Application of Support Vector Machine and Naive Bayes Method to Analyze Mobile Banking User Behavior Through social media Twitter.

Abstract

The purpose of this research is to use Naive Bayes and Support Vector Machines (SVM) to determine sentiment analysis on social media related mobile banking. The author used a quantitative technique in this study. Data crawling and a review of the literature were two of the methods used in the collection process. Author using SEMMA technique. The process of defining and articulating the problem leads to this research phase. RapidMiner software is used in this study at every step of the data processing process while using the SEMMA approach. 580 positive and 650 negative feelings were produced by the Naïve Bayes classification approach, whereas 720 positive and 410 negative sentiments were produced by the Support Vector Machine (SVM) classification method. The Naïve Bayes approach yields 90% accuracy with a 1% margin of error, 91% precision for positive predictions, 89% precision for negative predictions, 90% recall for positive data, and 90% recall for negative data. An accuracy value of 89% with a margin of error of 3%, positive predictive precision of 67%, negative predictive precision of 99%, positive data recall of 95%, and negative data recall of 88% are obtained by the Support Vector Machine (SVM) approach. The data used is from the Naïve Bayes approach since it has a higher accuracy value of 90% than the Support Vector Machine (SVM) method, based on the accuracy, precision, and recall values of the two classification methods. Social media users tend to have more positive perceptions regarding mobile banking since positive sentiment outweighs negative sentiment.