Automatic search of information about cancer in the web

Abstract

Due to the dramatic increase on available information in the WEB, users are in the need of new tools to find, filtrate and analyze the huge amount of data returned by search engines. In this paper we present a new tool for WEB searching which starts from a small set of examples labeled by the user. From them a process of selection and weighting of words to form new queries ensures the recovery of more relevant documents. The results obtained on a search for cancer related documents illustrate the effectiveness of the method. © Springer-Verlag Berlin Heidelberg 2007.