Features

• How do you calculate similarity? (We'll do this in a bit.)
• How do you describe the cases?
• The real work in actually implementing a CBR system is figuring out a good way to describe the cases.
• Typically you pick a set of features and describe their possible values.
• This is quite similar to defining working memory in a rule based system.
• However the values you allow are important in calculating similarity.
• You need to select the relevant features.
• Do all features need to be equally relevant?