s-news
[Top] [All Lists]

Big Data Library

To: s-news@lists.biostat.wustl.edu
Subject: Big Data Library
From: James Leszczenski <jleszcze@andrew.cmu.edu>
Date: Sat, 03 Feb 2007 10:16:43 -0500
User-agent: Thunderbird 1.5.0.9 (Windows/20061207)
Hi-

I'm fairly new to this mailing list, but figured I'd give it a shot since I've had a problem that I haven't been able to figure out despite my best efforts. My question is with regards to the functionality for Big Data, under version 7.0.6 of SPLUS(Windows, Enterprise Edition). I have a dataset with one numeric response(dependent) variable and three factor variables as independent. There are around 360,000 rows. I have no problem importing the data as a bdFrame, and have verified that the frame itself was correctly constructed. However, when I try to perform logistic regression using bdGlm, it still complains about insufficient dynamic memory. The documentation implies that the call uses the same parameters as the regular glm function, which I have successfully used on smaller datasets.

Are there any known issues with using factors for this? I've run the boston_housing example perfectly fine, but this doesn't have the same amount of data, and also only has numeric variables. I'd appreciate any pointers you might be able to provide.

Thanks,
   James

--
James Leszczenski
Researcher, School of Computer Science
Carnegie Mellon University
jleszcze@andrew.cmu.edu


<Prev in Thread] Current Thread [Next in Thread>
  • Big Data Library, James Leszczenski <=