Experiments on statistical approaches to compensate for limited linguistic resources

Abstract

Information Retrieval systems can benefit from advanced linguistic resources when carrying out tasks such as word-stemming or query translation. The main goal of our experiments has been the development of methodologies that minimize the human labor needed for creating linguistic resources for new languages. For this purpose, we have applied statistical techniques to extract information directly from the collections. © Springer-Verlag Berlin Heidelberg 2005.