Towards Understanding the Psychological Effects of the COVID-19 Pandemic on the Indian Population

ABSTRACT

This article explains the preliminary results of the analysis of a public survey carried out in India, assessing the psychological effects on people during the second wave of the COVID-19 pandemic. A survey was designed to categorize the population on the basis of various socio-economic demographics and respondents were then asked to fill out the DASS-21 questionnaire to get their levels of severity of anxiety, depression and stress. The dataset obtained was then further analyzed using various classification machine learning models with the level of severity as the target variable and respondent's attributes as independent variables. A Multinomial Logistic Regression was found to give the best results with an AUC score of 0.94 and was thus, used to predict the severity levels of these three categories, to find various insights from this publicly-sourced dataset. Additionally, the significance of the various socio-demographic attributes asked in the survey was analyzed in order to identify key drivers of mental ailments among the general Indian population. Further, a brief description of segmenting the population using K-Means clustering is provided which attempts to identify population groups that belong to similar socio-economic demographics and suffer from similar mental health issues during the pandemic. Thus, high-risk or high-severity groups can be identified and then could be targeted by the government to provide them relief schemes. This paper applies machine learning on a public dataset to explore the various facets of COVID-induced problems in the Indian Society. © 2021 IEEE.