Regression Analysis Of Fuzzy Granularity Based On Rough Sets

Machine learning and knowledge discovery is one of the most important issues to be addressed in artificial intelligence. And uncertainty and inconsistency are the key problems in knowledge discovery from complex data. Rough set theory, which simulates the capability of granulation and approximation in human cognition, has proven to be an effective mathematical tool to characterize incosistency in classification data. Rough set theory is without any prior knowledge of the case for the data set, and divided into upper and lower approximation, and then to study the knowledge from data set.In recent years, attribute reduction algorithm of decision (classification) problem based on rough set is relatively mature. The theoretical basis is based on attribute importance. But for the attribute reduction algorithm of regression problem with continuous output, the related research is less. Basically, the traditional regression algorithms are based on the known samples to build or simulate regression function. For a new data, put it into the regression function, so we can get the new output. Artificial regression equation for the subjective setting, there are inevitable disadvantages.The paper has two purposes: first, to construct the attribute reduction problem in the regression problem refe to decision making (classification) problem based on rough set theory. The concept of attribute importance of regression problem is gotten, and thus the design of the two calculation methods of attribute important to be used. Combined with actual data he design of the two methods are tested.Second, it analyzes the characteristics of traditional regression analysis and the main steps proposed. Design the ideas of regression analysis based on rough set theory, comparing the difference with the traditional regression analysis. Combined with actual data the regression analysis based on rough set theory has been tested.