A Case for Usage of Case Usage in Case-Based Reasoning.
Mohamed Abdul Karim Sadiq, Deepak Khemani · 2005
Some of the issues in case retrieval and maintenance of case-bases are discussed. Conventionally, nearest neighbor algorithm is applied for retrieval of similar cases. Uniqueness of feature-values is not rewarded in such similarity matching functions. In order to refine the retrieval process, a metric based on successful case usage in the past is incorporated. This metric is shown to be useful in case-base maintenance. Applications in Homoeopathy remedy and personal tour selection are presented with comparative results.