An Intuitionistic Fuzzy Set Approach for Multi-attribute Information Classification and Decision-Making

Abstract

This article introduced a new multi-attribute information classification method by employing intuitionistic fuzzy set (IFS) approach. The proposed method was referred as four-way intuitionistic decision space (4WIDS). In the 4WIDS, IFS theory was used to model the inherent uncertainty of multi-attribute information. For generating more precise level of decision-rules, granular computing (GrC) approach was employed. The proposed 4WIDS method was appropriate for the classification of the multi-attribute information into four different regions as positive IFS, negative IFS, uncertain IFS and gray IFS regions. Detail methodology of the 4WIDS was explained by presenting its representation in a precise way. This study also presented various definitions, properties and theorems in the support of the 4WIDS method. The 4WIDS was applied in benchmark datasets that included Pima Indians diabetes, Thyroid disease, Fisher’s Iris and Spambase datasets. Experimental results including statistical analysis indicated that the proposed 4WIDS outperformed existing classification methods, such as Naive Bayes, Decision tree, PART, J48, logistic model trees (LMT), rough set (RS), gray multi-granulation rough set (GMGRS) and multi-granulation fuzzy rough set (MGFRS).