Document Categorization Based on Usage of Features Reduction with Synonyms Clustering in Weak Semantic Map

摘要

Nowadays the number of huge companies and corporations has in their disposition various non-structured texts, documents and other data, but most of this data is still just text documents with different subject matters and content. The work-flow organization on this data format is complicated because of their characteristics, and requires modern tools for processing and analysis. Possible problem solution is machine learning algorithms and natural language processing methods envolving, with existing clustering and classification algorithms improvement. For document classification, we propose a proprietary approach based on the us-age of a semantic map as a feature reduction tool. In this paper we are going to investigate the impact of this approach on the quality of classification of documents and describe its application to the implementation of the document categorization.