Intuitionistic Fuzzy Similarity-Based Information Measure in the Application of Pattern Recognition and Clustering

IFS is the further extension of ordinary fuzzy set, widely used in distinct applications to tackle uncertainty and fuzzy problems. Generally, similarity/distance measures are significant technique to discriminate between two sets and further they can applied to the problems of pattern recognition and decision-making. Though various similarity measures have already been suggested, still a lot of scope is there because some of them could not satisfy the properties of similarity measures and provide contradictory results. In this paper, we suggested new similarity measures having ability to contrast intuitionistic fuzzy (IF) sets. Furthermore, we have also analyzed their properties to validate the existence of proposed measure. After that, we provided some experimental analysis to verify the effectiveness of proposed measures which includes numerical experiment pattern recognition and clustering analysis. In experimental analysis, firstly we included numerical experiment by taking different cases to study the performance of proposed measures. Then, we dealt with the problems of pattern recognition and incorporated a performance index in terms of "Degree of Confidence" (DOC). Furthermore, we introduced a IF-MST clustering algorithm to deal with IFSs using the notion of MST ("Maximum Spanning Tree"). Additionally, we have modified similarity measure into knowledge measure and contrasted its performance with others knowledge measures to show the superiority of proposed measure.