Random Forest Algorithm-based Modelling and Neural Network Analysis Between Social Anxiety Disorder of Childhood and Parents' Socioeconomic Attributes

Abstract

Using the random forest algorithm in machine learning, the problem of children's social phobia is transformed into a classification prediction problem. There are many reasons for social anxiety disorder in childhood (SADC). Thus, we study the influence of parents' socioeconomic attributes on SADC. Based on the data obtained from the questionnaire survey of children and their parents in an early education institution, we build a prediction model between SADC and parents' socioeconomic attributes with the bivariate correlation method, the logistic regression, and the random forest method. The study result shows that the parents' socio-economic attributes are strongly related to SADC, and the model can be applied to the personalized care and psychological intervention of this early education institution. The result also shows that the accuracy reaches 80.5%. The model can be applied to preschool prediction and screening of children's social phobia tendencies and provides a reference for teachers to give personalized care and psychological intervention to children with a high tendency in follow-up teaching activities.