SANTOSO, KENLY KRISAGUINO (2025) COMPARING LOGISTIC REGRESSION AND NAÏVE BAYES ALGORITHM FOR DEPRESSION IDENTIFICATION FROM SOCIAL MEDIA USAGE. S1 thesis, UNIVERSITAS KATOLIK SOEGIJAPRANATA.

Preview	Text 21.K1.0017_KENLY KRISAGUINO SANTOSO_COVER.pdf Download (493kB) | Preview
	Text 21.K1.0017_KENLY KRISAGUINO SANTOSO_BAB I.pdf Restricted to Registered users only Download (473kB)
	Text 21.K1.0017_KENLY KRISAGUINO SANTOSO_BAB II.pdf Restricted to Registered users only Download (472kB)
	Text 21.K1.0017_KENLY KRISAGUINO SANTOSO_BAB III.pdf Restricted to Registered users only Download (738kB)
	Text 21.K1.0017_KENLY KRISAGUINO SANTOSO_BAB IV.pdf Restricted to Registered users only Download (591kB)
	Text 21.K1.0017_KENLY KRISAGUINO SANTOSO_BAB V.pdf Restricted to Registered users only Download (405kB)
Preview	Text 21.K1.0017_KENLY KRISAGUINO SANTOSO_DAPUS.pdf Download (471kB) | Preview
	Text 21.K1.0017_KENLY KRISAGUINO SANTOSO_LAMP.pdf Restricted to Registered users only Download (530kB)

Abstract

Social media has changed how people live today, bringing positive and negative effects on their mental health. It can get the user addicted, anxious, feeling alone, and depressed which affects their relationships and responsibilities in real life. A Lot of studies use the Naïve Bayes Algorithm for detecting mental health problems but the Logistic Regression Algorithm is used the least. Some studies show that Naïve Bayes has a high accuracy, reaching 92.5%. This study aims to compare both the Naive Bayes Algorithm and Logistic Regression to detect mental health problems through social media usage. The datasets this study uses are 481 user data obtained from Kaggle containing their use of social media patterns and mental health conditions. By using both Naïve Bayes and Logistic Regression, this study hopes to analyze depression signs by the users’ social media usage and to find which algorithm performs the best in terms of accuracy, precision, recall, and F1 scores.

Actions (login required)

View Item