An improving method of CBR retrieval based on self-organizing map
Case retrieval is the most crucial part in CBR. However, traditional case retrieval methods have many disadvantages on accuracy and efficiency. In order to cope with this problem, an improving method based on self-organizing maps (SOM) was proposed in this paper. Firstly, cluster previous cases into several groups using of SOM networks; secondly, input the new case into SOM networks, and identify the most similar case group according the visual clustering output; finally, to decide the most similar case according similarity. The advantage of this method is Cases visual clustering result provided by SOM networks greatly facilitating retrieval process and decreasing the retrieval time. Experimental results show that the proposed method may improve the efficiency of case retrieval.
Case-based reasoning (CBR) Case retrieval Self-organizing maps (SOM)
Du Hui
Nanjing University of Aeronautics & Astronautics,Nanjing China;Physical & Electrical Engineering department of Zaozhuang university,Shandong China
国际会议
上海
英文
616-620
2009-11-20（万方平台首次上网日期，不代表论文的发表时间）