将特征向量传递给sklearn的Logistic回归函数

Question

我正在使用LogisticRegression对kaggle.com的泰坦尼克号问题进行建模。我想使用多个变量，如年龄，性别等来模拟我的S型函数。如果只与Sex这样的一个变量一起使用，同样的方法也可以很好地工作，但当与多个变量一起使用时，它会抛出以下错误

TypeError:foat（）参数必须是字符串或数字，而不是"method"

我的猜测是我没有正确使用整形方法。附言：我是python和sklearn库的初学者。请对我宽容一点。

import pandas as pd
from sklearn.linear_model import LogisticRegression
import numpy as np

df = pd.read_csv(r'C:UsersabhiDownloadstrain.csv')
df.Age = df.Age.fillna(df.Age.mean)
df.Embarked = df.Embarked.fillna(df.Embarked.median)
x1 = df.Pclass
x2 = df.Sex
for i in range(len(x2)):
    if x2[i]=='male':
        x2[i]=1
    else: 
        x2[i]=0
#female,male 0,1
x3 = df.Age
x4 = df.SibSp
x5 = df.Parch
x6 = df.Ticket
x7 = df.Fare
x9 = df.Embarked
for i in range(len(x9)):
  if x9[i]=='C':
      x9[i]=0
  elif x9[i]=='Q': 
      x9[i]=1
  else :x9[i]=2
# C,Q,S = 0,1,2
# Creating a feature vector of multiple vectors
i2 = pd.DataFrame()
i2['Pclass'] = x1
i2['Sex'] = x2
i2['Age'] = x3
i2['SibSp'] = x4
i2['Parch'] = x5
i2['Fare'] = x7
i2['Embarked'] = x9
i2 = np.array(i2)
i2 = i2.reshape(-1,1)
ytrain = df.Survived
ytrain = np.array(ytrain)
ytrain = ytrain.reshape(-1,1)
c1 = LogisticRegression(penalty='l2',solver='liblinear')
c1.fit(i2,ytrain,sample_weight=None)
c1.score(i2,ytrain,sample_weight=None)

Answer 1

您能在删除此行的情况下运行代码吗？i2 = i2.reshape(-1,1)

将i2重塑为(-1,1)将i2变换为具有i2中的元素总数的长度的一维阵列。这可能不是你想要做的。