Case-Based Reasoning Algorithm Based on Qualitative Causality

Abstract

With the limitation of low predictive performance and poor reliability in current case-based reasoning method, this paper proposes a case-based reasoning algorithm based on qualitative causality (CBR-QC). On the basis of traditional quantitative retrieval method, it introduces relevant knowledge about qualitative reasoning, describes the case qualitatively with qualitative method, constructs qualitative equations, and then gets case groups with requirements through qualitative retrieval. Euclidean distance similarity computation method and KNN method are applicable in this article to retrieve quantitatively similar cases from retrieved qualitatively similar cases, finally get prediction values through adapting the retrieved cases.