Page 182 - Demo

1st Int. Transborder Conf. of the Timor Island: Timor %u2013 Science without borderDili, 7-8 May 2025181Text Mining and Word Cloud Visualization of Public Sentiment on Timor-Leste%u2019s ASEAN Membership Using Machine Learning Marcelino C. Noronha*, Frederico S. Cabral, Jose S. Pinto, Ferdinando da Concei%u00e7%u00e3o Soares and Quintino SaoresDepartment of Informatics Engineering, Faculty of Engineering Science and Technology , National University of East Timor, Timor-Lestee-mail: marcelino.noronha@untl.edu.tl; pintosaran81@gmail.com, ferdinando.soares@untl.edu.tl.Abstract This study examines public sentiment regarding Timor-Leste's potential ASEAN membership through an innovative approach combining text mining, word cloud visualization, and machine learning techniques. Using YouTube comment data as our primary source, we implement Natural Language Processing (NLP) methods for text preprocessing and feature extraction, followed by sentiment classification using Support Vector Machine (SVM) algorithm. The research incorporates word cloud visualization to identify key themes and frequently occurring terms in public discourse. Our SVM model achieved 80.85% accuracy in classifying sentiments into positive, negative, and neutral categories. Word cloud analysis revealed economic benefits, sovereignty concerns, and regional cooperation as dominant themes shaping public opinion. The study demonstrates how the combination of computational text analysis and visual analytics can provide comprehensive insights into complex political sentiments. These findings offer valuable perspectives for policymakers regarding public perception of regional integration, while also presenting a methodological framework for social media-based political sentiment analysis in developing nation contexts. Keywords: text mining, sentiment analysis, word cloud, Timor-Leste, ASEAN membership, machine learning, SVM.