Multiple Linear Regression and Bagging-based Analysis and Modeling of Influence of Mother's Socio-economic Attributes on Anxiety of Online Education

Abstract

Many mothers expressed varying degrees of anxiety about their children's online lessons at home. To investigate the related phenomena, data samples were obtained through questionnaires on mothers. The dependent variable of each sample was the anxiety degree towards online classes, and the independent variable was the various socioeconomic attributes of mothers. The results indicated that there was a strong correlation between the mother's anxiety about online classes and socioeconomic attributes such as age, educational background, income, and the elderly person at home accompanying children in online classes. The results of linear regression modeling indicate that it is difficult to fit a simple linear relationship between dependent and independent variables. The integrated learning model based on Bagging indicates that the dependent and independent variables can be fitted into a more complex numerical relationship. The experimental results show that the classification accuracy is 91.6%. The portrait characteristics of mothers with high anxiety about online classes include unaccompanied children at home during online classes, family income, and the mother's educational background. When there was only one child in the family, the age difference between mother and child was significantly larger than the average difference of all subjects.