Heuristic Attribute Reduction Research On Interval-valued Intuitionistic Fuzzy Rough Set

Fuzzy sets (Fs) and Rough sets (Rs) theory are two different ways of processing information; both theories have their own advantages in terms of attribute reduction. Attribute reduction is the core problem of data mining domain and the key technique for any sector decision-making knowledge acquisition. Once a breakthrough is achieved in the algorithms research, it Omit can improve the enterprises and companies’decision efficiency greatly. At the same time, the knowledge we need can be achieved faster and more comprehensive. This paper has Omit improved attribute reduction algorithm based on intuitionistic fuzzy rough sets (IFRs). and a heuristic attribute reduction algorithm of interval-valued intuitionistic fuzzy rough sets (IVIFRs) is presented. The new algorithm has more advantages than the original algorithm whether in time complexity or in space complexity. The main contents are as follows:(1)Theoretical knowledge of rough set and fuzzy set are introduced accordingly; the advantages and disadvantages of the two theories are presented. It points IVIFRs can not only maintain data integrity of the original information systems, but also have more feature selection than a single theory.(2)On the basis of attribute reduction algorithm which is focused on the original intuitionist fuzzy rough set, the paper puts forward the interval-valued intuitionist fuzzy rough set attribute reduction algorithm which is on account of mutual information. In terms of keeping the ability of classifying original information constant, we try to make a two-way return by taking dual metrics as standard——attribute importance and attribute dependence, and get the decision tables. Moreover, the experimental analysis proves that attribute reduction algorithm based on mutual information has an advantage. The corresponding experimental analysis shows the more advantages of attribute reduction algorithm based on mutual information.(3)Attribute reduction theory based on a heuristic interval intuitionistic fuzzy rough set is further proposed, which can delicate original data more exquisitely, accomplish the feature selection more effectively, and retain more knowledge from original data in real applications. Accomplish the feature selection more effectively, and retain more knowledge from original data in real applications. Define dependence and non-dependence intervals of IVIFRs which combine into relatively positive domain, and determine the importance by decision interval-valued intuitionistic fuzzy numbers, then complete the attribute reduction. Finally, the example analysis shows the effectiveness of the method.