I'm wondering as in the first step in case selection is case representation. Any idea how can the case indexing be done? other than just a pointer you do it roughly which points on the main feature within the case?
Indexing a case is considered as one of the open problems in the CBR. Among the techniques used to explore and solve the problem of indexing in a case is the explanation-based technique which is used to identify the relevant features of the case.
Try also reading more in Heuristic search techniques and Quantitative Models.
Selecting the appropriate attributes that define each case is the most important and hard task to do. Case indexing is the 1st step in the case representation step in in CBR problem solving life cycle....
Do any believe case indexing can be better done in anyway than DB technology? Like defining the structure and description of each attribute within the case? Please share your experience.