AmazonHelpfulReviewClassifier
The classifier system capable of predicting if review will be helpful on Amazon.
Uses text mining to do it.
How to use:
- Download files from here: http://jmcauley.ucsd.edu/data/amazon/
- Run ConvertAndCleanup.py to extract reviews from raw files into csv.
- Import both Metadata and Reviews files into the MS SQL Server database
- Run SQLPart.sql to merge metatdata with reviews and save into csv.
- Run Sentiments.py to add sentiments of reviews into file
- Run LDA.py to add LDA topic probabilities to each review
- Run Classification.py to train classification algorithms which will predict how helpful the review will be