Analisis Sentimen Pembangunan IKN (Ibu Kota Nusantara) Pada Twitter Menggunakan Metode K- Nearest Neighbor, Naive Bayes Dan Support Vector Machines

Abstract

This research investigates the Nusantara Capital City (IKN) relocation, which has generated diverse opinions, including concerns over the chosen location and the swift ratification of related laws. Recently, the Indonesian government has called on the public to support IKN's development. To assess public sentiment regarding this relocation, sentiment analysis was performed on a dataset of tweets. After data cleaning, 502 tweets were analyzed, yielding 337 positive and 163 negative comments. The analysis utilized Support Vector Machine (SVM), Naive Bayes (NB), and K-Nearest Neighbor (K-NN) algorithms, incorporating feature selection through Particle Swarm Optimization (PSO). This study compares the performance of Naive Bayes, SVM, and K-NN without feature selection against those methods with feature selection, specifically analyzing their Area Under Curve (AUC) values to identify the most effective algorithm. The results indicate that the PSO-based SVM algorithm achieved the highest performance, with an accuracy of 97.63% and an AUC of 0.997. This research successfully identifies an optimal algorithm for classifying positive and negative comments regarding the relocation of the Nusantara Capital City, contributing valuable insights to public sentiment analysis in this context.