Analisis Sentimen Aplikasi Inarisk Personal di Google Play Store Menggunakan Algoritma NLP

Abstract

From a geographical perspective, Indonesia is recognized as a country that is prone to natural disasters, BNPB as one of the disaster management institutions in Indonesia has the Inarisk application as an application that provides information about disasters. To improve the effectiveness of services to users, sentiment analysis is carried out to find out user reviews on the InaRisk application. The purpose of this study is to explore user opinions on the InaRisk application on the Google Play Store platform. Researchers used a total of 597 reviews as data to be analyzed to determine polarity using NLP models. The results showed that as many as 34.7% of reviews had a positive sentiment, 52.6% of reviews had a neutral sentiment, and 12.7% of reviews had a negative sentiment. This shows that even though the application is good enough, it still needs some improvements in the application. In classifying predictions carried out using 6% of data as test data and 94% as training data where regression models were used and succeeded in producing an accuracy of 69.81%, this indicates that the logistic regression model used to classify predictions is good enough so that it can be used for sentiment analysis of other studies.