SHOPSMART: product recommendations through technical specifications and user reviews

This paper describes a new method for providing recommendations tailored to a user's preferences using text mining techniques and online technical specifications of products. We first learn a model that can predict the price of a product given automatically-determined features describing technical specifications and users' opinions. We then use this model to rank a list of products based on individual users' preferences about various features. On a data set collected from Amazon reviews and online technical specifications, rankings produced by this model rank the best product for a user in the 87th percentile of products in its category, on average. Our approach outperforms several comparison systems by 21 percentiles or more.

• Type: Conference
• Year: 2008
• Where: CIKM
• Authors: Alexander Yates, James Joseph, Ana-Maria Popescu, Alexander D. Cohn, Nick Sillick
• Keywords: CIKM 2008, Information Management, Online Technical Specifications, Technical Specifications, Text Mining Techniques