A Case Retrieval Algorithm Based on Bayesian Estimation

Implementing knowledge engineering in DSS has brought a problem of inefficient reasoning due to huge searching space. The use of case-based reasoning technique has ameliorated the situation. The validity of case retrieval directly influences the effect of case-based reasoning, and the key is to measure the similarity between cases. In this paper, an algorithm for cases similarity computation based on Bayesian Estimation is proposed. First, we designate the priori distribution parameters by the semantic distance-based similarity algorithm, and then calculate the posteriori encountering probability using Bayesian Estimation, thereby, the cases semantic similarity integrating the subjective experience with the objective statistic is acquired. This approach can effectively improve the success rate of case retrieval in the situation of incomplete sample information.