如何使用TimeSeriesSplit和GridSearchCV对象在scikit-learn中调整模型?
问题描述:
我已经搜索了sklearn docs for TimeSeriesSplit
和docs for cross-validation,但我一直未能找到一个工作示例。如何使用TimeSeriesSplit和GridSearchCV对象在scikit-learn中调整模型?
我正在使用sklearn版本0.19。
这是我的设置
import xgboost as xgb
from sklearn.model_selection import TimeSeriesSplit
from sklearn.grid_search import GridSearchCV
import numpy as np
X = np.array([[4, 5, 6, 1, 0, 2], [3.1, 3.5, 1.0, 2.1, 8.3, 1.1]]).T
y = np.array([1, 6, 7, 1, 2, 3])
tscv = TimeSeriesSplit(n_splits=2)
for train, test in tscv.split(X):
print(train, test)
给出:
[0 1] [2 3]
[0 1 2 3] [4 5]
如果我尝试:
model = xgb.XGBRegressor()
param_search = {'max_depth' : [3, 5]}
my_cv = TimeSeriesSplit(n_splits=2).split(X)
gsearch = GridSearchCV(estimator=model, cv=my_cv,
param_grid=param_search)
gsearch.fit(X, y)
它给:TypeError: object of type 'generator' has no len()
我得到的问题:GridSearchCV
是Ť正在调用len(cv)
,但my_cv
是一个没有长度的迭代器。然而,docs for GridSearchCV
状态,我可以用一个
INT,交叉验证发电机或迭代,可选
我试着用TimeSeriesSplit
没有.split(X)
,但它仍然没有奏效。
我确定我忽略了一些简单的东西,谢谢!
答
事实证明,问题是我使用sklearn.grid_search
的GridSearchCV
,这已被弃用。从sklearn.model_selection
导入GridSearchCV
解决了这个问题:
import xgboost as xgb
from sklearn.model_selection import TimeSeriesSplit, GridSearchCV
import numpy as np
X = np.array([[4, 5, 6, 1, 0, 2], [3.1, 3.5, 1.0, 2.1, 8.3, 1.1]]).T
y = np.array([1, 6, 7, 1, 2, 3])
tscv = TimeSeriesSplit(n_splits=2)
model = xgb.XGBRegressor()
param_search = {'max_depth' : [3, 5]}
my_cv = TimeSeriesSplit(n_splits=2).split(X)
gsearch = GridSearchCV(estimator=model, cv=my_cv,
param_grid=param_search)
gsearch.fit(X, y)
给出:
GridSearchCV(cv=<generator object TimeSeriesSplit.split at 0x11ab4abf8>,
error_score='raise',
estimator=XGBRegressor(base_score=0.5, colsample_bylevel=1, colsample_bytree=1, gamma=0,
learning_rate=0.1, max_delta_step=0, max_depth=3,
min_child_weight=1, missing=None, n_estimators=100, nthread=-1,
objective='reg:linear', reg_alpha=0, reg_lambda=1,
scale_pos_weight=1, seed=0, silent=True, subsample=1),
fit_params=None, iid=True, n_jobs=1,
param_grid={'max_depth': [3, 5]}, pre_dispatch='2*n_jobs',
refit=True, return_train_score=True, scoring=None, verbose=0)
尝试使用'my_cv = [(火车,测试),用于火车,测试在TimeSeriesSplit(n_splits = 2).split(X) ]' –
这工作,谢谢!但是,该函数不应该与迭代器一起工作吗?当观察次数很多时(如果折叠次数较多,我会更糟糕),我宁愿不要在内存中保存那些大阵列 – cd98
是的。你应该在scikit-learn github页面上发布一个问题。 –