使用了RamdomedSearchCV迭代100次,从参数组里面选择出当前最佳的参数组合
在RamdomedSearchCV的基础上,使用GridSearchCV在上面最佳参数的周围选择一些合适的参数组合,进行参数的微调
1. RandomedSearchCV(estimator=rf, param_distributions=param_random, cv=3, verbose=2,random_state=42, n_iter=100) # 随机选择参数组合
参数说明:estimator使用的模型, param_distributions表示待选的参数组合,cv表示交叉验证的次数,verbose表示打印的详细程度,random_state表示随机种子, n_iter迭代的次数
2.GridSearchCV(estimator = rf, param_grid=grid_param, cv=3, verbose=2)
参数说明:estimator使用的模型, param_grid 待选择的参数组合, cv交叉验证的次数,verbose打印的详细程度
3. pprint(rf.get_params())
参数说明:pprint按顺序进行打印, rf.get_params() 表示获得随机森林模型的当前输入参数
代码:
第一步:导入数据
第二步:对数据的文本标签进行one-hot编码
第三步:提取特征和标签
第四步:使用train_test_split将数据分为训练集和测试集
第五步:构建随机森林训练集进行训练
第六步:获得模型特征重要性进行排序,选取前5重要性的特征rf.feature_importances_
第七步:重新构建随机森林的模型
第八步:使用RandomedSearchCV() 进行参数组的随机选择
第九步:根据获得的参数组,使用GridSearchCV() 进行参数组附近的选择,从而对参数组进行微调
import pandas as pd import numpy as np import matplotlib.pyplot as plt from sklearn.model_selection import train_test_split from sklearn.ensemble import RandomForestRegressor import time # 第一步读取数据 data = pd.read_csv('data/temps_extended.csv') # 第二步:对文本标签使用one-hot编码 data = pd.get_dummies(data) # 第三步:提取特征和标签 X = data.drop('actual', axis=1) feature_names = np.array(X.columns) y = np.array(data['actual']) X = np.array(X) # 第四步:使用train_test_split进行样本的拆分 train_x, test_x, train_y, test_y = train_test_split(X, y, test_size=0.3, random_state=42) # 第五步:建立模型和预测 rf = RandomForestRegressor(random_state=42, n_estimators=1000) rf.fit(train_x, train_y) pre_y = rf.predict(test_x) # MSE mse = round(abs(pre_y - test_y).mean(), 2) error = abs(pre_y - test_y).mean() # MAPE mape = round(((1 - abs(pre_y - test_y) / test_y)*100).mean(), 2) print(mse, mape) # 第六步:选取特征重要性加和达到95%的特征 # 获得特征重要性的得分 feature_importances = rf.feature_importances_ # 将特征重要性得分和特征名进行组合 feature_importances_names = [(feature_name, feature_importance) for feature_name, feature_importance in zip(feature_names, feature_importances)] # 对特征重要性进行按照特征得分进行排序 feature_importances_names = sorted(feature_importances_names, key=lambda x: x[1], reverse=True) # 获得排序后的特征名 feature_importances_n = [x[0] for x in feature_importances_names] # 获得排序后的特征重要性得分 feature_importances_v = [x[1] for x in feature_importances_names] feature_importances_v_add = np.cumsum(feature_importances_v) little_feature_name = feature_importances_n[:np.where([feature_importances_v_add > 0.95])[1][0]+1] # 第七步:选择重要性前5的特征重新建立模型 X = data[little_feature_name].values y = data['actual'].values # 使用train_test_split进行样本的拆分 train_x, test_x, train_y, test_y = train_test_split(X, y, test_size=0.3, random_state=42) rf = RandomForestRegressor(random_state=42, n_estimators=1000) # 第八步:使用RandomizedSearchCV随机选择参数组合 # 使用pprint打印rf的参数 from pprint import pprint pprint(rf.get_params()) from sklearn.model_selection import RandomizedSearchCV #树的个数 n_estimators = [int(x) for x in range(200, 2000, 100)] min_samples_leaf = [2, 4, 6] min_samples_split = [1, 2, 4] max_features = ['auto', 'sqrt'] bootstrap = [True, False] max_depth = [int(x) for x in range(10, 100, 10)] param_random = { 'n_estimators': n_estimators, 'max_depth': max_depth, 'max_features': max_features, 'min_samples_leaf': min_samples_leaf, 'min_samples_split': min_samples_split, 'bootstrap': bootstrap } rf = RandomForestRegressor() rf_random = RandomizedSearchCV(estimator=rf, param_distributions=param_random, cv=3, verbose=2, random_state=42) rf_random.fit(train_x, train_y) # 获得最好的训练模型 best_estimator = rf_random.best_estimator_ # 定义用于计算误差和准确度的函数 def Calculation_accuracy(estimator, test_x, test_y): pre_y = estimator.predict(test_x) error = abs(pre_y - test_y).mean() accuraccy = ((1 - abs(pre_y - test_y)/test_y)*100).mean() return error, accuraccy # 计算损失值和准确度 error, accuraccy = Calculation_accuracy(best_estimator, test_x, test_y) print(error, accuraccy) # 打印最好的参数组合 print(rf_random.best_params_) # 最好的参数组合 {'n_estimators': 800, 'min_samples_split': 4, 'min_samples_leaf': 4, 'max_features': 'auto', # 'max_depth': 10, 'bootstrap': 'True'} # 第九步:根据RandomizedSearchCV获得参数,使用GridSearchCV进行参数的微调 from sklearn.model_selection import GridSearchCV n_estimators = [600, 800, 1000] min_samples_split = [4] min_samples_leaf = [4] max_depth = [8, 10, 12] grid_param = { 'n_estimators': n_estimators, 'min_samples_split': min_samples_split, 'min_samples_leaf': min_samples_leaf, 'max_depth': max_depth } rf = RandomForestRegressor() rf_grid = GridSearchCV(rf, param_grid=grid_param, cv=3, verbose=2) rf_grid.fit(train_x, train_y) best_estimator = rf_grid.best_estimator_ error, accuraccy = Calculation_accuracy(best_estimator, test_x, test_y) print(error, accuraccy) print(rf_grid.best_params_) # {'max_depth': 8, 'min_samples_leaf': 4, 'min_samples_split': 4, 'n_estimators': 1000}