Implementation of Naïve Bayes Algorithm for Sentiment Classification of Public Youtube Opinions Related to Nickel Mining Issues in Raja Ampat

Abstract

Indonesia currently holds the world’s largest nickel reserves. However, extractive activities in the Raja Ampat conservation area pose significant ecological threats and trigger social polarization on social media. The massive volume of opinion data creates challenges for policymakers in mapping public perception quickly and objectively. This study aims to classify public sentiment regarding nickel mining activities in Raja Ampat using the Naïve Bayes algorithm with TF-IDF feature weighting. The methodology employed in this research is CRISP-DM, which consists of Business Understanding, Data Understanding, Data Preparation, Modeling, Evaluation, and Deployment stages. The dataset consists of 10,903 YouTube comments collected from the Ferry Irwandi and Kompas.com channels. The results indicate that negative sentiment dominates public discourse at 46.4%, followed by neutral sentiment at 31%, and positive sentiment at 22.6%. The classification model achieved an accuracy rate of 71.59%. Furthermore, a Streamlit-based visualization dashboard was developed to assist stakeholders in monitoring public opinion trends systematically.