支持重用的层次智能CBR检索模型

检索是ＣＢＲ中的关键技术，直接影响ＣＢＲ的推理效率和质量，检索出事例的质量好 坏决定着事例重用与修改的难易。注意到事例的检索与事例的表示和存储密切相关，同 时事例的重用必须考虑事例的适用环境，提出一种层次智能检索模型，将事例描述分为 主题属性和环境属性两部分，基于人工神经网络技术和环境约束对事例进行聚类和检 索。该模型可提高事例的检索效率；避免仅靠相似度计算进行事例重用的不可靠性。

A Hierarchical Intelligent CBR Retrieval Model Supporting Reuse

Retrieval is the key technology in CBR. It imposes a direct effect on the efficiency and quality of CBR, and the quality of the retrieved case determines the difficulty of case reuse and adaptation. On the observation that case retrieval and representation have a close relationship with case memory, and the proper context must be taken into account during case reuse, a hierarchical intelligent retrieval model is proposed in this paper. In this model, case representation is composed of two parts, subject attributes and context attributes, and the clustering and retrieval of cases are based on ADN and context constraints. This model may improve the efficiency of case retrieval, and avoid the unreliabbility of case reuse by calculating the similarity only.