Sentiment-analysis to detect early depressive symptom in Bangla language from social media: a review study

• Main Authors: Hassan, Md. Hasibul, Kamaruddin, Azrina, Azmi Murad, Masrah Azrifah
• Format: Article
• Published: Little Lion Scientific R&D 2021
• Online Access: http://psasir.upm.edu.my/id/eprint/95044/; http://www.jatit.org/volumes/ninetynine16.php

Social media platforms hold a vast volume of raw data that has been posted by people in the forms of texts, images, audio and video. People use this medium to express their thoughts and opinions. As a result, the data can be captured, categorized, and analyzed using Sentiment Analysis approaches to identify users' behavior, customer's feedback or gauge public opinions. WHO reported that the numbers on existing mental health disorders are a troubling phenomenon. The identification of mental health can be detected using several data domains such as: sensors, text, structured data, and multi-modal system use. Several researches focus on specific public sentiments for example Malays, English, Arabic, Chinese and Korean. However, very little research was conducted about sentiment analysis approaches implementation in Bangla language. The purpose of this paper is to explain the knowledge gap and the proposed model by using Bangla language sentiment analysis. In this paper we have reviewed 50 articles from which 18 are listed here that have the most similarity with our research. The review shows that the mostly used method in Sentiment-analysis is Machine Learning in the field of Opinion-mining. Furthermore, we have identified another variable that can be included to improve the existing algorithm.