Re: Digest for mlcomp@googlegroups.com - 2 Messages in 1 Topic

[Joshua Ainsley]

Simple enough. Thanks for the suggestion!

However, I still get an error:

=== START program2: ./run inspect ../dataset1/train
=== END program2: ./run inspect ../dataset1/train --- OK [0s]
=== START program2: ./run inspect ../dataset1/test
Traceback (most recent call last):
  File "./run", line 89, in <module>
    inspect(sys.argv[2])
  File "./run", line 31, in inspect
    y = int(y)
ValueError: invalid literal for int() with base 10: '1:0.0920242'
=== END program2: ./run inspect ../dataset1/test --- FAILED [1s]
Failed: Inspecting test datashard failed.

Does the test data set require a column for the classifier?

On Sat, Dec 29, 2012 at 5:26 AM, <mlc...@googlegroups.com> wrote:

  Today's Topic Summary

Group: http://groups.google.com/group/mlcomp/topics

• Failed data processing [2 Updates]

 Failed data processing

Josh <jain...@gmail.com> Dec 28 11:45AM -0800  

Hello,
I have just started with mlcomp.org, and I'm running into an issue with the
first processing step for the data (Inspect Split Strip). I have some
continuous data sets. There are 57 lines in the training set and 5186 in
the test set. I've used an SVM for binary classification using R, but I'd
like to see what some other algorithms can achieve with mlcomp. Does anyone
have some advice on what I can fix?
Thanks!
 
=== START program2: ./run inspect ../dataset1/train
Traceback (most recent call last):
File "./run", line 89, in <module>
inspect(sys.argv[2])
File "./run", line 31, in inspect
y = int(y)
ValueError: invalid literal for int() with base 10:
'-1,1:0.0677988,2:9.10265,3:0.015101,4:4.00298'
=== END program2: ./run inspect ../dataset1/train --- FAILED [0s]
Failed: Inspecting train datashard failed.
 
head train:
-1,1:0.0677988,2:9.10265,3:0.015101,4:4.00298
-1,1:39.3518,2:55.8266,3:54.1971,4:306.397
1,1:3.00313,2:0.311742,3:29.0759,4:5.59234
-1,1:25.8011,2:6.47332,3:0,4:0.0975788
-1,1:7.76623,2:62.5829,3:0,4:0
-1,1:3.78907,2:2.14078,3:1.08095,4:12.655
1,1:46.7778,2:0.577116,3:0.122145,4:1.02428
-1,1:1.14223,2:5.76031,3:2.64855,4:23.7908
1,1:713.41,2:179.671,3:543.475,4:271.076
1,1:1.11077,2:0.137706,3:12.7526,4:0.146553
 
head test:
1:0.0920242,2:0.570197,3:1.85306,4:0.593986
1:0,2:0,3:8.21495,4:19.0214
1:1.22572,2:0,3:0.977919,4:0.148526
1:0,2:0.0701989,3:0,4:4.59519
1:0.323363,2:0.805351,3:0,4:1.19509
1:0,2:7.28946,3:0,4:0.843023
1:0,2:49.8327,3:0,4:0
1:0,2:0.12655,3:0,4:8.82058
1:2.51452,2:0.4295,3:2.77731,4:2.86104
1:0.0107031,2:2.61035,3:0.00843482,4:4.13331


 

Percy Liang <pli...@cs.stanford.edu> Dec 28 12:52PM -0800  

Your file format is incorrect. Use spaces rather than commas.
 
-Percy
 


 

You received this message because you are subscribed to the Google Group mlcomp.
You can post via email.
To unsubscribe from this group, send an empty message.
For more options, visit this group.

[Percy Liang]

Yes, your test data should be in the exact same format as the training
data, i.e., have the label as the first column.

-Percy