s-news
[Top] [All Lists]

Intrepretation of warning/error from stepwise

To: snews <s-news@wubios.wustl.edu>
Subject: Intrepretation of warning/error from stepwise
From: Kim.Elmore@noaa.gov
Date: Sun, 22 Mar 2009 21:38:58 -0500
I have a data set with 841 rows and 13 columns and a response variable
as a vector of 841 elements. on which I have run stepwise. If I ask for
method = "exhaustive" I get an error message that I don't understand: 

> stepwise(training.data, training.response, method = "exhaustive")
Warning messages:
  length of data is longer than product of dimensions, data truncated
in: array(data, dim[2:1])
Problem in stepwise(training.data, training.response, method =
"exhaustive"): The Fortran code has returned as error
      -- check the consistency of the arguments supplied 
Use traceback() to see the call stack

If I run any other method, I still get the following warning:

Warning messages:
  length of data is longer than product of dimensions, data truncated
in: array(data, dim[2:1])

According to the documentation, stepwise expects x to be either a matrix
or dataframe of explanatory variables, and y to be a vector of
responses. In my case, "x" is a 841 x 13 matrix, and "y" is a vector of
length 841. All variables are continuous.

What might be wrong with the data I'm passing to stepwise?

Kim Elmore

<Prev in Thread] Current Thread [Next in Thread>
  • Intrepretation of warning/error from stepwise, Kim . Elmore <=