Diagnosa Tingkat Depresi Mahasiswa Selama Masa Pandemi Covid-19 Menggunakan Algoritma Random Forest

Abstract

Covid-19 virus has become a pandemic across the world, including Indonesia. Based on the data from the Covid-19 Handling Officer Unit, the number of Covid-19 sufferers in Indonesia until February 15, 2021 reaches 1.2 million people. The number of daily cases that continues to grow has forced the government to enforce policies to work, study, and worship from home to minimize the Covid-19 transmission. The policy and many Covid-19 sufferers Indonesia affect the mental health of people, including students of Singaperbangsa Karawang University. Therefore, this research aims to diagnose the initial level of depression in students of Singaperbangsa Karawang University during Covid-19 pandemic by using data mining with Random Forest algorithm. The method used in this research is KDD (Knowledge Discovery in Database) with data used come from PHQ-9 questionnaire given to 392 respondents according to calculation of Slovin formula. Evaluation model used is 10-fold cross validation, with accuracy, sensitivity and specificity parameters. The results of the research show the depression level prediction model using Random Forest algorithm has an accuracy of 85.94%. From 392 students, 1.02% of students are normal, 47.96% have mild depressive symptoms, 36.73% have mild depression, 8.16% have moderate depression, and 6.12% have major depression.