Aspect Based Sentiment Analysis Survey Kepuasan Pelanggan terhadap Perusahaan Penyedia Jasa Solusi Perangkat Lunak Menggunakan Multinomial Naïve Bayes

Abstract

A software developer company is a company that provides services in the form of software technology development to assist business people in achieving their business goals. In order for system maintenance services to be maximized, it is necessary to carry out an aspect-based sentiment analysis so that the company can find out what needs to be improved and developed so that in the future customers will continue to subscribe and develop the system in the future. This study aims to conduct an aspect-based sentiment analysis from a customer satisfaction survey of software solution service providers using the Multinomial Naïve Bayes method. This study used 1300 data in the form of user reviews. The data that has been collected will later be cleaned through 7 stages of preprocessing namely cleaning, case folding, punctual removal, normalization, stopword removal, stemming, and tokenization. Next, feature extraction will be carried out using TF-IDF for the word weighting process. Then, the data will be classified based on aspect-based sentiment using Multinomial Naive Bayes. Evaluation results are presented through a confusion matrix and get an average accuracy based on the four product aspects of 88.75%. From the accuracy that has been obtained, it can be said that the system is good enough at predicting reviews based on aspects.