mahmoudmohamedmahmoud/amazon_review repository overview

⁠Amazon Review Rating Prediction with BERT base Model

This Docker image is designed to perform sentiment analysis on Amazon product reviews using a BERT-based classification model to predict rating. The model is trained on a dataset comprising 500,000 reviews and achieves a training loss of 0.99 and 1 evaluation loss. It has been rigorously tested on separate test data of 600K review, demonstrating an accuracy of 60%.

⁠Usage

To analyze a single text input, run the following command:
docker run -it --rm -v /path/to/local/output:/project/output amazon_review "your_review"

To process reviews from a CSV file, use the following command. Make sure your CSV file has a column named "text" containing reviews:
docker run -it --rm -v /path/to/local/csv:/project/input -v /path/to/local/output:/project/output amazon_review /project/input/your_file.csv

Replace /path/to/local/output with the local directory where you want to store the output, and /path/to/local/csv with the directory containing your CSV file.

This image leverages a powerful BERT-based model to predict sentiment and assign ratings between 1 and 5 to Amazon product reviews. The provided Docker commands ensure seamless interaction with the image, allowing users to effortlessly analyze both individual text inputs and CSV files .