Classification of Indonesian University Entrance Tweets Using Machine Learning

Abstract

Entrance selection for State Universities (PTN) in Indonesia, specifically SNBP, SNBT, and Independent Selection (Mandiri), is a trending topic on social media, particularly Twitter. However, the high volume of tweets creates noise, making it difficult for prospective students to find relevant information. This study aims to classify tweets into three categories (SNBP, SNBT, Mandiri) and compare the performance of four machine learning models: Naïve Bayes, Support Vector Machine (SVM), Artificial Neural Network (ANN), and Long Short-Term Memory (LSTM). The data consists of 12,442 tweets collected using specific keywords. The methodology involves preprocessing (cleaning, normalization, stemming), feature extraction using TF-IDF, and model evaluation. The results show that SVM achieved the best performance with an accuracy of 89.41% and an F1-Score of 89.44%, outperforming the deep learning models (ANN and LSTM) for this specific dataset. These findings indicate that traditional machine learning models can be more effective for text classification with moderate dataset sizes compared to complex deep learning architectures.