined with the relevant theories of psychology,an analytical index system is established.Using k-modes clustering method based on feature importance,according to the five-level division of mental health cognitive attitude research,mental health cognitive attitudes are divided into positive,more positive,neutral,more negative and negative.Abstract the corresponding characteristics of various types.People with negative mental attitudes are at a low level of cognition and expression of mental health.These people are high-risk groups of mental health and need special attention and attention.Therefore,it is necessary to further build an early warning model of mental health cognitive attitude risk.For the three commonly used machine learning classification algorithms: random forest classification algorithm,logistic regression algorithm and KNN classification algorithm.Through data modeling comparison,it is found that the logistic regression algorithm in this example has high classification accuracy;when exploring to improve classification accuracy,two methods of random forest feature selection and chi-square test feature selection are used for feature screening,and it is found that random forest features are used Choosing a method can help the logistic regression algorithm to improve the classification accuracy,and the improved classification accuracy can reach 96.18%.Since the data in this example is qualitative data,the features are multi-valued features,and the classification results are greater than 2,so it can be considered that for this type of data,the logistic regression algorithm is the better algorithm.This article builds a user portrait model of mental health cognitive attitude,which provides a new analytical method and idea for psychology research;at the same time,in the model classification algorithm and feature screening,it provides a better model algorithm for qualitative data For reference,the application of machine learning algorithms is usefully explored.Therefore,the study of the thesis has both theoretical and practical significance.