/Product-Catalog-Size-Recommendation-Framework

Framework based on latent factor model and metric learning technique to predict fit of different catalog sizes of clothing products so as to recommend appropriately sized products to customers.

Primary LanguageJupyter Notebook

Product-Catalog-Size-Recommendation-Framework

This framework is based on the algorithm proposed in the paper Decomposing Fit Semantics for Product Size Recommendation in Metric Spaces published at RecSys 2018.

It consists of collaborative filtering and metric learning components for predicting the fit of different catalog sizes of clothing products so as to recommend appropriately sized products to customers.

We also contributed two clothing fit datasets from Modcloth and Renttherunway retail websites. Details of the datasets can be found here.

Prerequisites

  • Install PyLMNN library from here.
  • Download the datasets from here.

Publication

If you use this code or the datasets, please cite the following:

Decomposing Fit Semantics for Product Size Recommendation in Metric Spaces,
by Rishabh Misra, Mengting Wan, Julian McAuley, 
in Proc. of 2018 ACM Conference on Recommender Systems (RecSys’18), Vancouver, Canada, Oct. 2018

Citation format available here: