AUTOMATING PRODUCT CLASSIFICATION: THE IMPACT OF STANDARDIZATION
Author(s):
Joerg Leukel
Paper abstract: Approaches to automating product classification adopt mainly algorithms and tools for classification in general, such as
Vector Space Model, Bayesian Classification and K-Nearest Neighbor Classification, or take classification systems as
(database) schemas that have to be integrated. Much work has been carried out on applying, evaluating and improving
these tools for this specific domain, product data in B2B e-commerce. The formal specification of a product classification
system (PCS) is seen as input data; hence this data has to be imported and often converted into an internal representation.
This paper takes a closer look at this input data and examines its structure, semantic richness and degree of
standardization. The reason is that standards for product classification systems and standardized specifications address
the data exchange issues of these systems – and more important, are able to increase the success of automated product
classification. To do so the paper will derive and analyze major standardization trends, show their interdependencies and
evaluate their impact on automated classification.
Keywords:
B2B, E-Commerce, E-Procurement, ISO 13584, Standardization, XML.
Type:
Full Paper
Full Contents (click to dowload):
First Page: 147
Last Page: 154
Year:
2004
Editors:
Pedro Isaías, Piet Kommers and Maggie McPherson
ISBN:
972-98947-5-2
Language:
English
Conference Name:
IADIS International Conference e-Society 2004
