if not len(sys.argv) == 1:
leave_k_out = int(sys.argv[1])
print 'Use default leave k out: k=' + str(leave_k_out)
if not len(sys.argv) <= 2:
lafactor = int(sys.argv[2])
print 'Use default latent factor: ' + str(lafactor)
max_rank = 2000
# number of repetitions.
total_iteration = 2
# recommendation algorithms RandUV(latent_factor=lafactor),
method_list = [ HierLat(latent_factor=lafactor, cold_start = HierLat.CS_EQUAL_PROB), \
LMaFit(latent_factor=lafactor), NMF(latent_factor=lafactor),
PMF(latent_factor=lafactor), TriUHV(latent_factor=lafactor) ]
# main method.
result = experiment_coldstart_map(exp_name, daily_data_file, \
min_occ_user, min_occ_prog, num_user, num_prog,\
method_list, leave_k_out, total_iteration, max_rank, binary = False)
matlab_output = {}
for method_name, method_iter_perf in result.items():
print 'Method: ' + method_name
rmse = sum(x['RMSE'] for x in method_iter_perf) / len(method_iter_perf)
print '>>Average RMSE : %.5f' % rmse
perf_recall = np.zeros(len(method_iter_perf[0]['recall']))
exp_name = 'test_exp_mid_prec_rec'; # something meaningful.
# filtering criteria
min_occ_user = 35;
min_occ_prog = 300;
top_n = 50; # performance computed on top N;
leave_k_out = 10; # perform leave k out.
# number of repetitions.
total_iteration = 3;
# recommendation algorithms
method_list = [ LMaFit(latent_factor=lafactor), RandUV(latent_factor=lafactor), \
HierLat(latent_factor=lafactor), NMF(latent_factor=lafactor),
PMF(latent_factor=lafactor), TriUHV(latent_factor=lafactor) ];
# main method.
result = experiment_leave_k_out(exp_name, daily_data_file, min_occ_user, min_occ_prog, \
method_list, leave_k_out, total_iteration, top_n);
# display results (average RMSE).
for method_name, method_iter_perf in result.items():
print 'Method: '+ method_name;
print '>>Average precision : %.5f' % (sum( x['prec'] for x in method_iter_perf)/len(method_iter_perf));
print '>>Average recall : %.5f' % (sum( x['recall'] for x in method_iter_perf)/len(method_iter_perf));
print '>>Average rmse : %.5f' % (sum( x['rmse'] for x in method_iter_perf)/len(method_iter_perf));
#print method_iter_perf;
#print result;
#feedback_data = reader.read_file_with_minval(filename, 25, 300);
feedback_data = reader.read_file_with_minval(filename, 35, 300);
print feedback_data;
print 'Maximum Genre.'
print np.max(feedback_data.meta['pggr_gr']) + 1;
print 'Normalizing data.'
feedback_data.normalize_row();
# build model with 3 latent factors.
r = 5;
# the L_2 norm regularizer
lamb = 0.001;
# the stopping delta value
delta = 0.01;
# the maximum iteration number
maxiter = 500;
HierLat_model = HierLat(r,lamb,delta,maxiter, verbose = True);
#HierLat_model.train(feedback_data, simplex_projection = False);
HierLat_model.train(feedback_data, simplex_projection = True);
'''
# test.
loc_row = [200, 4, 105];
loc_col = [ 10, 22, 4];
print 'Prediction:'
print HierLat_model.predict(loc_row, loc_col);
'''
# filtering criteria.
min_occ_user = 50
min_occ_prog = 1000
num_user = 10000
num_prog = 3000
total_iteration = 2
iteration = 1
# iteration out of total_iteration.
leave_k_out = 20
lafactor = 5
method = HierLat(latent_factor=lafactor)
hash_file_str = str(hash(tuple(daily_data_file)))
reader = DailyWatchTimeReader()
feedback_data = reader.read_file_with_minval(daily_data_file, min_occ_user,
min_occ_prog, num_user,
num_prog)
exp_id = 'lko_bi_' + exp_name + '_data' + hash_file_str\
+ '_mu' + str(min_occ_user) + '_mp' + str(min_occ_prog) \
+ '_nu' + str(num_user) + '_np' + str(num_prog) \
+ '_k' + str(leave_k_out) + '_toiter' + str(total_iteration)
result_resource_str = 'exp' + exp_id + \
'_method' + method.unique_str() + \
'_iter' + str(iteration)
# filtering criteria.
min_occ_user = 50;
min_occ_prog = 1000;
num_user = 10000;
num_prog = 3000;
total_iteration = 2;
iteration = 1; # iteration out of total_iteration.
leave_k_out = 20;
lafactor = 5;
method = HierLat(latent_factor=lafactor);
hash_file_str = str(hash(tuple(daily_data_file)));
reader = DailyWatchTimeReader();
feedback_data = reader.read_file_with_minval(daily_data_file, min_occ_user, min_occ_prog, num_user, num_prog);
exp_id = 'lko_bi_' + exp_name + '_data' + hash_file_str\
+ '_mu' + str(min_occ_user) + '_mp' + str(min_occ_prog) \
+ '_nu' + str(num_user) + '_np' + str(num_prog) \
+ '_k' + str(leave_k_out) + '_toiter' + str(total_iteration);
result_resource_str = 'exp' + exp_id + \
'_method' + method.unique_str() + \
'_iter' + str(iteration);
sub_folder = exp_id + '/models/' + method.unique_str(); # use a sub folder to store the experiment resource.
# filtering criteria
min_occ_user = 4;
min_occ_prog = 1;
top_n = 15; # performance computed on top N;
leave_k_out = 1; # perform leave k out.
# number of repetitions.
total_iteration = 3;
# latent factor
lf = 5;
# recommendation algorithms
method_list = [ LMaFit(latent_factor = 5), RandUV(latent_factor = 5), \
HierLat(latent_factor = 5) , NMF(latent_factor = 5) ];
# main method.
result = experiment_leave_k_out(exp_name, daily_data_file, min_occ_user, min_occ_prog, \
method_list, leave_k_out, total_iteration, top_n);
# display results (average RMSE).
for method_name, method_iter_perf in result.items():
print 'Method: '+ method_name;
print '>>Average precision : %.5f' % (sum( x['prec'] for x in method_iter_perf)/len(method_iter_perf));
print '>>Average recall : %.5f' % (sum( x['recall'] for x in method_iter_perf)/len(method_iter_perf));
print '>>Average rmse : %.5f' % (sum( x['rmse'] for x in method_iter_perf)/len(method_iter_perf));
#print method_iter_perf;
#print result;
# filtering criteria
min_occ_user = 35
min_occ_prog = 300
# specify the percentage of training and (1 - training_prec) is testing.
training_prec = 0.5
# number of repetitions.
total_iteration = 5
# latent factor
lf = 10
# recommendation algorithms
method_list = [
LMaFit(latent_factor=lf),
RandUV(latent_factor=lf),
HierLat(latent_factor=lf)
]
# main method.
result = experiment_rand_split(exp_name, daily_data_file, min_occ_user, min_occ_prog, \
method_list, training_prec, total_iteration)
# display results (average RMSE).
for method_name, method_iter_perf in result.items():
print 'Method: ' + method_name
print '>>Average performance RMSE: %.5f' % (
sum(x for x in method_iter_perf) / len(method_iter_perf))
#print result;
