Machine Learning Technique to Detect and Classify Mental Illness on Social Media Using Lexicon-Based Recommender...

Area of Science:

• Natural Language Processing
• Computational Linguistics
• Sentiment Analysis

Background:

• Social media enables user expression of feelings, influencing decision-making.
• Analyzing emotions in natural language presents challenges, particularly a lack of precise classification resources.
• Existing sentiment analysis often focuses on polarity, which is insufficient for nuanced emotion classification.

Purpose of the Study:

• To develop a grammar rules-based classification system for Indian language tweets.
• To address key challenges in classifying emotions within informal and morphologically rich Indian languages.
• To enhance the accuracy of sentiment analysis for Indian language social media data.

Main Methods:

• A novel approach combining grammar rules (adjectives, negations) for informal tweet classification.
• Utilizing 'field tags' to incorporate non-grammatical elements like slang and abbreviations.
• Integrating grammar rules with N-gram techniques and machine learning for complex Indian languages.

Main Results:

• Identified informal nature, slang/abbreviations, and morphological richness as key challenges.
• Proposed methods to effectively handle these challenges in Indian language tweet analysis.
• Developed a system that functionally predicts sentiments using syntactic words in Indian languages.

Conclusions:

• The proposed grammar rules-based method offers a robust solution for emotion classification in Indian language tweets.
• The techniques developed can improve the accuracy and applicability of sentiment analysis in diverse linguistic contexts.
• This research contributes to better understanding and analyzing user emotions expressed on social media platforms in India.