Volume 29, Issue 2
DOI: 10.24205/03276716.2020.400
MENTAL HEALTH ASSESSMENT OF SOCIAL NETWORK USERS BASED ON CONVOLUTIONAL NEURAL NETWORK
Abstract
In the age of the Internet, more and more users mention their mental health problems anonymously on social network sites. These users should be identified and intervened timely to prevent their mental health from deteriorating. This paper constructs a mental health assessment model based on convolutional neural network (CNN), which automatically evaluates the mental health of social network users, and the urgency of their intervention demand. Under the guidance of psychological knowledge, this model relies on the CNN to mine the statistical features of word frequency in different categories of posts, and then pinpoints the posts that need to be intervened. Experimental results show that, compared with other methods, our model can achieve a desirable recognition effect for social network users who need urgent psychological intervention. The research results reflect the value of psychological knowledge in feature extraction through deep learning.
Keywords
Mental Health Assessment, Psychological Intervention, Convolutional Neural Network (CNN), Social Network.