Linguistic Summaries Generation with Hybridization Method Based on Rough and Fuzzy Sets

摘要

In this paper authors propose a new algorithm for linguistic data summarization based on hybridization of rough sets and fuzzy sets techniques. The new algorithm applies rough sets theory for feature selection in early stages of linguistic summaries' generation. The rough sets theory was used to reduce on significant way, the amount on summaries obtained by others algorithms. The algorithm combines lower approximation, k grade dependency and fuzzy sets to get linguistic summaries. The results of proposed algorithm are compared with association rules approach. In order to validate the algorithm proposed, authors apply both qualitative and quantitative methods. Authors used two databases in order to validate the algorithm; theses databases belong to "Repository of Project Management Research". The first database is associated to personality traits and human performance in software projects. The second database is associated to analysis of revenue assurance in different organization. Considering quantitative approach, the algorithm proposed, obtains better results than the algorithm based on association rules; while regards execution time, the best algorithm was the algorithm based on association rules, because rough sets theory was high time-consuming technique.