A Kind of Case Similarity Evaluation Model Based on Case-Based Reasoning

Reasonable evaluation of case similarity is one of the key technologies based on Case-Based Reasoning (CBR). In order to solve some defects of case similarity model, such as have not fully considered the diversity of attribute type, lack of commonality and so on, this paper presents a new evaluation model of case similarity based on CBR. In the model, the similarity of symbolic attributes, numeric attributes and interval attributes are designed in detail, the weight of each attribute is evaluated by G1 method, and the case similarity is measured by combination of global similarity and attributes similarity. Application of the evaluation model illustrates that it has good versatility and extensibility.