Analisis Sentimen Masyarakat Indonesia terhadap Pemindahan Ibu Kota Negara Indonesia pada Twitter

Abstract

The relocation state capital of Indonesia raises various responses, especially from the Indonesian people. The discussion related to these issues is very interesting to study, how are the positive and negative sentiments of the Indonesian towards the government's decision. This study aims to analyze the sentiments of the Indonesian people regarding the relocation state capital of Indonesia, including the chosen name of Nusantara on Twitter. In this study, a comparison of 3 algorithms is used, namely the Support Vector Machine (SVM), Naïve Bayes, and K-Nearest Neighbor (KNN) algorithms. From this study, the results obtained are 1,141 positive comments, while negative sentiments are 591 comments. This shows that the Indonesian people have a positive opinion towards the new capital city of Indonesia. In the classification and model testing phase, 10-fold cross validation is used. From these tests, the SVM algorithm obtained an accuracy value of 85.71%, the Naïve Bayes algorithm obtained an accuracy value of 76.70%, the KNN algorithm obtained an accuracy value of 52.74%. This study shows that the SVM algorithm can work better than the Naïve Bayes algorithm and KNN. The accuracy value for the KNN algorithm obtains a low value, this is because the KNN algorithm is sensitive to features that are less relevant.

Hoax News Analysis for the Indonesian National Capital Relocation Public Policy with the Support Vector Machine and Random Forest Algorithms

“…The results showed that the SVM (95.24%) algorithm is better than RF (86.90%). These results corroborate the research of Lestari et al (2022), that the accuracy of SVM (85.71%) is better than NB (76.70%) and KNN (52.74%) [8]. Also research by Lestari, et al (2022) on SVM with Chi-Square Feature selection shows that SVM is a very good algorithm with an accuracy value of 90%, an averaecision value of 90%, 86% recall, and an f1-score of 88% [10].…”

“…Various studies have explored public opinion on the policy of relocating the Indonesian state capital. For instance, Lestari et al (2022) [8] used Support Vector Machine (SVM) and K-Nearest Neighbor (KNN) algorithms with 10-fold crossvalidation and achieved 1,141 positive and 591 negative sentiments. Their results indicated that SVM (accuracy = 85.71%) outperformed Naïve Bayes (NB) (76.7%) and KNN (52.74%).…”

“…Arsi and Waluyo (2021) [9] analyzed 1,236 tweets using SVM and reported very good accuracy (96.68%). Similarly, Lestari (2022) [10] used SVM with Chi-square feature selection and achieved an accuracy and precision of 90%. Other studies have also explored sentiment analysis with different algorithms, including NB, Logistic Regression, and Neighbor Weighted K-Nearest Neighbor (NWKNN).…”

“…This study seeks to address this gap by comparing the performance of SVM and Random Forest (RF) algorithms in hoax analysis on Twitter. SVM and RF were chosen due to their excellent performance in text mining [8], [9], [11], [18]. The study's contribution is twofold: first, it scientifically evaluates the performance of SVM and RF for hoax analysis, and second, it offers practical recommendations on government policy regarding the capital relocation.…”

Analisis Sentimen Relokasi Ibukota Nusantara Menggunakan Algoritma Naïve Bayes dan KNN

“…Dengan menggunakan metode NB, Mendapatkan nilai akurasi sebesar 63.21% sedangkan nilai KNN sebesar 58.10% yang berarti metode NB paling akurat dibandingkan metode KNN [12]. Penelitian tentang Sistem Analisis Sentimen pada ulasan produk dengan menggunakan total 1500 data dari femaledaily.com dan nilai akhir sebesar 77.78%, data tersebut kasifikasi menggunakan Naive Bayes [13], [14].…”