Azzahra, Hanifa (2025) Klasifikasi Cyberbullying Flaming Pada Media Sosial X Menggunakan Metode Logistic Regression. Other thesis, Universitas Islam Riau.

Text skripsi_203510534_watermark.pdf - Submitted Version Restricted to Registered users only Download (4MB) | Request a copy

Abstract

The influence of technological developments that are increasingly easy to use, has created many negative effects that have emerged. One of them is cyberbullying, Flaming is one form of cyberbullying. Cyberbullying Flaming is defined as an act in which the perpetrator sends messages containing harsh words, hateful and insulting messages, threats with the aim of provoking or hurting others or even hate speech to an online group or to the victim directly by utilizing technology such as social media, including X. The purpose of this study is to determine the performance and accuracy results of the Logistic Regression method in classifying the type of Cyberbullying Flaming based on trained data. Data was taken from various tweets containing elements of Cyberbullying Flaming. The data used in this study were 2391 Indonesian language tweets collected using Tweet-Harvest. The types of Cyberbullying Flaming consist of Direct and Intentional Flaming, Satirical Flames, Hot Flames, Cold Flames and Neutral. Based on the research results, the success rate of Cyberbullying Flaming classification on social media X using KFold Cross Validation with the Logistic regression method obtained an accuracy result of 82%, precision 80%, recall 81% and f1-score 80%. K

Actions (login required)

View Item