def run_tests(na, nc, ncat, na_gen, pcerrors, nseeds, vetot, vetom, vparam,
filename):
# Create the CSV writer
f = open(filename, 'wb')
writer = csv.writer(f)
# Write the test options
writer.writerow(['algorithm', algo.__name__])
writer.writerow(['veto type', vetot])
writer.writerow(['veto mode', vetom])
writer.writerow(['veto weights', indep_veto_weights])
writer.writerow(['veto param', vparam])
writer.writerow(['na', na])
writer.writerow(['nc', nc])
writer.writerow(['ncat', ncat])
writer.writerow(['na_gen', na_gen])
writer.writerow(['pcerrors', pcerrors])
writer.writerow(['nseeds', nseeds])
writer.writerow(['', ''])
# Create a test_results instance
results = test_results()
# Initialize the seeds
seeds = range(nseeds)
# Run the algorithm
initialized = False
for _na, _nc, _ncat, _na_gen, _vparam, _pcerrors, seed \
in product(na, nc, ncat, na_gen, vparam, pcerrors, seeds):
t1 = time.time()
t = test_mip_mrsort_vc(seed, _na, _nc, _ncat, _na_gen, _vparam,
_pcerrors)
t2 = time.time()
if initialized is False:
fields = t.get_attributes()
writer.writerow(fields)
initialized = True
t.tocsv(writer, fields)
f.flush()
print("%s (%5f seconds)" % (t, t2 - t1))
results.append(t)
# Perform a summary
writer.writerow(['', ''])
t = results.summary(
['na', 'nc', 'ncat', 'na_gen', 'veto_param', 'pcerrors'], [
'na_err', 'nv_m1_learning', 'nv_m2_learning', 'nv_m1_gen',
'nv_m2_gen', 'ca_best', 'ca_errors', 'ca_gen', 'ca_gen_err',
't_total'
])
t.tocsv(writer)
def run_tests(na, nc, ncat, na_gen, pcerrors, nseeds,
vetot, vetom, vparam, filename):
# Create the CSV writer
f = open(filename, 'wb')
writer = csv.writer(f)
# Write the test options
writer.writerow(['algorithm', algo.__name__])
writer.writerow(['veto type', vetot])
writer.writerow(['veto mode', vetom])
writer.writerow(['veto weights', indep_veto_weights])
writer.writerow(['veto param', vparam])
writer.writerow(['na', na])
writer.writerow(['nc', nc])
writer.writerow(['ncat', ncat])
writer.writerow(['na_gen', na_gen])
writer.writerow(['pcerrors', pcerrors])
writer.writerow(['nseeds', nseeds])
writer.writerow(['', ''])
# Create a test_results instance
results = test_results()
# Initialize the seeds
seeds = range(nseeds)
# Run the algorithm
initialized = False
for _na, _nc, _ncat, _na_gen, _vparam, _pcerrors, seed \
in product(na, nc, ncat, na_gen, vparam, pcerrors, seeds):
t1 = time.time()
t = test_mip_mrsort_vc(seed, _na, _nc, _ncat, _na_gen,
_vparam, _pcerrors)
t2 = time.time()
if initialized is False:
fields = t.get_attributes()
writer.writerow(fields)
initialized = True
t.tocsv(writer, fields)
f.flush()
print("%s (%5f seconds)" % (t, t2 - t1))
results.append(t)
# Perform a summary
writer.writerow(['', ''])
t = results.summary(['na', 'nc', 'ncat', 'na_gen', 'veto_param',
'pcerrors'],
['na_err', 'nv_m1_learning', 'nv_m2_learning',
'nv_m1_gen', 'nv_m2_gen', 'ca_best', 'ca_errors',
'ca_gen', 'ca_gen_err', 't_total'])
t.tocsv(writer)
def run_tests(na, nc, ncat, na_gen, pcerrors, nseeds, filename):
# Create the CSV writer
f = open(filename, 'wb')
writer = csv.writer(f)
# Write the test options
writer.writerow(['algorithm', algo.__name__])
writer.writerow(['na', na])
writer.writerow(['nc', nc])
writer.writerow(['ncat', ncat])
writer.writerow(['na_gen', na_gen])
writer.writerow(['pcerrors', pcerrors])
writer.writerow(['nseeds', nseeds])
writer.writerow(['random_model_type', random_model_type])
writer.writerow(['', ''])
# Create a test_results instance
results = test_results()
# Initialize the seeds
seeds = range(nseeds)
# Run the algorithm
initialized = False
for _na, _nc, _ncat, _na_gen, _pcerrors, seed \
in product(na, nc, ncat, na_gen, pcerrors, seeds):
t1 = time.time()
t = test_meta_electre_tri_global(seed, _na, _nc, _ncat, _na_gen,
_pcerrors)
t2 = time.time()
if initialized is False:
fields = [
'seed', 'na', 'nc', 'ncat', 'na_gen', 'pcerrors', 'na_err',
'ca_best', 'ca_errors', 'ca2_best', 'ca2_errors', 'ca_gen',
'ca_gen_err', 't_total'
]
writer.writerow(fields)
initialized = True
t.tocsv(writer, fields)
f.flush()
print("%s (%5f seconds)" % (t, t2 - t1))
results.append(t)
# Perform a summary
writer.writerow(['', ''])
t = results.summary(['na', 'nc', 'ncat', 'na_gen', 'pcerrors'], [
'na_err', 'ca_best', 'ca_errors', 'ca2_best', 'ca2_errors', 'ca_gen',
'ca_gen_err', 't_total'
])
t.tocsv(writer)
def run_tests(na, nc, ncat, na_gen, pcerrors, nseeds, filename):
# Create the CSV writer
f = open(filename, 'wb')
writer = csv.writer(f)
# Write the test options
writer.writerow(['algorithm', algo.__name__])
writer.writerow(['na', na])
writer.writerow(['nc', nc])
writer.writerow(['ncat', ncat])
writer.writerow(['na_gen', na_gen])
writer.writerow(['pcerrors', pcerrors])
writer.writerow(['nseeds', nseeds])
writer.writerow(['random_model_type', random_model_type])
writer.writerow(['', ''])
# Create a test_results instance
results = test_results()
# Initialize the seeds
seeds = range(nseeds)
# Run the algorithm
initialized = False
for _na, _nc, _ncat, _na_gen, _pcerrors, seed \
in product(na, nc, ncat, na_gen, pcerrors, seeds):
t1 = time.time()
t = test_meta_electre_tri_global(seed, _na, _nc, _ncat, _na_gen,
_pcerrors)
t2 = time.time()
if initialized is False:
fields = ['seed', 'na', 'nc', 'ncat', 'na_gen', 'pcerrors',
'na_err',
'ca_best', 'ca_errors', 'ca2_best', 'ca2_errors',
'ca_gen', 'ca_gen_err', 't_total']
writer.writerow(fields)
initialized = True
t.tocsv(writer, fields)
f.flush()
print("%s (%5f seconds)" % (t, t2 - t1))
results.append(t)
# Perform a summary
writer.writerow(['', ''])
t = results.summary(['na', 'nc', 'ncat', 'na_gen', 'pcerrors'],
['na_err', 'ca_best', 'ca_errors', 'ca2_best',
'ca2_errors', 'ca_gen', 'ca_gen_err',
't_total'])
t.tocsv(writer)
def run_tests(nseeds, data, pclearning, nloop, nmodels, nmeta, filename):
# Create the CSV writer
f = open(filename, 'wb')
writer = csv.writer(f)
# Write the test options
writer.writerow(['data', data.name])
writer.writerow(['heur_init_profiles', heur_init_profiles.__name__])
writer.writerow(['lp_weights', lp_weights.__name__])
writer.writerow(['heur_profiles', heur_profiles.__name__])
writer.writerow(['nloop', nloop])
writer.writerow(['nmodels', nmodels])
writer.writerow(['nmeta', nmeta])
writer.writerow(['pclearning', pclearning])
# Create a test_results instance
results = test_results()
# Initialize the seeds
seeds = range(nseeds)
# Run the algorithm
initialized = False
for _pclearning, _nloop, _nmodels, _nmeta, seed in product(
pclearning, nloop, nmodels, nmeta, seeds):
t1 = time.time()
t = run_test(seed, data, _pclearning, _nloop, _nmodels, _nmeta)
t2 = time.time()
if initialized is False:
fields = t.get_attributes()
writer.writerow(fields)
initialized = True
t.tocsv(writer, fields)
f.flush()
print("%s (%5f seconds)" % (t, t2 - t1))
results.append(t)
# Perform a summary
writer.writerow(['', ''])
t = results.summary([
'na', 'nc', 'ncat', 'pclearning', 'na_learning', 'na_test', 'nloop',
'nmodels', 'nmeta'
], [
'ca_learning', 'ca_test', 'ca_all', 'auc_learning', 'auc_test',
'auc_all', 't_total'
])
t.tocsv(writer)
def run_tests(nseeds, data, pclearning, nloop, nmodels, nmeta, filename):
# Create the CSV writer
f = open(filename, 'wb')
writer = csv.writer(f)
# Write the test options
writer.writerow(['data', data.name])
writer.writerow(['heur_init_profiles', heur_init_profiles.__name__])
writer.writerow(['lp_weights', lp_weights.__name__])
writer.writerow(['heur_profiles', heur_profiles.__name__])
writer.writerow(['nloop', nloop])
writer.writerow(['nmodels', nmodels])
writer.writerow(['nmeta', nmeta])
writer.writerow(['pclearning', pclearning])
# Create a test_results instance
results = test_results()
# Initialize the seeds
seeds = range(nseeds)
# Run the algorithm
initialized = False
for _pclearning, _nloop, _nmodels, _nmeta, seed in product(pclearning,
nloop, nmodels, nmeta, seeds):
t1 = time.time()
t = run_test(seed, data, _pclearning, _nloop, _nmodels, _nmeta)
t2 = time.time()
if initialized is False:
fields = t.get_attributes()
writer.writerow(fields)
initialized = True
t.tocsv(writer, fields)
f.flush()
print("%s (%5f seconds)" % (t, t2 - t1))
results.append(t)
# Perform a summary
writer.writerow(['', ''])
t = results.summary(['na', 'nc', 'ncat', 'pclearning', 'na_learning',
'na_test', 'nloop', 'nmodels', 'nmeta'],
['ca_learning', 'ca_test', 'ca_all',
'auc_learning', 'auc_test', 'auc_all', 't_total'])
t.tocsv(writer)
def run_test(m, n, lbda_min, lbda_max, filename):
# Create the CSV writer
f = open(filename, 'wb')
writer = csv.writer(f)
# Create a test_results instance
results = test_results()
initialized = False
# Run the algorithm
c = generate_criteria(n)
coalitions = [ i for i in powerset(c.keys()) ]
for i in range(m):
t = one_loop(c, lbda_min, coalitions)
t.test_name = "%d-%d" % (i, n)
t['i'] = i
t['nc'] = n
t['lbda_min'] = lbda_min
t['lbda_max'] = lbda_max
t['n'] = m
if initialized is False:
fields = ['i', 'nc', 'lbda_min', 'lbda_max'] \
+ [pprint_coalition(j) for j in coalitions] \
+ ["%d_criteria" % i for i in range(len(c) + 1)] \
+ ["ncoalitions"]
writer.writerow(fields)
initialized = True
results.append(t)
t.tocsv(writer, fields)
f.flush()
print("%s" % t)
# Perform a summary
writer.writerow(['', ''])
t = results.summary(['nc', 'lbda_min', 'lbda_max', 'n'], [pprint_coalition(j)
for j in coalitions], [], [])
t.tocsv(writer)
writer.writerow(['', ''])
t = results.summary(['nc', 'lbda_min', 'lbda_max', 'n'],
["%d_criteria" % i for i in range(len(c) + 1)] \
+ ['ncoalitions'], ['ncoalitions'],
['ncoalitions'])
t.tocsv(writer)
def run_tests(na, nc, ncat, ns, na_gen, pcerrors, nseeds, filename):
# Create the CSV writer
writer = csv.writer(open(filename, 'wb'))
# Write the test options
writer.writerow(['na', na])
writer.writerow(['nc', nc])
writer.writerow(['ncat', ncat])
writer.writerow(['ns', ns])
writer.writerow(['na_gen', na_gen])
writer.writerow(['pcerrors', pcerrors])
writer.writerow(['nseeds', nseeds])
writer.writerow(['', ''])
# Create a test results instance
results = test_results()
# Initialize the seeds
seeds = range(nseeds)
# Run the algorithm
initialized = False
for _na, _nc, _ncat, _ns, _na_gen, _pcerrors, seed \
in product(na, nc, ncat, ns, na_gen, pcerrors, seeds):
t1 = time.time()
t = test_lp_avfsort(seed, _na, _nc, _ncat, _ns, _na_gen, _pcerrors)
t2 = time.time()
if initialized is False:
writer.writerow(t.get_attributes())
initialized = True
t.tocsv(writer)
print("%s (%5f seconds)" % (t, t2 - t1))
results.append(t)
# Perform a summary
writer.writerow(['', ''])
t = results.summary(['na', 'nc', 'ncat', 'ns', 'na_gen', 'pcerrors'], [
'na_err'
'obj', 'ca', 'ca_errors', 'altered', 'ca2', 'ca2_errors', 'ca_gen',
'ca_gen_err', 't_total', 't_const', 't_solve'
])
t.tocsv(writer)
def run_tests(na, nc, ncat, na_gen, pcerrors, nseeds, filename):
# Create the CSV writer
writer = csv.writer(open(filename, 'wb'))
# Write the test options
writer.writerow(['na', na])
writer.writerow(['nc', nc])
writer.writerow(['ncat', ncat])
writer.writerow(['na_gen', na_gen])
writer.writerow(['pcerrors', pcerrors])
writer.writerow(['nseeds', nseeds])
writer.writerow(['random_model_type', random_model_type])
writer.writerow(['learned_model_type', learned_model_type])
writer.writerow(['', ''])
# Create a test results instance
results = test_results()
# Initialize the seeds
seeds = range(nseeds)
# Run the algorithm
initialized = False
for _na, _nc, _ncat, _na_gen, _pcerrors, seed \
in product(na, nc, ncat, na_gen, pcerrors, seeds):
t1 = time.time()
t = test_lp_learning_weights(seed, _na, _nc, _ncat, _na_gen,
_pcerrors)
t2 = time.time()
if initialized is False:
writer.writerow(t.get_attributes())
initialized = True
t.tocsv(writer)
print("%s (%5f seconds)" % (t, t2 - t1))
results.append(t)
# Perform a summary
writer.writerow(['', ''])
t = results.summary(['na', 'nc', 'ncat', 'na_gen', 'pcerrors'],
['obj', 'ca', 'ca_errors', 'ca2', 'ca2_errors',
'ca_gen', 't_total', 't_const', 't_solve'])
t.tocsv(writer)
def run_tests(na, nc, ncat, pcexamples, pcerrors, nseeds, filename):
# Create the CSV writer
writer = csv.writer(open(filename, 'wb'))
# Write the test options
writer.writerow(['na', na])
writer.writerow(['nc', nc])
writer.writerow(['ncat', ncat])
writer.writerow(['pcexamples', pcexamples])
writer.writerow(['pcerrors', pcerrors])
writer.writerow(['nseeds', nseeds])
writer.writerow(['', ''])
# Create a test results instance
results = test_results()
# Initialize the seeds
seeds = range(nseeds)
# Run the algorithm
initialized = False
for _na, _nc, _ncat, _pcexamples, _pcerrors, seed \
in product(na, nc, ncat, pcexamples, pcerrors, seeds):
t1 = time.time()
t = test_heur_mrsort_coalitions(seed, _na, _nc, _ncat, _pcexamples,
_pcerrors)
t2 = time.time()
if initialized is False:
writer.writerow(t.get_attributes())
initialized = True
t.tocsv(writer)
print("%s (%5f seconds)" % (t, t2 - t1))
results.append(t)
# Perform a summary
writer.writerow(['', ''])
t = results.summary(['na', 'nc', 'ncat', 'pcexamples', 'pcerrors'],
['ncoal', 'ncoal_ni', 'ncoal_add'])
t.tocsv(writer)
def run_tests(na, nc, ncat, pcerrors, nseeds, filename):
# Create the CSV writer
writer = csv.writer(open(filename, 'wb'))
# Write the test options
writer.writerow(['na', na])
writer.writerow(['nc', nc])
writer.writerow(['ncat', ncat])
writer.writerow(['pcerrors', pcerrors])
writer.writerow(['nseeds', nseeds])
writer.writerow(['', ''])
# Create a test results instance
results = test_results()
# Initialize the seeds
seeds = range(nseeds)
# Run the algorithm
initialized = False
for _na, _nc, _ncat, _pcerrors, seed \
in product(na, nc, ncat, pcerrors, seeds):
t1 = time.time()
t = test_heur_mrsort_init_profiles(seed, _na, _nc, _ncat, _pcerrors)
t2 = time.time()
if initialized is False:
writer.writerow(t.get_attributes())
initialized = True
t.tocsv(writer)
print("%s (%5f seconds)" % (t, t2 - t1))
results.append(t)
# Perform a summary
writer.writerow(['', ''])
t = results.summary(['na', 'nc', 'ncat', 'pcerrors'],
['ca_rdom', 'ca_heur'])
t.tocsv(writer)
