Gojek and Grab User Sentiment Analysis on Google Play Using Naive Bayes Algorithm And Support Vector Machine Based Smote Technique

Abstract

As the online Ojek services, people often talk about them by giving their opinions and opinions through various media, one of which is Google Play opinion given by the public to the services of online Ojek also diverse. Users provide review reviews or comments about the application, of course users will choose an app that has a good review. But monitoring the reviews of the general public is not easy, because the amount is very much to be processed so that researchers want to know the extent of the user review analysis of Gojek and Grab applications based on the review of user comments using the classification technique is using the NB algorithm and SVM based technique Smote. The results of the test with the highest accuracy result 81.09% and AUC value = 0.922 is the application Gojek while for application test results grab accuracy value of 73.20% and AUC value = 0.848. To that end, the implementation of the Support Vector Machine based Smote technique in this study has higher accuracy so that it can be used to provide solution to the sentiment analysis problems in the review user comments online Ojek application

Factor analysis influencing Mobile JKN user experience using sentiment analysis

“…The application where NB algorithms are frequently utilized is sentiment analysis [28]. The NB classifier uses probability theory to identify the highest probability of classification by examining the occurrence of each category in the training dataset [17], [30].…”

“…SVM is a machine learning algorithm that utilizes the principle of structural risk minimization (SRM) to identify the optimal hyperplane that distinguishes between the two classes in the input space. In short, SVM looks for the best hyperplane that functions as a separator for two classes of data [30]- [32].…”

“…SMOTE can optimize the performance of the classification model algorithm when performing Sentiment analysis [19], [20]. This approach creates new minority class samples by combining neighboring models in a convex manner to balance the dataset [21], [30].…”

Enhancing Sentiment Analysis Accuracy in Delivery Services Through Synthetic Minority Oversampling Technique

“…For the Gojek application, the highest accuracy result was 81.09%, and the AUC value = 0.922. Also, for the Grab application, the accuracy value of 73.20% and AUC value = 0.848 [11].…”

Twitter Sentiment Analysis of Kanjuruhan Disaster using Word2Vec and Support Vector Machine

“…Various methods have been used to analyze sentiment that obtained from social media. Research by Hermanto, Kuntoro, et al [2] about sentiment on Google Play Store review using Naïve Bayes and Support Vector Machine (SVM) with SMOTE oversampling has concluded that SVM's performance is higher than Naïve-Bayes.…”