Knowledge-oriented methodologies for causal inference relations using fuzzy cognitive maps: A systematic review

The latest research in the field of casual relations using fuzzy cognitive maps utilizes various adaptive supervised learning techniques. These supervised methods provide promising data understanding and processing the fuzzy outcomes. This article contributes in the modern research by providing an adaptive review of different knowledge-oriented methods for the establishment of Fuzzy Cognitive Maps (FCM) based Causal Inference Relations exploiting various domains. A comprehensive review is presented in this article that significantly introduces the artificial intelligence-based methodologies and their progress in the field of casual inference relations. An ensembled-FCM based approach is presented in this article which is analysed and compared to the conventional Hebbian based, error-driven and hybrid approaches. The performance parameters evaluated for the testing phase of E-FCM provides the precision, recall and accuracy rates of 98.05 %, 97.83 % and 94.54 % respectively, while prediction errors observed MSE value of 0.034 +/- 0.005 with relevant reduction in RSME and MSE of 0.052 +/- 0.007 and 0.085 +/- 0.009 respectively. Further, a systematic categorization of existing techniques is presented based on the taxonomy of different techniques from varying perspectives like its extension and application domains. Finally, an outline of the potential research directions in this field is presented, establishing a clear understanding of development in this domain.

Automated summary

This article contributes to the field of casual relations by providing an adaptive review of different knowledge-oriented methods for the establishment of Fuzzy Cognitive Maps based Causal Inference Relations. It introduces artificial intelligence-based methodologies and compares their performance. The article also presents an ensembled-FCM based approach and evaluates its performance parameters.