def train_model(path_list,dst_folder):
train_file = path_list[0]
test_file = path_list[1]
result_all = {}
#random the file
if rand_num > 1:
rand_file = train_file + '_rand'
else:
rand_file = train_file
rand_file_cache = rand_file + '_cache'
for k in range(0,rand_num):
if rand_num > 1:
print 'shuffle datset...'
sol_shuffle.sol_shuffle(train_file, rand_file)
cmd_data = dataset.get_cmd_data_by_file(rand_file, test_file, is_cache)
dataset.analyze(rand_file);
for opt in opt_list:
print '-----------------------------------'
print ' Experiment on %s' %opt + ' Random %d' %k
print '-----------------------------------'
if opt == 'vw':
result_file = 'vw_result_%d' %k + '.txt'
result_once = run_vw.run_vw(rand_file, test_file, ds, result_file, is_cache)
elif opt == 'liblinear':
result_file = 'liblinear_result_%d' %k + '.txt'
result_once = run_liblinear.run_liblinear(rand_file, test_file, ds, result_file)
elif opt == 'fgm':
result_file = 'fgm_result_%d' %k + '.txt'
result_once = run_fgm.run_fgm(rand_file, test_file, ds, result_file)
elif opt == 'mRMR':
result_file = dst_folder + '/%s' %opt + '_result_%d' %k + '.txt'
result_once = run_mRMR.run_mRMR(rand_file, test_file, ds, result_file)
print '\nparsing result...'
#write the result to file
parse_file = dst_folder +'/%s' %opt + '_%d' %k + '.txt'
result_once2 = run_util.parse_result(result_file, parse_file);
bs_num = len(result_once)
if bs_num != len(result_once2):
print 'inconsistent parsing result'
for m in range(0,bs_num):
result_once[m][0] = result_once2[m][0]
result_once[m][1] = result_once2[m][1]
if result_once[m][2] == 0:
result_once[m][2] = result_once2[m][2]
if result_once[m][3] == 0:
result_once[m][3] = result_once2[m][3]
else:
result_file = dst_folder + '/%s' %opt + '_result_%d' %k + '.txt'
cmd = cmd_data
cmd += extra_cmd
if is_default_param == False:
cmd += dataset.get_model_param(ds, opt)
run_experiment.run_experiment(opt,result_file,ds, cmd)
print '\nparsing result...'
#write the result to file
parse_file = dst_folder +'/%s' %opt + '_%d' %k + '.txt'
result_once = run_util.parse_result(result_file, parse_file);
result_all = add_to_dict(opt,result_all, result_once)
#remove previous file
if rand_num > 1:
os.system('rm -f %s' %rand_file_cache)
os.system('rm -f %s' %rand_file)
#average the result
for opt in opt_list:
rows = len(result_all[opt])
cols = len(result_all[opt][0])
for k in range(0,rows):
for m in range(0,cols):
result_all[opt][k][m] /= rand_num
return result_all
return result_list
result_item_num = 4
result_list_all = [[0 for y in range(0,result_item_num)] for x in range(0,grid_size)]
for split_item in split_list:
test_file = split_item
train_file_list = filter(lambda x : x != split_item,split_list)
train_file = test_file + '_train'
os.system('rm -f %s' %train_file)
for x in train_file_list:
merge_cmd = 'cat %s' %x + ' >> %s' %train_file
os.system(merge_cmd)
cmd_data = dataset.get_cmd_data_by_file(train_file, test_file)
result_list_one = run_one_data(exe_cmd + cmd_data)
#raw_input('type to continue' )
for k in range(0,grid_size):
for m in range(0,result_item_num):
result_list_all[k][m] += float(result_list_one[k][m])
#delete the temp files
os.system('rm -f %s' %train_file)
os.system('rm -f %s' %train_file + '_cache')
#average the results
for k in range(0,grid_size):
for m in range(0,result_item_num):
result_list_all[k][m] /= fold_num
def train_model(path_list, dst_folder):
train_file = path_list[0]
test_file = path_list[1]
result_all = {}
#random the file
if rand_num > 1:
rand_file = train_file + '_rand'
else:
rand_file = train_file
rand_file_cache = rand_file + '_cache'
for k in range(0, rand_num):
if rand_num > 1:
print 'shuffle datset...'
sol_shuffle.sol_shuffle(train_file, rand_file)
cmd_data = dataset.get_cmd_data_by_file(rand_file, test_file, is_cache)
dataset.analyze(rand_file)
for opt in opt_list:
for mp_method in mp_list:
cmd_mp = ' -mpt %s ' % mp_method
for m in range(0, run_time):
print '-----------------------------------'
print ' Experiment on %s' % opt + ' Random %d' % k + ' Multi-Pass %s' % mp_method + ' Round %d' % m
print '-----------------------------------'
result_file = dst_folder + '/%s' % opt + '_rand_%d' % k + '_mp_%s' % mp_method + '_round_%d' % m + '.txt'
cmd = cmd_data
cmd += cmd_mp
cmd += extra_cmd
if is_default_param == False:
cmd += dataset.get_model_param(ds, opt)
run_mpol.run_mpol(opt, result_file, ds, cmd)
print '\nparsing result...'
#write the result to file
parse_file = dst_folder + '/%s' % opt + '_%d' % k + '_%s' % mp_method + '_%d' % m + '.txt'
result_once = run_util.parse_result(
result_file, parse_file)
result_all = add_to_dict(opt, mp_method, result_all,
result_once)
if mp_method == 'none':
break
#remove previous file
if rand_num > 1:
os.system('rm -f %s' % rand_file_cache)
os.system('rm -f %s' % rand_file)
#average the result
for opt in opt_list:
for mp in mp_list:
rows = len(result_all[opt][mp])
cols = len(result_all[opt][mp][0])
divid = rand_num
if mp != 'none':
divid *= run_time
for k in range(0, rows):
for m in range(0, cols):
result_all[opt][mp][k][m] /= divid
return result_all
return result_list
result_item_num = 4
result_list_all = [[0 for y in range(0,result_item_num)] for x in range(0,grid_size)]
for split_item in split_list:
test_file = split_item
train_file_list = filter(lambda x : x != split_item,split_list)
train_file = test_file + '_train'
os.system('rm -f %s' %train_file)
for x in train_file_list:
merge_cmd = 'cat %s' %x + ' >> %s' %train_file
os.system(merge_cmd)
cmd_data = dataset.get_cmd_data_by_file(train_file, test_file)
result_list_one = run_one_data(exe_cmd + cmd_data)
#raw_input('type to continue' )
for k in range(0,grid_size):
for m in range(0,result_item_num):
result_list_all[k][m] += float(result_list_one[k][m])
#delete the temp files
os.system('rm -f %s' %train_file)
os.system('rm -f %s' %train_file + '_cache')
#average the results
for k in range(0,grid_size):
for m in range(0,result_item_num):
result_list_all[k][m] /= fold_num
def train_model(path_list,dst_folder):
train_file = path_list[0]
test_file = path_list[1]
result_all = {}
#random the file
if rand_num > 1:
rand_file = train_file + '_rand'
else:
rand_file = train_file
rand_file_cache = rand_file + '_cache'
for k in range(0,rand_num):
if rand_num > 1:
print 'shuffle datset...'
sol_shuffle.sol_shuffle(train_file, rand_file)
cmd_data = dataset.get_cmd_data_by_file(rand_file, test_file, is_cache)
dataset.analyze(rand_file);
for opt in opt_list:
for mp_method in mp_list:
cmd_mp = ' -mpt %s ' %mp_method
for m in range(0,run_time):
print '-----------------------------------'
print ' Experiment on %s' %opt + ' Random %d' %k + ' Multi-Pass %s' %mp_method + ' Round %d' %m
print '-----------------------------------'
result_file = dst_folder + '/%s' %opt + '_rand_%d' %k + '_mp_%s' %mp_method + '_round_%d' %m + '.txt'
cmd = cmd_data
cmd += cmd_mp
cmd += extra_cmd
if is_default_param == False:
cmd += dataset.get_model_param(ds, opt)
run_mpol.run_mpol(opt,result_file,ds, cmd)
print '\nparsing result...'
#write the result to file
parse_file = dst_folder +'/%s' %opt + '_%d' %k + '_%s' %mp_method + '_%d' %m + '.txt'
result_once = run_util.parse_result(result_file, parse_file);
result_all = add_to_dict(opt,mp_method,result_all, result_once)
if mp_method == 'none':
break
#remove previous file
if rand_num > 1:
os.system('rm -f %s' %rand_file_cache)
os.system('rm -f %s' %rand_file)
#average the result
for opt in opt_list:
for mp in mp_list:
rows = len(result_all[opt][mp])
cols = len(result_all[opt][mp][0])
divid = rand_num
if mp != 'none':
divid *= run_time
for k in range(0,rows):
for m in range(0,cols):
result_all[opt][mp][k][m] /= divid
return result_all
def run_mRMR(train_file, test_file,ds, result_file):
result_all = []
train_exe_name = exe_path.mRMR
#make the result dir
dst_folder = './%s' %ds
run_util.create_dir(dst_folder)
data_valid_dim = run_util.get_valid_dim(train_file)
data_num = run_util.get_data_num(train_file)
#bs_list = l1_def.get_lambda_list(ds,'mRMR')
if 'synthetic' in ds:
bs_list = l1_def.get_lambda_list(ds,'mRMR')
else:
lambda_list = l1_def.get_lambda_list(ds,'mRMR')
bs_list = []
b_num = len(lambda_list)
for i in range(0,b_num):
dim = int(data_valid_dim * (1 - lambda_list[i]))
if dim > 0 and dim <= 500:
bs_list.append(dim)
bs_list = l1_def.get_lambda_list(ds,'mRMR')
#clear the file if it already exists
open(result_file,'w').close()
for bs in bs_list:
result_once = [0,0,0,0]
model_file = dst_folder + '/mRMR_model%d' %bs
parse_file = dst_folder + '/mRMR_model_parse%d' %bs
if os.path.exists(model_file) == False:
print model_file + ' not exist'
csv_train_file = train_file + '.csv'
if os.path.exists(csv_train_file) == False:
#convert data
print 'convert data'
cmd = exe_path.csv_converter + ' -i %s' %train_file + ' -o %s' %csv_train_file
cmd += ' -sdt libsvm -ddt csv'
print cmd
os.system(cmd)
#run mRMR
prev_cmd = train_exe_name + ' -v %d' %data_valid_dim + ' -t 0.5 -i %s' %csv_train_file
cmd = prev_cmd + ' -n %d' %bs + ' > %s' %model_file
print cmd
start_time =time.time()
os.system(cmd)
end_time = time.time()
#parse learning time
train_time = (float)(end_time - start_time)
result_once[3] = train_time
#parse result
parse_model_file(model_file,parse_file);
#run OGD
cmd_data = dataset.get_cmd_data_by_file(train_file, test_file,True)
cmd = exe_path.SOL_exe_name + cmd_data + ' -m %s' %parse_file + ' -k %d' %bs
cmd += dataset.get_model_param(ds,'SGD-FS')
cmd += ' -opt mRMR_OGD -norm -loss Hinge >> %s' %result_file
print cmd
os.system(cmd)
result_once[2] = bs
result_all.append(result_once)
return result_all