readme

变色龙层次聚类算法

This program is developed according to the CHAMELEON clustering algorithm.

In using this program, first there are several constant numbers need to be 
noticed. In application, user may need to adjust these constants in the 
program. Of course, these constants can be changed to become input parameters,
but in my experiments they are mostly don't need to be changed, so I just
made them constants. These constants include:
 
MAXSIZE -- maximum number of data points
K_nearest -- k-nearest graph
BC -- balance constraints in graph partition
MINSIZE -- minimum size in graph partition

following three constants relating to the clustering of 2-dimentional spatial
data:
WIDTH -- width of the 2-dimentional space
HEIGHT -- height of the 2-dimentional spac
DIAG -- diagonal distance of the 2-dimentional spac

ALPHA -- This constant is related to the relative weight of closeness and 
connections in the clustering goodness function:

Once the above constants are set, the usage of the program is like this:
"chameleon N fileName stopClusterNO"
the meanning of each parameter is:
N -- number of data points
fileName -- name of the data file
stopClusterNO -- cluster number when the cluster process stop