使用R中的计分卡时出错（“数据”必须是向量类型，为“ NULL”）

Question

我正在使用python中的库 rpy2 将 R 包应用于熊猫数据框。

我想从包 记分卡 （来自 R < / em>）到熊猫datframe，但是当我遇到错误并且不知道为什么

这是我的代码：

# python
import pandas as pd
import numpy as np
import rpy2
from rpy2.robjects.packages import importr
from rpy2.robjects import pandas2ri
from rpy2.robjects.vectors import DataFrame

# R
base = importr('base')
score = importr("scorecard")

# Create pandas df
df = pd.DataFrame( np.random.randn(5,4), # 5 rows, 2 columns
               columns = ["A","B","C","D"], # name of columns
               index = ["Max", "Nathy", "Tom", "Joe", "Kathy"] )
df["C"] = [0,0,1,0,1] # "BGI"

pandas2ri.activate()
# Convert pandas to r
df_r = pandas2ri.py2ri(df)
df_r = base.as_data_frame(df_r)
print(type(df_r))
pandas2ri.deactivate()

bins = score.woebin(df_r, 
                    y = "C",
                    x = base.c("A","B") )

在las命令中出现以下错误

Error in matrix(unlist(value, recursive = FALSE, use.names = FALSE), nrow = nr,  : 
  'data' must be of a vector type, was 'NULL'

Answer 1

这是pyper

的一个选项

import pandas as pd
import numpy as np
from pyper import *

df = pd.DataFrame( np.random.randn(5,4), # 5 rows, 2 columns
               columns = ["A","B","C","D"], # name of columns
               index = ["Max", "Nathy", "Tom", "Joe", "Kathy"] )
df["C"] = [0,0,1,0,1]


r=R(use_pandas=True)

r.assign("df_r", df)  
r("library(scorecard)")

r('bins <- woebin(df_r, y = "C", c("A", "B"))')

binsN = r.get('bins')

-检查输出

print(binsN)
#{'A':   variable                    bin   count    count_distr    good    bad   #\
#0     b'A'  b'[-Inf,1.777599442)'        3            0.6       2      1   
#1     b'A'  b'[1.777599442, Inf)'        2            0.4       1      1   

#    badprob       woe    bin_iv    total_iv           breaks  \
#0   0.333333 -0.287682  0.047947    0.115525  b'1.777599442'   
#1   0.500000  0.405465  0.067578    0.115525          b'Inf'   

#    is_special_values   
#0                False  
#1                False  , 'B':   variable                     bin   count    #count_distr    good    bad   \
#0     b'B'  b'[-Inf,0.2711706509)'        3            0.6       2      1   
#1     b'B'  b'[0.2711706509, Inf)'        2            0.4       1      1   

#    badprob       woe    bin_iv    total_iv            breaks  \
#0   0.333333 -0.287682  0.047947    0.115525  b'0.2711706509'   
#1   0.500000  0.405465  0.067578    0.115525           b'Inf'   

#    is_special_values   
#0                False  
#1                False  }

---

这也可以通过R和python来获取reticulate个对象来完成。创建了一个python脚本（'pytmp.py'）

#pytmp.py

import pandas as pd
import numpy as np


df = pd.DataFrame( np.random.randn(5,4), # 5 rows, 2 columns
               columns = ["A","B","C","D"], # name of columns
               index = ["Max", "Nathy", "Tom", "Joe", "Kathy"] )
df["C"] = [0,0,1,0,1] # "BGI"

df

-在R

中调用它

library(reticulate)
library(scorecard)
use_python("/usr/local/bin/python")
use_virtualenv("~/r-reticulate")

source_python("pytmp.py")
bins <- woebin(df, y = "C", x = c("A","B") )
bins
#$A
#   variable                bin count count_distr good bad   badprob        woe     bin_iv  total_iv      breaks is_special_values
#1:        A [-Inf,0.895928754)     3         0.6    2   1 0.3333333 -0.2876821 0.04794701 0.1155245 0.895928754             FALSE
#2:        A [0.895928754, Inf)     2         0.4    1   1 0.5000000  0.4054651 0.06757752 0.1155245         Inf             FALSE

#$B
#   variable                 bin count count_distr good bad   badprob        woe     bin_iv  total_iv       breaks is_special_values
#1:        B [-Inf,0.2356073663)     3         0.6    2   1 0.3333333 -0.2876821 0.04794701 0.1155245 0.2356073663             FALSE
#2:        B [0.2356073663, Inf)     2         0.4    1   1 0.5000000  0.4054651 0.06757752 0.1155245          Inf             FALSE

注意：我们没有设置种子，因此每次运行的值都不同