E-Wallet Sentiment Analysis Using Na¨ıve Bayes and Support Vector Machine Algorithm

research
  • 19 Aug
  • 2022

E-Wallet Sentiment Analysis Using Na¨ıve Bayes and Support Vector Machine Algorithm

Nowadays most of consumers in urban areas are accustomed to using digital
wallets. The habit of transaction in cashless has been widely applied to the transportation
system, restaurants and shops in the mall or supermarket. Apart of the ease of conducting
transactions, various promotions in the form of points and cashback offered from various digital
wallet application developers or e-wallets have become very attractive to users. One of the most
widely used e-wallets by the public is OVO and DANA. This phenomena encourages researchers
to do a research and make it as an object of study due to both are widely discussed by various
groups, especially in the capital of Jakarta lately. As it is used, many customers write product
and service reviews based on their experience on the Google Play store. Sentiment analysis is a
technique that can find the right solution in creating a system that can automatically analyse
these reviews and extract information that is most relevant to users. Researchers collected OVO
and DANA review data on the Google Play store with a total of 2000 datasets. In this study,
researchers compared the two algorithms namely Na¨ıve Bayes and Support Vector Machine
(SVM). The stages carried out in this study are data collection, initial data processing, modelling
with the chosen method, experimental & model testing as well as evaluation and validation of
result. Evaluation is carried out using 10 Fold Cross Validation. The result showed that OVO
is the most popular e-wallet application by the public with an accuracy measurement using the
Confusion Matrix reaching 91.00% for the SVM algorithm. The ROC curve showed the best
AUC result of 0.986 (Excellent Classification)

Unduhan

 

REFERENSI

Cari

Bidang Ilmu

Kategori

Kiriman Terbaru