本文整理汇总了Python中sklearn.ensemble.GradientBoostingRegressor.predict_proba方法的典型用法代码示例。如果您正苦于以下问题：Python GradientBoostingRegressor.predict_proba方法的具体用法？Python GradientBoostingRegressor.predict_proba怎么用？Python GradientBoostingRegressor.predict_proba使用的例子？那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类sklearn.ensemble.GradientBoostingRegressor的用法示例。
在下文中一共展示了GradientBoostingRegressor.predict_proba方法的1个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Python代码示例。
示例1: main
# 需要导入模块: from sklearn.ensemble import GradientBoostingRegressor [as 别名]
# 或者: from sklearn.ensemble.GradientBoostingRegressor import predict_proba [as 别名]
def main(options, args):
if options.utildate == None:
assert(False)
fsampleY = open(options.input + "/" + options.utildate + "/features.csv", "r")
l_X = []
l_y = []
for line in fsampleY:
tokens = line.split(",")
features = []
for i in range(len(tokens)):
if i < len(tokens)-1:
features.append(float(tokens[i]))
else:
l_y.append(int(tokens[i]))
l_X.append(features)
print "features loaded!"
X = np.array(l_X)
y = np.array(l_y)
assert(X.shape[0] == y.shape[0])
if int(options.short) > 0:
print "using short data for test purpose"
X = X[0:int(options.short)]
y = y[0:int(options.short)]
print "preparing models"
trainModel = globals()[options.trainmodel]()
print trainModel
if options.isregress == True:
if options.trainmodel == "Gdbc1":
model_predictor = GradientBoostingRegressor(max_features=0.6, learning_rate = 0.05, max_depth=5, n_estimators=300)
else:
model_predictor = trainModel.get_model()
else :
model_predictor = GradientBoostingClassifier(max_features=0.6, learning_rate=0.05, max_depth=5, n_estimators=300)
#model_predictor = GradientBoostingClassifier()
X_train, X_test, y_train, y_test = cross_validation.train_test_split(X, y, test_size=0.3, random_state=0)
# print cross_validation.cross_val_score(model_predictor, X, y)
clf = model_predictor.fit(X_train, y_train)
if options.isregress:
if options.trainmodel != "God":
pred = model_predictor.predict(X_test)
r2 = r2_score(y_test, pred)
print "the r2 score is ", r2
else:
pred = model_predictor.predict_proba(X_test)
# calculate the R2score
# tpred = model_predictor.predict(X_test)
# score = model_predictor.score(X_test, tpred)
# print "score=", score
#assert(len(pred) == X_test.shape[0])
#{{{ prediction
print "prediction ..."
stock_predict_dir = options.output + "/"+ options.trainmodel +"/" + options.utildate
if not os.path.exists(stock_predict_dir) : os.makedirs(stock_predict_dir)
stock_predict_out = file(stock_predict_dir + "/predicted.csv", "w")
metastr = file(options.input + "/" + options.utildate + "/meta.json", "r").readlines()[0]
metajson = json.loads(metastr)
for line in file(options.input + "/" + options.utildate + "/last.csv", "r"):
tokens = line.split(",")
l_features = []
for i in range(len(tokens)):
if 0 == i:
print >> stock_predict_out, "%s," % tokens[i],
elif 1 == i:
print >> stock_predict_out, "%s," % tokens[i],
print >> stock_predict_out, "%d," % metajson["span"],
else:
l_features.append(float(tokens[i].strip()))
l_features2 = []
l_features2.append(l_features)
np_features = np.array(l_features2)
if np_features.shape[1] != X.shape[1] :
assert(false)
if options.isregress:
if options.trainmodel == "God":
pred = model_predictor.predict(np_features, tokens[0], tokens[1], metajson["span"])
else:
pred = model_predictor.predict(np_features)
print >> stock_predict_out, "%f" % pred
else:
pred = model_predictor.predict_proba(np_features)
print >> stock_predict_out, "%f" % pred[0,1]
stock_predict_out.close()