Penggunaan Seleksi Fitur Query Expansion Ranking dan Genetic Algorithm-Support Vector Machine untuk Analisis Sentimen pada Aplikasi Perbankan Jenius

Haksar, Haksar
Siswanto, Siswanto
Ilyas, Nirwan

ESTIMASI: Journal of Statistics and Its Application
Vol. 7, No. 1, Januari, 2026 : Estimasi

Jenius is a digital banking product from BTPN launched in 2016. In 2021, various opinions emerged on Twitter regarding cases of lost customer funds, necessitating sentiment analysis to understand public perception. This study used the Support Vector Machine (SVM) method with two feature selection approaches: Query Expansion Ranking (QER) and Genetic Algorithm (GA). The data used were 2,008 manually labeled tweets. The results showed that the QER-SVM method produced an accuracy of 87.81%, while the GA-SVM achieved an accuracy of 88.31% with improvements in precision and F-measure. Thus, the combination of Genetic Algorithm and SVM was more effective in classifying sentiment towards the Jenius application on Twitter.