GridSearchCV(sklearn) 中的多个估计器 [英] More than one estimator in GridSearchCV(sklearn)

问题描述

我正在查看有关 GridSearchCV 的 sklearn 文档网页.GridSearchCV 对象的属性之一是 best_estimator_.所以这是我的问题.如何将多个估计量传递给 GSCV 对象?

I was checking sklearn documentation webpage about GridSearchCV. One of attributes of GridSearchCV object is best_estimator_. So here is my question. How to pass more than one estimator to GSCV object?

使用如下字典:{'SVC()':{'C':10, 'gamma':0.01}, 'DecTreeClass()':{....}}?

推荐答案

GridSearchCV 适用于参数.它将使用 param_grid 中指定的不同参数组合训练多个估计器(但相同的类(SVC 或 DecisionTreeClassifier 或其他分类器之一).best_estimator_ 是执行最好的数据.

GridSearchCV works on parameters. It will train multiple estimators (but same class (one of SVC, or DecisionTreeClassifier, or other classifiers) with different parameter combinations from specified in param_grid. best_estimator_ is the estimator which performs best on the data.

因此本质上 best_estimator_ 是使用最佳找到的参数初始化的同一个类对象.

So essentially best_estimator_ is the same class object initialized with best found params.

因此在基本设置中，您不能在网格搜索中使用多个估计器.

So in the basic setup you cannot use multiple estimators in the grid-search.

但作为一种解决方法，当使用估计器是 GridSearchCV 可以设置的 "parameter" 的管道时，您可以有多个估计器.

But as a workaround, you can have multiple estimators when using a pipeline in which the estimator is a "parameter" which the GridSearchCV can set.

像这样:

from sklearn.pipeline import Pipeline
from sklearn.svm import SVC
from sklearn.tree import DecisionTreeClassifier
from sklearn.model_selection import GridSearchCV
from sklearn.datasets import load_iris
iris_data = load_iris()
X, y = iris_data.data, iris_data.target


# Just initialize the pipeline with any estimator you like    
pipe = Pipeline(steps=[('estimator', SVC())])

# Add a dict of estimator and estimator related parameters in this list
params_grid = [{
                'estimator':[SVC()],
                'estimator__C': [1, 10, 100, 1000],
                'estimator__gamma': [0.001, 0.0001],
                },
                {
                'estimator': [DecisionTreeClassifier()],
                'estimator__max_depth': [1,2,3,4,5],
                'estimator__max_features': [None, "auto", "sqrt", "log2"],
                },
               # {'estimator':[Any_other_estimator_you_want],
               #  'estimator__valid_param_of_your_estimator':[valid_values]

              ]

grid = GridSearchCV(pipe, params_grid)

您可以根据需要在 params_grid 列表中添加任意数量的 dict，但请确保每个 dict 都具有与 'estimator' 相关的兼容参数.

You can add as many dicts inside the list of params_grid as you like, but make sure that each dict have compatible parameters related to the 'estimator'.

这篇关于GridSearchCV(sklearn) 中的多个估计器的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！