Practical Applications of Data Mining

310 Chapter 7 Clustering

converges, and the five schools are finally classified into the two clusters

(B, C ) and (A, D, E ). The table is divided into two parts: one representing

larger schools (i.e., schools with a larger number of students) and the other

representing smaller schools. It should be pointed out that if other attributes

such as teachers and TAS were considered together with students, the pat-

tern of clusters might be different.

From the example above we can see that the k-means algorithm can

reach optimal clusters fairly quickly, which makes it very suitable for pro-

cessing large databases. However, the shortcomings ...