A web-based application that will be able to scrape online reviews and make accurate predictions using machine learning models.
- User can run the sentiment analysis on a product from the given choices or he/she can feed
in the amazon URL of a new product.
- Once the product is selected then the reviews are scraped from amazon.
- This scraped data is preprocessed. Preprocessing involves handling missing data, removing
punctuations, removing stop words, tokenization and lemmatization.
- Next the features are extracted using TF-IDF vectorizer.
- Then 3 classification algorithms namely Support Vector Machine, Logistic Regression and
Random Forest are applied on extracted features.
- These algorithms predict the sentiment, 1 for positive and 0 for negative and send this data
to the frontend.
- Visualizations including overall sentiment and sentiment over time are displayed to the
user.