[R] Penalized L1 regression error in lasso2 package

From: Miguel Ratón Almansa <cibermike_at_telefonica.net>
Date: Wed, 09 Jan 2008 15:11:33 +0100


Hi everbody,

I need to apply lasso regression using the function l1ce from the lasso2 package:

fr <- l1ce( fitness_T ~ ., data = dr, absolute.t=TRUE, standardize=FALSE );

Where fitness_T is my response, "." represents all the terms of my dataframe dr. This function fails when I use a lot of variables (around 200-300 terms) in the dataframe. Is there any constraint about this? This is the error message I get:

Error en l1ce(fitness_T ~ ., data = dr, absolute.t = TRUE, standardize = FALSE) :
  Oops, something went wrong in .C("lasso",..): -1

Sometimes I get the same error when I introduce dataframes like the following (with only 30 terms):

          X1                  X2           X3              X4           
    X5             X6              X7              X8
1 0.7320944 0.4567412 0.44856288 0.05964432 0.05954413 0.6156447 0.7659501 0.573173534
2 0.7773032 0.3449025 0.04422306 0.50439358 0.39123515 0.8047629 0.4401352 0.064500660
3 0.7773032 0.3449025 0.04422306 0.50439358 0.39123515 0.8047629 0.4401352 0.064500660
4 0.5768691 0.4639482 0.25751515 0.33793192 0.01037575 0.1798244 0.5707092 0.735116023
5 0.9802398 0.5033950 0.48795317 0.97431465 0.51465128 0.2914794 0.6774021 0.602885327
6 0.9802398 0.5033950 0.48795317 0.97431465 0.51465128 0.2914794 0.6774021 0.602885327
7 0.9802398 0.5033950 0.48795317 0.97431465 0.51465128 0.2914794 0.6774021 0.602885327
8 0.8252928 0.4653600 0.85582162 0.53349585 0.81383935 0.8111897 0.5070405 0.006989377
9 0.8252928 0.4653600 0.85582162 0.53349585 0.81383935 0.8111897 0.5070405 0.006989377
10 0.4685068 0.6208291 0.20864631 0.81919070 0.88421721 0.4488032 0.3170425 0.268553276

...

          X25              X26             X27             X28          
   X29          X30             fitness_T
1 0.51582545 0.2883907 0.004668415 0.50093324 0.5488026 0.36071348
-30911635603

2 0.50049201 0.5951309 0.388231982 0.87264326 0.7180451 0.09474729
-30592911895

3 0.50049201 0.5951309 0.388231982 0.87264326 0.7180451 0.09474729
-30592911895

4 0.48488880 0.5687469 0.945986723 0.04843440 0.3921170 0.68509765
-23433191157

5 0.74335685 0.6656366 0.295069967 0.65278954 0.8403278 0.01226897
-8358708674

6 0.74335685 0.6656366 0.295069967 0.65278954 0.8403278 0.01226897
-8358708674

7 0.74335685 0.6656366 0.295069967 0.65278954 0.8403278 0.01226897
-8358708674

8 0.55207468 0.5004715 0.396061154 0.78939996 0.9649179 0.10526002 12377155187
9 0.55207468 0.5004715 0.396061154 0.78939996 0.9649179 0.10526002 12377155187
10 0.01548059 0.5422345 0.408542004 0.05060681 0.8768663 0.49197307 115852466198

I get the same error. Is there any constraint about the values that can take the response in the formula of l1ce?

Can somebody help me?

Thank you for your help,
Miguel Ratón Almansa



R-help_at_r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. Received on Wed 09 Jan 2008 - 14:20:42 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Wed 09 Jan 2008 - 15:30:05 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-help. Please read the posting guide before posting to the list.

list of date sections of archive