Graded Mean Integration Representation and Intuitionistic Fuzzy Weighted Arithmetic Mean for Similarity Measures in Case-Based Reasoning.

An essential component of case-based reasoning (CbR) is the similarity function. The existence of incorrect similarity values, one of the primary issues of similarity function, occurs when two cases seem to have a lot in similarity but actually differ greatly from one another. The aim of this study is to suggest a fuzzy similarity function that might be used with CbR, which has been executed in three primary phases. Firstly, a knowledge base is created. Secondly, the inference engines are created. To determine fuzzy similarity values, the proposed approach employs Graded Mean Integration Representation (GMIR), which takes into consideration the impact of each evidence on the case. Finally, model testing is determined by comparing GMIR with the additive function and additive weighting and using them as the sensitivity test. A comparison of similarity methods was carried out using the k-NN, Euclidean, and Pearson similarity. Expert evaluation carries out external validity checking. The degree of similarity between the CbR output solution and the expert solution is determined using the Intuitive Fuzzy Weighted Arithmetic Mean (IFWAM). The results demonstrate that the GMIR function applied with thresholds of 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, and 0.8 provides the fewest instances of similar cases when compared to the additive and additive weighting approaches. GMIR is able to provide a variety of alternatives using the threshold value. IFWAM performs 93.33% when used in 15 test scenarios, and only one of the solutions from the CbR output differs from that offered by the pshycologs. The main reason is the challenge of selecting keywords for every item of evidence provided. When CbR's validity was examined, it was also demonstrated that IFWAM was reliable in estimating the degree of similarity between CbR output and expert solutions. [ABSTRACT FROM AUTHOR]