Product Review System With BERT for Sentiment Analysis and Implementation of Administrative Privileges on Node-RED

Abstract

There are many ways to collect the product review, either by using a website or by manually asking the customers to leave their comments after buying the product. The traditional product feedback process is time-consuming. Nowadays, although industries use online review system to rate their product, the process lacks modernization. In this paper, BERT sentiment analysis is used to process the product review from the customers and automatically convert the review into a numeric score. This average score value is used for processing the sentiment analysis. The system is built entirely on Node-RED as a backbone to gather the review from the users with certain administrator privileges. An interactive dashboard is created to analyze the feedback of the product review. The above method of review system also reduces malpractice that the manufacturer or service provider may indulge into increase their ratings.

Keywords

• BERT model
• natural language processing
• sentiment analysis
• Node-RED