Comparative Study Of Public Sentiment Analysis On Ikn (Nusantara Capital City) Across X And Youtube Using Pso-Optimized Svm

Awaliah Fitri Nur Ananda (a:1:{s:5:"en_US"s:27:"Universitas Negeri Makassar"})

International Journal of Health, Engineering and Technology
Vol. 5 No. 1 (2026): IJHET MAY 2026

This study aims to compare public sentiment regarding the IKN on Platform X and YouTube using a Polynomial Kernel SVM algorithm optimized by PSO. The method used is an experimental study. Data was obtained through web scraping using Google Colab, yielding 1,413 tweets for the X dataset and 814 for the YouTube dataset. The collected data underwent data cleaning, followed by sentiment labeling of each data point into three classes. Following this, TF-IDF vectorization, PSO-based feature selection, and classification using the Polynomial Kernel SVM were performed. The results of the study showed an accuracy of 76% for the PSO-SVM on the X platform and 75% for the YouTube platform. These results indicate that the PSO-SVM algorithm performs better on the X platform compared to the YouTube platform.