Sentiment Analysis on Twitter Social Media Regarding Depression Disorder Using the Naive Bayes Method

Lavenia, Nur Lickha
Permatasari, Reisa

Coreid Journal
Vol. 1 No. 2 (2023): July 2023

Depression disorder is a serious issue in mental health that affects many individuals worldwide. This research analyzes the sentiments related to depression disorder on Twitter using the Naïve Bayes method. Depression-related tweet data was collected through snscrape and processed to eliminate irrelevant information. Three Naïve Bayes methods, namely Multinomial, Gaussian, and Bernoulli, were compared to classify positive, negative, or neutral sentiments in each tweet. The results of the study indicate that Multinomial Naïve Bayes exhibited the best performance with an accuracy rate of 90.13%, followed by Gaussian Naïve Bayes (88.38%), and Bernoulli Naïve Bayes (85.37%).