Implementation of Particle Swarm Optimization on Sentiment Analysis of Cyberbullying using Random Forest
Social media has exerted a significant influence on the lives of the majority of individuals in the contemporary era. It not only enables communication among people within specific environments but also facilitates user connectivity in the virtual realm. Instagram is a social media platform that pla...
Gespeichert in:
Veröffentlicht in: | Juita : jurnal informatika (Online) 2023-11, Vol.11 (2), p.301-309 |
---|---|
Hauptverfasser: | , , , , |
Format: | Artikel |
Sprache: | eng ; ind |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | Social media has exerted a significant influence on the lives of the majority of individuals in the contemporary era. It not only enables communication among people within specific environments but also facilitates user connectivity in the virtual realm. Instagram is a social media platform that plays a pivotal role in the sharing of information and fostering communication among its users through the medium of photos and videos, which can be commented on by other users. The utilization of Instagram is consistently growing each year, thereby potentially yielding both positive and negative consequences. One prevalent negative consequence that frequently arises is cyberbullying. Conducting sentiment analysis on cyberbullying data can provide insights into the effectiveness of the employed methodology. This research was conducted as an experimental research, aiming to compare the performance of Random Forest and Random Forest after applying the Particle Swarm Optimization feature selection technique on three distinct data split compositions, namely 70:30, 80:20, and 90:10. The evaluation results indicate that the highest accuracy scores were achieved in the 90:10 data split configuration. Specifically, the Random Forest model yielded an accuracy of 87.50%, while the Random Forest model, after undergoing feature selection using the Particle Swarm Optimization algorithm, achieved an accuracy of 92.19%. Therefore, the implementation of Particle Swarm Optimization as a feature selection technique demonstrates the potential to enhance the accuracy of the Random Forest method. |
---|---|
ISSN: | 2086-9398 2579-8901 |
DOI: | 10.30595/juita.v11i2.17920 |