This is a joint research project on text categorization, based on the idea of synergy between a modern and powerful tool for search and optimization - Genetic Algorithms - and various algorithms for clustering and classification applied to text categorization.
Text categorization is a central issue in modern information management. Methods for automatic text categorization have been proposed (LSI and k-NN or LSI and centroid method), but their use for large amounts of data is prohibitive due to their complexity.
Genetic Algorithms have become lately a well-known tool for search, optimization and machine learning. By means of simulating natural evolution, such algorithms achieve high accuracy and efficiency in solving very difficult optimization problems.
Hybridizing the focus of text categorization algorithms with the power of Genetic Algorithms is likely to provide very efficient tools for solving this problem. To date, few (if any) attempts to perform such an interdisciplinary approach have been reported. The joint team of Greek and Romanian scholars brings together the competence of the Greek partner in text categorization and the expertise of the Romanian team in Genetic Algorithms.