Fuzzy Rough Data Model: A New Technique for Analyzing Data

A new technique for analyzing data,fuzzy rough data model,FRDM,is proposed.By means of the dynamic adaptive fuzzy clustering techniques,the approach turn the grid hard partition of input data space in Kowalczyk's rough data model(KRDM) to the fuzzy partition,and(identify) the fuzzy pattern clusters of input data space.Then,the FRDM is built through utilizing the definition of type mapping relation ftype:C_(i)→y from each fuzzy pattern clusters to the decision categories as well as the concept DoF(x),which is the degree of fulfillment of an input data relative to the classification rules for the pattern clusters.Finally,different experimental databases are calculated and the results demonstrate that above approach has better generalization ability,more powerful ability to handle data contaminated by noise and higher searching efficiency compared with the Kowalczyk's RDM.