训练一个机器学习模型的基本步骤

本小白的学习机器学习实战这本书时，按照书中的步骤总结了一下简单训练一个机器学习的基本步骤。

1.导入数据

2.将数据划分成训练集和测试集

3.训练机器学习模型

4.对模型进行评估

5.进行预测并对测试集评估

from sklearn.datasets import fetch_california_housing # 引用相关数据库
from sklearn.model_selection import train_test_split,RandomizedSearchCV,GridSearchCV # 引入随机搜索和网格搜索
from sklearn.preprocessing import StandardScaler # 引用标准化库

# 调用数据库
housing = fetch_california_housing()
x = housing['data']
y = housing['target']

# 分好训练集和测试集
x_train,x_test,y_train,y_test = train_test_split(x,y,test_size=0.1,random_state=42)

# 标准化数据（非必要）
scaler = StandardScaler()
x_train_scaled = scaler.fit_transform(x_train)
x_test_scaled = scaler.fit_transform(x_test)

# 导入相关的机器学习模型,这里使用的时线性SVM
from sklearn.svm import LinearSVR
lin_svr = LinearSVR(random_state=42)
lin_svr.fit(x_train_scaled,y_train)

# 计算得分,评估模型
from sklearn.metrics import mean_squared_error,accuracy_score
# 一般分类用accuracy_score(),回归用mean_squared_error
y_pred = lin_svr.predict(x_test_scaled)
mse = mean_squared_error(y_train,y_pred)
# print(mse)
ac_scores = accuracy_score(y_train,y_pred)

# 线性结果不理想，下面使用非线性方法调整超参数，有随机搜索和网格搜索两种，RandomizedSearchCV,GridSearchCV
from sklearn.svm import SVR
from scipy.stats import reciprocal,uniform

# 参数集
param_distributions = {'gamma':reciprocal(0.001,0.1),'C':uniform(1,10)}
# 随机搜索
rnd_search_cv = RandomizedSearchCV(SVR(),param_distributions,n_iter=10,verbose=2,cv=3,random_state=42)
rnd_search_cv.fit(x_train_scaled,y_train)

# 输出最好的参数组合
print(rnd_search_cv.best_estimator_)
# 使用最好的参数进行训练
y_pred = rnd_search_cv.best_estimator_.predict(x_train_scaled)
# 再次评估模型
mse_train = mean_squared_error(y_train,y_pred)
# ac_score = accuracy_score(y_train,y_pred)
# 预测并评估
# y_pred = rnd_search_cv.best_estimator_.predict(x_test_scaled)
# mse_test = mean_squared_error(y_test,y_pred)
# ac_scores = accuracy_score(y_test,y_pred)

以上就是俺这个小白的学习心得，请各路大神指正。