Scale Conversion

In many applications, the variables describing the objects to be clustered will not be measured in the same scales. They may often be variables of completely different types, some interval, others categorical. There are three approaches to cluster objects described by variables of different types. One is to use a similarity coefficient, which can incorporate information from different types of variable, such as the general coefficients described in Section 6.5. The second is to carry out separate analyses of the same set of objects, each analysis involving variables of a single type only, and then to synthesize the results from different analyses. The third is to convert the variables of different types to variables of the same type, such as converting all variables describing the objects to be clustered into categorical variables.