Automatic classification in product catalogs

In this paper, we present the AutoCat system for product classification. AutoCat uses a vector space model, modified to consider product attributes unavailable in traditional document classification. We present a method of feature source weight optimization. We present key features of our user interface, developed to assist users with evaluating and editing the output of the classification algorithm. Finally, we present observations about the use of this technology in the field. KEYWORDS Text classification, categorization, product classification, user interface design.

• Type: Journal
• Year: 2002
• Where: SIGIR
• Authors: Ben Wolin
• Keywords: Information Technology, Product Classification, SIGIR 2002, Traditional Document Classification, User Interface