PRODUCT REVIEW SENTIMENT ANALYSIS AT ONLINE STORE JINISO OFFICIAL SHOP USING NAIVE BAYES CLASSIFIER (NBC) METHOD

Efta, Riana Putri and Herny, Februariyanti PRODUCT REVIEW SENTIMENT ANALYSIS AT ONLINE STORE JINISO OFFICIAL SHOP USING NAIVE BAYES CLASSIFIER (NBC) METHOD. Manajemen, Teknologi Informatika dan Komunikasi (Mantik).

PDF Download (2MB)

Abstract

"It is undeniable that online shopping is the choice of many people during the pandemic, because online shopping is an easy and safe solution for people who are required to carry out social distancing by the government. Jiniso Official shop is one of the online stores that currently sell through Shopee. This study was conducted to categorize and analyze customer views of the product by using Jiniso product review data from the comments column on the Shopee application. In this study, researchers will utilize the text mining process using classification techniques, the algorithm that will be used is the Naive Bayes algorithm. Naive Bayes allows classification based on the assumption of separate conditions between the predicted attributes of a particular class. For this reason, the Naive Bayes Classifier is a very competent classification, it works quite well in classification tasks so many researchers are trying to improve the performance of Naive Bayes [3]. The results of sentiment analysis using NBC produces an accuracy rate of 0.941747572815534 or 94%. From the results of this study, positive sentimental reviews can be used as a reference to maintain things that make customers feel satisfied. while negative reviews can be used as motivation to improve services and products."

Item Type:	Article
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science Q Science > QA Mathematics > QA76 Computer software
Faculty / Institution:	Fakultas Teknologi Informasi > Program Studi Sistem Informasi
Depositing User:	Jati Sasongko Wibowo
Date Deposited:	09 Dec 2022 08:47
Last Modified:	09 Dec 2022 08:47
URI:	https://eprints.unisbank.ac.id/id/eprint/8987

Actions (login required)

View Item