Authors:
G. Kalpana, A. S. Aathira, K. Shantichitra, B. Jaiganesh
Addresses:
Department of Computer Science, Faculty of Science and Humanities, SRM Institute of Science and Technology, Kattakulathur, Chennai, Tamil Nadu, India. Department of English, SRM Institute of Science and Technology, Kattakulathur, Chennai, Tamil Nadu, India. Department of English, Faculty of Science and Humanities, SRM Institute of Science and Technology, Kattakulathur, Chennai, Tamil Nadu, India. Department of Tamil, Faculty of Science and Humanities, SRM Institute of Science and Technology, Kattakulathur, Chennai, Tamil Nadu, India.
Abstract:
Twitter is the ultimate micro-blogging social networking tool, generating 6000 Tweets per second, or 500 million per day. Twitter generates and stores a lot of data because businesses and politicians market their brands there. Classifying and identifying tweet sentiments is sentiment analysis. To represent public opinion, sentiment analysis analyses available data to capture feelings. Twitter sentiment analysis is harder than wide sentiment analysis due to misspellings, slang, and repeated words/characters. Word emotion must be determined. SDSNN ML techniques provide sentiment analysis for current feedback/reviews on impending news-based data, making it more accurate and efficient. Region-level sentiment analysis shows how domain information impacts classification. Sentiment analysis involves data collection, processing, RST feature selection, categorisation, and prediction. An innovative feature model for tweet classification into neutral, positive, and negative categories, and for news-related public opinion extraction, is advised to identify the best Sentiment Analysis approach for large datasets. This study analyses sentiment analysis as a classification and computational linguistics problem. In the model's interpretative framework, Twitter's morphological noise, slang, and pragmatic intricacy of sarcasm are examined. It makes the SDSNN model human-centric in communication.
Keywords: Sentiment Analysis; Feature Selection; Rough Set Theory; Semi-Deep Structural Neural Network (SDSNN); Classification and Prediction; Support Vector Machine (SVM).
Received on: 12/11/2024, Revised on: 02/01/2025, Accepted on: 01/04/2025, Published on: 03/01/2026
DOI: 10.64091/ATICL.2026.000251
AVE Trends in Intelligent Computer Letters, 2026 Vol. 2 No. 1 , Pages: 1–12