Depression Detection Using Linear Regression Model

According to the World Health Organization (WHO), globally around 5% of adults suffer from depression. Depression in teenagers and working people is a lot common these days. Depression not only affects the lives of individuals but also affects the person around them. Witnessing loved ones struggling with depression is emotionally challenging. This study is based on analyzing the types of depression, and their factors. Through an extensive literature review, we found several factors that lead to different types of depression. These factors include loss of interest, sleeping problems, feelings of worthlessness, and many more which result in different types of depression such as major depression, psychotic disorder, seasonal affective disorder. A causal effect analysis is done by using a decision making trial and evaluation laboratory to study the interrelationships among these factors and their impact on depression. Based on the findings a web application was developed to check if someone is suffering from depression using a linear regression model. The performance and accuracy of the machine learning model have been evaluated using the r-squared method and MSE. The r-squared value of this model is 0.8668 which represents 86.68% of variance in target and MAE is 0.2641 which indicates the absolute difference between absolute value and predicted value. This web application is created for students, people who have recently started to work and people who are unemployed. The web application asks multiple questions related to their day-to-day lives and according to the user input it finds which type of depression a person has and what grade it is.