def predictMan(res):
try:
X, _ = load_svmlight_file(res.data_file, n_features=res.nfeat)
except Exception:
X, _ = load_svmlight_file(res.data_file)
X = X.toarray()
if res.meth[0] == 'Classification':
ens = EnsembleSelectionClassifier(db_file=res.db_file, models=None)
elif res.meth[0] == 'Regression':
ens = EnsembleSelectionRegressor(db_file=res.db_file, models=None)
else:
msg = "Invalid method passed (-T does not conform to ['Regression','Classification']"
raise ValueError(msg)
if (res.pred_src == 'best'):
preds = ens.best_model_predict_proba(X)
else:
preds = ens.predict_proba(X)
if res.meth[0] == 'Classification':
if (not res.return_probs):
preds = np.argmax(preds, axis=1)
for p in preds:
if (res.return_probs):
mesg = " ".join(["%.5f" % v for v in p])
else:
mesg = p
print(str(mesg))
return preds
action='store_true',
default=False,
help='predict probabilities')
return parser.parse_args()
if (__name__ == '__main__'):
res = parse_args()
X, _ = load_svmlight_file(res.data_file)
X = X.toarray()
ens = EnsembleSelectionClassifier(db_file=res.db_file, models=None)
if (res.pred_src == 'best'):
preds = ens.best_model_predict_proba(X)
else:
preds = ens.predict_proba(X)
if (not res.return_probs):
preds = np.argmax(preds, axis=1)
for p in preds:
if (res.return_probs):
mesg = " ".join(["%.8f" % v for v in p])
else:
mesg = p
print(mesg)
parser.add_argument('-p', dest='return_probs',
action='store_true', default=False,
help='predict probabilities')
return parser.parse_args()
if (__name__ == '__main__'):
res = parse_args()
X, _ = load_svmlight_file(res.data_file)
X = X.toarray()
ens = EnsembleSelectionClassifier(db_file=res.db_file, models=None)
if (res.pred_src == 'best'):
preds = ens.best_model_predict_proba(X)
else:
preds = ens.predict_proba(X)
if (not res.return_probs):
preds = np.argmax(preds, axis=1)
for p in preds:
if (res.return_probs):
mesg = " ".join(["%.5f" % v for v in p])
else:
mesg = p
print(mesg)