'eval_metric' : 'auc'
}
num_round = 100
kf = cross_validation.KFold(len(y), n_folds=4, random_state=33)
for train_index, test_index in kf:
X_train, X_test = X[train_index], X[test_index]
y_train, y_test = y[train_index], y[test_index]
dtrain = xgb.DMatrix(X_train, y_train)
dtest = xgb.DMatrix(X_test, y_test)
watchlist = [(dtest,'eval'), (dtrain,'train')]
clf = xgb.train(param, dtrain, num_round, watchlist)
# check importance
importance = clf.get_fscore()
tuples = [(k, importance[k]) for k in importance]
tuples = sorted(tuples, key=lambda x: x[1], reverse=True)
labels, values = zip(*tuples)
print values
for i in range(len(values)):
if values[i] == 1:
print labels[i],
print
break
#preds = clf.predict(dtest)
