The algorithm collects reviews of a user-specified restaurant or café and assigns a numerical rating (1 to 5 stars) based on the sentiment of the user reviews. A website's reviews are scraped and turned into numeric tokens for the machine learning model. Then, the scraped checks are subjected to sentiment analysis using Bidirectional Encoder Representations from Transformers (BERT). The rating derived from the extracted sentiment of the user review is then compared to the rating the user provided on the website. PyTorch, BeautifulSoup, and Hugging Face comprised this project's technology stack. This project demonstrates how the user's views expressed in the form of a review encode the rating that the user will ultimately assign to a service.
vikramjitsinghrathee / bert_nlp-sentiment_analysis_website_scraped_reviews Goto Github PK
View Code? Open in Web Editor NEWUtilize Bidirectional Encoder Representations from Transformers (BERT), to perform sentiment analysis on the reviews scraped from a website.
License: MIT License