Sentiment Analysis Of Mental Health Using K-Nearest Neighbors On Social Media Twitter

Abstract

Mental health issues are still significant health problems in the modern world. Poor understanding, stigma, and low mental health awareness contribute to efforts to educate people about mental health. The issue of mental health is widely discussed on social media, one of which is Twitter. So it is necessary to analyze sentiment on mental health issues on Twitter. The dataset in this study uses community reviews on May 16, 2022, with the search word "Mental Health." The research method used in this study has several stages, such as processing data using the K-Nearest Neighbors algorithm by comparing the Support Vector Machine classification algorithm and Decision Tree Processing research data using Rapid Miner tools. The conclusion of this study is, based on experimental results with a dataset of 639 positive and 193 negative sentiment reviews. The results of modeling processing using the K-Nearest Neighbors algorithm obtained the best results when using the split data method 70:30 with k value at number 5, producing precision of 60.87% and recall of 44.03%, respectively, and accuracy of 58.39%.