def generateMinPDist(transaction_list, trans4lcm, threshold, set_method, lcm_path, \
max_comb, permute_num, outlog, alternative):
# sys.stderr.write("--- original dataset ---\n")
# for j in transaction_list:
# j.output()
# sys.stderr.write("------\n")
starttime = time.time()
# Initialize the apriori and functinos using LAMP.
fre_pattern, lam_star, max_lambda, correction_term_time, func_f \
= lamp.runMultTest( transaction_list, trans4lcm, threshold, set_method, \
lcm_path, max_comb, outlog, alternative )
# calculate the set of minimum p-values using permuted data
min_p_list = [] # the list stores the minimum p-values
org_values_list = getValuesList(
transaction_list) # Raw (non-permuted) dataset
# estimate the probability distribution of the minimum p-value using permuted datasets.
for i in xrange(0, permute_num):
per_start = time.time()
permute_transaction_list, org2shuffled_list = permuteData(
transaction_list,
org_values_list) # generate the permuted dataset.
# for j in permute_transaction_list:
# sys.stderr.write("%s %s " % (j.id, j.name))
# sys.stderr.write("%s" % j.itemset)
# sys.stderr.write(" %s\n" % j.value)
func_f.calTime = 0
min_p, low_sup, freq_time = calculateMinimumPValue(
permute_transaction_list, trans4lcm, fre_pattern, func_f, max_comb,
org2shuffled_list)
per_time = time.time() - per_start
if (i == 0):
per_time = time.time() - starttime
min_p_list.append(
tuple([
min_p, low_sup,
fre_pattern.getTotal(low_sup), freq_time, per_time,
func_f.calTime
]))
outlog.write( "[permute %s] minP %s, minSupport %s, totalTest %s, freqTime %s, totalTime %s, #ofPvalue %s\n" \
% (i, min_p_list[i][0], min_p_list[i][1], min_p_list[i][2], \
min_p_list[i][3], min_p_list[i][4], min_p_list[i][5]))
return min_p_list, fre_pattern, func_f
def generateMinPDist(transaction_list, trans4lcm, threshold, set_method, lcm_path, \
max_comb, permute_num, outlog, alternative):
# sys.stderr.write("--- original dataset ---\n")
# for j in transaction_list:
# j.output()
# sys.stderr.write("------\n")
starttime = time.time()
# Initialize the apriori and functinos using LAMP.
fre_pattern, lam_star, max_lambda, correction_term_time, func_f \
= lamp.runMultTest( transaction_list, trans4lcm, threshold, set_method, \
lcm_path, max_comb, outlog, alternative )
# calculate the set of minimum p-values using permuted data
min_p_list = [] # the list stores the minimum p-values
org_values_list = getValuesList( transaction_list ) # Raw (non-permuted) dataset
# estimate the probability distribution of the minimum p-value using permuted datasets.
for i in xrange( 0, permute_num ):
per_start = time.time()
permute_transaction_list, org2shuffled_list = permuteData( transaction_list, org_values_list ) # generate the permuted dataset.
# for j in permute_transaction_list:
# sys.stderr.write("%s %s " % (j.id, j.name))
# sys.stderr.write("%s" % j.itemset)
# sys.stderr.write(" %s\n" % j.value)
func_f.calTime = 0
min_p, low_sup, freq_time = calculateMinimumPValue( permute_transaction_list, trans4lcm, fre_pattern,
func_f, max_comb, org2shuffled_list )
per_time = time.time() - per_start
if ( i == 0 ):
per_time = time.time() - starttime
min_p_list.append( tuple( [ min_p, low_sup, fre_pattern.getTotal( low_sup ), freq_time, per_time, func_f.calTime ] ) )
outlog.write( "[permute %s] minP %s, minSupport %s, totalTest %s, freqTime %s, totalTime %s, #ofPvalue %s\n" \
% (i, min_p_list[i][0], min_p_list[i][1], min_p_list[i][2], \
min_p_list[i][3], min_p_list[i][4], min_p_list[i][5]))
return min_p_list, fre_pattern, func_f
max_comb = lamp.convertMaxComb( max_comb, len(columnid2name) )
except ValueError, e:
return
except KeyError, e:
return
# run multiple test
transaction4lcm53 = transaction_file + ".4lcm53"
# run
try:
outlog = open( log_file, 'w' )
starttime = time.time()
sys.stderr.write( "Compute the optimal correction factor ..." )
fre_pattern, lam_star, max_lambda, correction_term_time, func_f \
= lamp.runMultTest(transaction_list, transaction4lcm53, threshold, \
set_method, lcm_path, max_comb, outlog, alternative)
k = fre_pattern.getTotal( lam_star )
sys.stderr.write( " %s\n" % k )
sys.stderr.write( "Compute P-values of testable combinations ...\n" )
enrich_lst, finish_test_time \
= lamp.fwerControll(transaction_list, fre_pattern, lam_star, max_lambda, \
threshold, func_f, columnid2name, outlog)
outlog.close()
except IOError, e:
outlog.close()
sys.stderr.write( "Output results ...\n" )
# output result
result = makeResult(transaction_file, flag_file, threshold, set_method, max_comb, \
columnid2name, lam_star, k, enrich_lst, transaction_list, func_f, alternative )
max_comb = lamp.convertMaxComb(max_comb, len(columnid2name))
except ValueError, e:
return
except KeyError, e:
return
# run multiple test
transaction4lcm53 = transaction_file + ".4lcm53"
# run
try:
outlog = open(log_file, 'w')
starttime = time.time()
sys.stderr.write("Compute the optimal correction factor ...")
fre_pattern, lam_star, max_lambda, correction_term_time, func_f \
= lamp.runMultTest(transaction_list, transaction4lcm53, threshold, \
set_method, lcm_path, max_comb, outlog, alternative)
k = fre_pattern.getTotal(lam_star)
sys.stderr.write(" %s\n" % k)
sys.stderr.write("Compute P-values of testable combinations ...\n")
enrich_lst, finish_test_time \
= lamp.fwerControll(transaction_list, fre_pattern, lam_star, max_lambda, \
threshold, func_f, columnid2name, outlog)
outlog.close()
except IOError, e:
outlog.close()
sys.stderr.write("Output results ...\n")
# output result
result = makeResult(transaction_file, flag_file, threshold, set_method, max_comb, \
columnid2name, lam_star, k, enrich_lst, transaction_list, func_f, alternative )
