Question

我试图使用带有黑色星期五数据集的蟒蛇来预测购买量这是我的代码

    train=pd.read_csv("C:\\Users\\User\\Documents\\data sets\\train.csv")
    test=pd.read_csv("C:\\Users\\User\\Documents\\data sets\\test.csv")
    import numpy as np
    frames=[train,test]
    data=pd.concat(frames)
    print(data.shape)
    data.head()
    data.isnull().any()
    data.fillna(999,inplace=True)
    data.head(20)
    data.Age[data["Age"]=="0-17"]="15"
    data["Age"].head(10)
    data.Age[data["Age"]=="18-25"]="21"
    data.Age[data["Age"]=="26-35"]="30"
    data.Age[data["Age"]=="36-45"]="40"
    data.Age[data["Age"]=="46-50"]="48"
    data.Age[data["Age"]=="51-55"]="53"
    data.Age[data["Age"]=="55+"]="60"
   data.Stay_In_Current_City_Years[data["Stay_In_Current_City_Years"]=="4+"]
   ="4"
   data["Age"]=data["Age"].astype(int)
   data["Stay_In_Current_City_Years"]=data["Stay_In_Current_City_Years"].
   astype(int)
   data.dtypes
   data["Marital_Status"]=data["Marital_Status"].astype(int)
   data["Occupation"]=data["Occupation"].astype(int)
   data["Product_Category_1"]=data["Product_Category_1"].astype(int)
   data["Product_Category_1"]=data["Product_Category_1"].astype(int)
   data["Product_Category_2"]=data["Product_Category_2"].astype(float)
   data["Product_Category_3"]=data["Product_Category_3"].astype(float)
   data["Purchase"]=data["Purchase"].astype(float)
   sex=pd.get_dummies(data["Gender"]).iloc[:,1:]
   data1=pd.concat([data,sex],axis=1)
   city=pd.get_dummies(data["City_Category"]).iloc[:,1:]
   data1=pd.concat([data,sex,city],axis=1)
   # cross validation and creating the features and the target variable 
   from sklearn.cross_validation import train_test_split
   y=data1["Purchase"]
   x=data1[["Age","City_Category","Gender","Marital_Status","Occupation",
 "Product_Category_1","Product_Category_2","Product_Category_3","Product_ID"
  ,"Stay_In_Current_City_Years","User_ID","M","B","C"]]
  x_train,x_test,y_train,y_test=train_test_split(x,y)
   # building the regration
   from sklearn import linear_model
   reg=linear_model.LinearRegression()
   reg.fit(x_train,y_train)

但是我继续得到这个：

    ValueError: could not convert string to float: 'P00100642'

是什么意思？我还需要转换为整数才能运行回归吗？我该如何解决？谢谢：）

Answer 1

机器学习算法仅采用数字数据。 Purchase_ID列没有数字数据，因为它以' P '开头。您正在尝试通过它，因为这样会导致错误。

注意值中的模式，您将看到每个条目均以“ P00”开头。由于它是一个字符串，因此可以用 nothing 代替它。

尝试一下：

data['Product_ID'] = data['Product_ID'].str.replace('P00', '')

之后，您可以使用StandardScaler缩小值。

黑色星期五数据集与Anaconda的LinearRegression

1 个答案: