•jpscan and jarpat both
perform Jarvis-Patrick clustering based on nearest neighbours (NN) data. Both programs use two
Jarvis-Patrick clustering parameters:
the number of neighbors to examine and the number required to be in common.
clusters data using all possible parameter combinations up to a given limit (typically set to the list length,
default is 16) and outputs tables
of statistics intended to help in selecting a pair of parameters appropriate to the problem at hand.
that the parameters be specified and outputs the clustering results.
•It is advisable to run jpscan and examine its output
before running jarpat.
•Both programs also allow control of the way the
clustering search is done :
–as published (the default)
–an exhaustive search (only useful for very small data
–a faster search which approximates the exhaustive search