Habibi, Thoriq Rifki (2025) Analisis Sentimen Publik di Media Sosial Terhadap Pemotongan Anggaran Pendidikan Menggunakan Algoritma SVM dan Logistic Regression. Other thesis, Universitas Islam Riau.

Text skripsi_203510591_watermark.pdf - Submitted Version Restricted to Registered users only Download (4MB) | Request a copy

Abstract

The Indonesian government's education budget cuts have sparked diverse responses from the public, particularly through social media. This study aims to analyze public sentiment toward this policy using data from two social media platforms: Twitter and YouTube. The sentiment analysis in this study classified data into three categories: positive, negative, and neutral. Two machine learning algorithms were used to classify sentiment: Support Vector Machine (SVM) and Logistic Regression. The research process began with data collection (crawling), followed by a text preprocessing stage that included data cleaning, tokenization, stopword removal, normalization, and stemming. Labeling is done manually (human) by researchers based on understanding the context and meaning of each text. Next, word weighting was performed using the TF-IDF method and data balancing with the SMOTE technique. Evaluation results showed that both algorithms performed quite well in sentiment classification. On the X (Twitter) data, SVM achieved an accuracy of 62.60% and Logistic Regression 64.17%. Meanwhile, on the YouTube data, SVM achieved an accuracy of 92.86% and Logistic Regression 92.50%.

Actions (login required)

View Item