A recommendation system using several techniques. For this reccomendation engine, we used a Amazon gourmet foods reviews dataset that was obtained from the Stanford Network Analysis Project (SNAP) database. The dataset contained over 35 million from users that have been collected over a period of 18 years. For this specific report, we used gourmet and fine food reviews on Amazon to build a product recommendation system. In this report, we discuss the different approaches we used to extract information from the data. We built an algorithm that performed collaborative filtering, topic modeling and sentiment analysis on 200k records of the dataset to produce a list of recommendations for a given user. From this analysis, we found out what most of the reviews were categorized as such as a positive, negative or neutral review.
This project was primarily implmented in the Jupyter Lab enviornment.
Python Libraries: Scikit-Learn and NLTK
The clean and raw data is available on GoogleDrive.
Clone the project
git clone https://github.com/Suchith3004/food-recommendations.git
Go to the project directory
cd food-reccomendations
- Suchith Suddala - @suchith3004
- Asha Cheruvu - [@ashac29]
- Uditi Goyal - [@ud2607]