A new strategy for case-based reasoning retrieval using classification based on association

This paper proposes a novel strategy, Case-Based Reasoning Using Association Rules (CBRAR) to improve the performance of the Similarity base Retrieval SBR, classed frequent pattern trees FP-CAR algorithm, in order to disambiguate wrongly retrieved cases in Case-Based Reasoning (CBR). CBRAR use class as-sociation rules (CARs) to generate an optimum FP-tree which holds a value of each node. The possible advantage offered is that more efficient results can be gained when SBR returns uncertain answers. We compare the CBR Query as a pattern with FP-CAR patterns to identify the longest length of the voted class. If the patterns are matched, the proposed strategy can select not just the most similar case but the correct one. Our experimental evaluation on real data from the UCI repository indicates that the proposed CBRAR is a better approach when com-pared to the accuracy of the CBR systems used in our experiments.