Digital Forensic Investigates Sexual Harassment on Telegram using Naïve Bayes

Abstract

The widespread use of Telegram in Indonesia has had both positive and negative effects. While the app offers strong security features, it has also become a platform for digital crimes, including sexual harassment. This study aims to address the need for effective forensic analysis and classification methods by employing the National Institute of Justice (NIJ) methodology and Naïve Bayes classifier to analyze conversations on Telegram. The research evaluates the performance of digital forensic tools and the effectiveness of the Naïve Bayes method in identifying instances of sexual harassment conversation. The data analyzed is about conversations on the telegram application that contain sexual harassment. Data collection involves extracting relevant conversations and subjecting them to forensic analysis using MOBIL edit Forensic Express and FTK Imager. Based on the test results, the naïve Bayes algorithm can be used to classify conversations into positive and negative about sexual harassment. The value obtained from naïve Bayes testing is the accuracy value of 91.3%, Precision 100%, and Recall 90%.