Abstract
With the advancement of web technology, huge information about products is available online in B2B Market. Quite often the available product data is only of short description. This short description product data is needed to be classi.ed to speci.c categories of a desired schema, before it can be used. Most of the classi.ers available presently require training and do not provide expected results when the data is of small description. Also, almost none of these consider the variation in the importance across the terms (words/phrases) along the input (sentences). For classification of such data, our decreasing-saw-tooth-priority based data classifier takes care of such variation. It makes use of information provided in schema only, hence does not require any training. Our classifier has been applied on such type of real life data taken from various domains. Results shown in the paper underscores the superiority of our classifier in terms of its accuracy as compared to others.