Skip to content

This project includes the implementation of a product recommender system based on product reviews and metadata history (product titles) from the Amazon Product co-purchasing Network metadata dataset.

License

Notifications You must be signed in to change notification settings

ramapinnimty/CS5664-Amazon-Product-Recommendation

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Amazon Product Recommendation using Network Analysis and Machine Learning

CS5664_TextPreProcessing_MetaData.ipynb contains:

  1. Parsing the metadata and produce essential pandas dataframes.
  2. Creating product-copurchase edge lists.

CS5664_NetworkAnalysis.ipynb contains:

  1. Network information such as degree centrality analysis.
  2. Node degree distribution.
  3. Powerlaw and heavy tail distribution.

CS5664_MetaData_and_ProductReviews_Analysis_EGO_Graph_Recommendations.ipynb contains:

  1. Text preprocessing.
  2. Meta data and Product Review processing.
  3. Sentiment Analysis.
  4. Topic Modeling.
  5. EgoGraph based Product Recommendations using product Titles.

CS5664_Product_Recommendations_MachineLearning.ipynb contains:

  1. Review rating analysis.
  2. KNN based SVD experimentation.
  3. Hyperparameter tuning and training SVD with best params.
  4. Product recommendations based on reviews.

Datasets:

  • Appliances.json - Reviews information for electronic appliances.
  • Magazine_Subscriptions.json - Reviews information for magazines.
  • amazon-books.csv - information on products (books) generated from amazon-meta.txt.
  • amazon-books-copurchase.edgelist - information generated from amazon-meta.txt contains purchase-copurchase similar edgelist.
  • amazon-meta.txt - Data downloaded from SNAP.
  • products_copurchases_links.csv - Purchase Copurchase list for all products generated from amazon-meta.txt used for network analysis.
  • products_data.csv - contains all product information.

Supplementary Materials & Visualizations folders include:

  1. Network Analysis plots.
  2. WordClouds - Reviews, Sentiments, Topics.
  3. Sentiment Analysis plots.
  4. LDA topics.

About

This project includes the implementation of a product recommender system based on product reviews and metadata history (product titles) from the Amazon Product co-purchasing Network metadata dataset.

Resources

License

Stars

Watchers

Forks