timecols = [11]
valcols = [3,4,5,6,7,8,9,11]
parser = Parser("prunedPhones.csv",namecols,timecols,valcols)
#Get either every dimension or a single dimension
val1s = [parser.getProperties()[0]]
val2s = [parser.getProperties()[1]]
weightFactors = [0.2]
Cs = [20]
gammas = [0.2]
#'''
od = ObservedDistribution(parser, val1s[0], contours, val2s[0], weightFactors[0])
ed = ExpectedDistribution(od,parallel=False)
edv = ExpectedDistributionVisualiser(ed,od,50,50)
fig = od.plotObservedContours(title='', alpha=0.25)
edv.save('dummy.pdf')
#'''
print "---------------------Parallel=True---------------------"
for val1 in val1s:
for val2 in val2s:
if val1 is not val2:
start_time = time.time()
#print "Modelling",val1,"(independent) against",val2,"(dependent)."
od = ObservedDistribution(parser, val1, contours, val2, weightFactors[0])
ed = ExpectedDistribution(od,{'C':Cs[0],'gamma':gammas[0]},parallel=True)
edv = ExpectedDistributionVisualiser(ed,od,50,50)
def __init__(self, parser_train, ind_attr, contours , dep_attr, weight_std_ratio=1, parser_test=None, retrain=True, prefix=None): ObservedDistribution.__init__(self, parser_train, ind_attr, contours , dep_attr, weight_std_ratio, retrain, prefix, save=False) if parser_test is not None: self.addTestData(parser_test) else: self.test_parser = None
