The km program implements the multidimensional supervised clustering kmeans algorithm 

It requires 3 arguments:  data file name, number of clusters and
choice for normalizing/not normalizing data.
The data file is a text file with or without header line. Columns must
be delimited by a valid character (for instance ;). Spaces,
tabulations and points (.) are not valid separators.
The third argument is 0 to woirk with data as they are, or 1 to
normalize data between 0 and 1 prior to classification.

The -a option displays detailed results.
Without that option, the program works silently.

In both cases, it creates two result files:
- centers.km, with as many rows as the required number of
clusters. Each row contains a cluster center coordinates.
- result.km,  with as many rows as the data file. Each row contains
the row number and its cluster number assignation.





