Sentiment Analysis of Unemployment in Indonesia During and Post COVID-19 on X (Twitter) Using Naïve Bayes and Support Vector Machine

Abstract

The COVID-19 pandemic has impacted health, economy, and society. Social distancing measures and quarantine policies have restricted economic activities, leading to downturns in COVID-19-affected regions and a subsequent rise in unemployment rates, particularly in urban areas. Concurrently, there has been a remarkable surge in the utilization of the X (Twitter) platform, with Indonesia ranking 6th globally in X (Twitter) users. This study aims to understand the diverse perspectives of society on unemployment and the factors influencing society's views on unemployment through sentiment analysis of X (Twitter) data. By analyzing 576,764 tweets from April 2020 to October 2023, tweets are categorized into positive, neutral, and negative classes. Classification model was built to classify tweet data by implementing TF-IDF for word weighting, and a pair of machine learning algorithms, Naïve Bayes and Support Vector Machine (SVM). Model evaluation yielded the highest accuracy of 81.5% using Naïve Bayes. The classification outcomes highlight prevalent negative perceptions of unemployment among Indonesians, totaling 50.03%. This research contributes to the literature by providing a large-scale analysis of social media data to uncover public sentiment trends and offering insights for policymakers to address unemployment and improve welfare.

Keywords

• unemployment
• tweet
• tf-idf
• naïve bayes
• support vector machine