Python的lasso回归分析

秒速五厘米 2021-12-19 22:43 312阅读 0赞

## 加载调用函数包 ##

import numpy as np # 快速操作结构数组的工具
    import pandas
    import matplotlib.pyplot as plt  # 可视化绘制
    from sklearn.linear_model import Lasso,LassoCV,LassoLarsCV

## 数据读取 ##

data = pd.read_csv('C://Users//TD//Desktop//data0629.csv')

#### 相关系数查看 ####

data.corr()

选择变量相关性系数较好的，可以作为进一步参考分析。

## 归一化处理 ##

利用均值和标准差进行数据归一化处理，克服量纲的影响。

import numpy as np
    arr_mean = np.mean(data) #求均值
    arr_std = np.std(data,ddof=1) #求标准差
    newdata=(data-arr_mean )/arr_std

## 分组抽样划分训练集和测试集 ##

X = newdata[["年龄","性别","重点病种名称编码","住院总费用","科室名称编码","确诊天数","门诊与入院诊断符合情况","住院是否超30天","术前住院日","手术季度","手术日期与出院日时差天数","手术级别程度编码","是否重返医院","检验时长"]]
    Y = newdata[['住院天数']]  # 选取列名为y的作为Y.
    from sklearn.model_selection import train_test_split
    X_train,X_test,y_train,y_test = train_test_split(X,Y,random_state=42)

## 通过设置不同的alpha值建立5个lasso实例 ##

lasso = Lasso().fit(X_train,y_train)
    lasso001 =Lasso(alpha=0.01).fit(X_train,y_train)
    lasso005 = Lasso(alpha=0.05).fit(X_train,y_train)
    lasso00001 = Lasso(alpha=0.0001).fit(X_train,y_train)
    lasso05 = Lasso(alpha=0.5).fit(X_train,y_train)

#### 方法一 ####

print('**********************************')
    print("Lasso alpha=1")
    print ("training set score:{:.2f}".format(lasso.score(X_train,y_train)))
    print ("test set score:{:.2f}".format(lasso.score(X_test,y_test)))
    print ("Number of features used:{}".format(np.sum(lasso.coef_!=0)))

#### 方法二 ####

print('**********************************')
    print("Lasso alpha=0.01")
    print ("training set score:{:.2f}".format(lasso001.score(X_train,y_train)))
    print ("test set score:{:.2f}".format(lasso001.score(X_test,y_test)))
    print ("Number of features used:{}".format(np.sum(lasso001.coef_!=0)))

#### 方法三 ####

print('**********************************')
    print("Lasso alpha=0.0001")
    print ("training set score:{:.2f}".format(lasso00001.score(X_train,y_train)))
    print ("test set score:{:.2f}".format(lasso00001.score(X_test,y_test)))
    print ("Number of features used:{}".format(np.sum(lasso00001.coef_!=0)))

#### 方法四 ####

print('**********************************')
    print("Lasso alpha=0.05")
    print ("training set score:{:.2f}".format(lasso005.score(X_train,y_train)))
    print ("test set score:{:.2f}".format(lasso005.score(X_test,y_test)))
    print ("Number of features used:{}".format(np.sum(lasso005.coef_!=0)))

#### 方法五 ####

print('**********************************')
    print("Lasso alpha=0.5")
    print ("training set score:{:.2f}".format(lasso05.score(X_train,y_train)))
    print ("test set score:{:.2f}".format(lasso05.score(X_test,y_test)))
    print ("Number of features used:{}".format(np.sum(lasso05.coef_!=0)))
    
     ### 建立岭回归实例
    ridge01 = Ridge(alpha=0.01).fit(X_train,y_train)

#### 计算R方 ####

ridge01.score(X_train,y_train)

#### 打印整个模型 ####

print("住院天数 = " + str(a.intercept_[0]) + " + " + str(a.coef_[0][0]) + " * CRIM")
    住院天数 = 0.17721718179269352 + 0.005671216373405902 * CRIM

## 进行预测和比较 ##

output=ridge01.predict(X_test)
    from pandas.core.frame import DataFrame
    output1=DataFrame(output)
    import pandas
    outputdata=pandas.concat([output1,y_test]) # 直接合并数据框
    outputdata.to_csv('C:\\Users\\TD\\Desktop\\some_csv.csv', index=False)

最终比较预测数据和实际数据的结果是否一致，得出MSE等值情况。