Search code examples
pythonrrpy2

Cubist regression under rpy2: "subscript out of bounds" error

When I use rpy2 to do the Cubist regression.I met the error:

Error in strsplit(tmp, "\"")[[1]] : subscript out of bounds

I try to Use as.matrix to change the data format,but it's still unwork. 

import rpy2.robjects as robjects
from rpy2.robjects.packages import importr
from rpy2.robjects.vectors import FloatVector
from rpy2.robjects import pandas2ri
Cubist = importr('Cubist')
lattice = importr('lattice')
r = robjects.r
# 准备样点数据
dt = r('mtcars')
Z = FloatVector(dt[3])
X = FloatVector(dt[5])
X1 = FloatVector(dt[6])
T = r['cbind'](X,X1)

regr = r['cubist'](x=T,y=Z,committees=10)
Solution

  • If a matrix, the x argument to cubist() seems to require a dimnames attribute.

    Setup in R:

    library(Cubist)

dt = mtcars
Z = dt[, 4]
X = dt[, 6]
X1 = dt[, 7]

    Now compare this (reproduces your error):

    > T = cbind(dt[, 6], dt[, 7])
> str(T)
 num [1:32, 1:2] 2.62 2.88 2.32 3.21 3.44 ...
> cubist(x=T, y=Z, committees=10)
cubist code called exit with value 1
Error in strsplit(tmp, "\"")[[1]] : subscript out of bounds

    vs.

    > T = cbind(X, X1)
> str(T)
 num [1:32, 1:2] 2.62 2.88 2.32 3.21 3.44 ...
 - attr(*, "dimnames")=List of 2
  ..$ : NULL
  ..$ : chr [1:2] "X" "X1"
> cubist(x=T, y=Z, committees=10)

Call:
cubist.default(x = T, y = Z, committees = 10)

Number of samples: 32
Number of predictors: 2

Number of committees: 10
Number of rules per committee: 1, 1, 1, 1, 1, 1, 1, 1, 1, 1

    There are multiple ways to ensure the dimnames get attached via rpy2. One easy way with your code is simply to explicitly name the variables:

    In [15]: T = r['cbind'](X=X,X1=X1)

In [16]: print(r['str'](T))
 num [1:32, 1:2] 2.62 2.88 2.32 3.21 3.44 ...
 - attr(*, "dimnames")=List of 2
  ..$ : NULL
  ..$ : chr [1:2] "X" "X1"
<rpy2.rinterface.NULLType object at 0x7f0d7c0f5608> [RTYPES.NILSXP]

In [17]: print(r['cubist'](x=T,y=Z,committees=10))

Call:
cubist.default(x = structure(c(2.62, 2.875, 2.32, 3.215, 3.44, 3.46,
 205, 215, 230, 66, 52, 65, 97, 150, 150, 245, 175, 66, 91, 113, 264, 175,
 335, 109), committees = 10L)

Number of samples: 32
Number of predictors: 2

Number of committees: 10
Number of rules per committee: 1, 1, 1, 1, 1, 1, 1, 1, 1, 1