def do_analysis(self):
log.info("Performing clustering analysis")
partnum = len(self.cfg.partitions)
subset_count = 2 * partnum - 1
scheme_count = partnum
self.cfg.progress.begin(scheme_count, subset_count)
# Start with the most partitioned scheme
start_description = range(len(self.cfg.partitions))
start_scheme = scheme.create_scheme(
self.cfg, "start_scheme", start_description)
# Analyse our first scheme
log.info("Analysing starting scheme (scheme %s)" % start_scheme.name)
self.analyse_scheme(start_scheme)
# Current scheme number
cur_s = 2
# Now we try out all clusterings of the first scheme, to see if we can
# find a better one
while True:
log.info("***Clustering algorithm step %d of %d***" %
(cur_s - 1, partnum - 1))
# Calculate the subsets which are most similar
# e.g. combined rank ordering of euclidean distances
# Could combine average site-rates, q matrices, and frequencies
scheme_name = "step_%d" % (cur_s - 1)
clustered_scheme = neighbour.get_nearest_neighbour_scheme(
start_scheme, scheme_name, self.cfg)
# Now analyse that new scheme
cur_s += 1
self.analyse_scheme(clustered_scheme)
# Stop when we've anlaysed the scheme with all subsets combined
if len(set(clustered_scheme.subsets)) == 1: # then it's the scheme with everything together
break
else:
start_scheme = clustered_scheme
self.cfg.progress.end()
txt = "Best scheme using Clustering Analysis, analysed with %s" % self.cfg.model_selection
self.cfg.reporter.write_best_scheme(txt, self.results)
def do_analysis(self):
log.info("Performing strict clustering analysis")
partnum = len(the_config.user_subsets)
subset_count = 2 * partnum - 1
scheme_count = partnum
the_config.progress.begin(scheme_count, subset_count)
# Start with the most partitioned scheme
start_description = range(partnum)
start_scheme = scheme.create_scheme(the_config, "start_scheme",
start_description)
# Analyse our first scheme
log.info("Analysing starting scheme (scheme %s)" % start_scheme.name)
self.analyse_scheme(start_scheme)
# Current scheme number
cur_s = 2
# Now we try out all clusterings of the first scheme, to see if we can
# find a better one
while True:
log.info("***Strict clustering algorithm step %d of %d***" %
(cur_s - 1, partnum - 1))
# Calculate the subsets which are most similar
# e.g. combined rank ordering of euclidean distances
# Could combine average site-rates, q matrices, and frequencies
scheme_name = "step_%d" % (cur_s - 1)
clustered_scheme = neighbour.get_nearest_neighbour_scheme(
start_scheme, scheme_name, the_config)
# Now analyse that new scheme
cur_s += 1
self.analyse_scheme(clustered_scheme)
# Stop when we've analysed the scheme with all subsets combined...
if len(set(clustered_scheme.subsets)) == 1:
# ... then it's the scheme with everything together
break
else:
# We keep going
start_scheme = clustered_scheme
the_config.progress.end()
the_config.reporter.write_best_scheme(self.results)