Analisis Perbandingan Algoritma SVM dan CNN dalam Mendeteksi Website Judi Online Berdasarkan Konten Teks

Abstract

This study aims to compare the effectiveness of Support Vector Machine (SVM) and Convolutional Neural Network (CNN) algorithms in detecting Indonesian-language online gambling websites. With the increasing number of online gambling players in Indonesia, it is essential to develop effective methods for identifying gambling content. The dataset used consists of 34,336 gambling websites and 36,529 non-gambling websites, collected through web scraping. The SVM model demonstrated an accuracy of 99%, with evaluation metrics including a precision of 1.00, recall of 0.99, and F1-score of 0.99. In contrast, the CNN model achieved perfect accuracy of 100%, with precision, recall, and F1-score all at 1.00. However, it is important to note that this perfect accuracy was achieved under certain conditions, including a relatively clean dataset and optimal training processes. Evaluation results using cross-validation techniques indicated that SVM maintained a consistent accuracy of approximately 99%, while CNN exhibited an average accuracy of 99.61% with a very low standard deviation. This research emphasizes the importance of data pre-processing in enhancing model accuracy and highlights the advantages of CNN in capturing complex patterns within text. These findings contribute significantly to the development of detection methods for online gambling websites in Indonesia and open avenues for further research in this field.