Practical Applications of Data Mining

318 Chapter 7 Clustering

To alleviate the problem of separating close values with the same behav-

ior when selecting a group of data values as an interval, we need a measure of

interval quality that is based on the distance range between the data points. The

semantics of interval data used in clustering a set of data values into a pattern

may influence not only the choice of data grouping but also the semantic

interpretation of the rules generated.

The relations S and T in Tables 7.27 and 7.28 illustrate the importance

of the semantics of interval data. From both relations the following associa-

tion rule can be generated:

(major = CS) ∧ (