Auto-Weka parameters do not give the same result when entered manually. Why?

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Auto-Weka parameters do not give the same result when entered manually. Why?

erkanozhan
This post was updated on .
Hi,

*Auto-WEKA result:*
best classifier: weka.classifiers.functions.SMO
arguments: [-C, 1.063722598005212, -N, 2, -K,
weka.classifiers.functions.supportVector.Puk -S 6.030922204438996 -O
0.8073168741747109]
attribute search: null
attribute search arguments: []
attribute evaluation: null
attribute evaluation arguments: []
metric: errorRate
estimated errorRate: 0.001567398119122257
training time on evaluation dataset: 3.916 seconds
---------------------------------------------------------------------------
You can use the chosen classifier in your own code as follows: (*What is the
Classify tab equivalent of this code?)*
----------------------------
Classifier classifier =
AbstractClassifier.forName("weka.classifiers.functions.SMO", new
String[]{"-C", "1.063722598005212", "-N", "2", "-K",
"weka.classifiers.functions.supportVector.Puk -S 6.030922204438996 -O
0.8073168741747109"});
classifier.buildClassifier(instances);


Correctly Classified Instances        3173               99.4671 %
Correctly Classified Instances        3173               99.4671 %
Incorrectly Classified Instances        17                0.5329 %
Kappa statistic                          0.9913
Mean absolute error                      0.2234
Root mean squared error                  0.2743
Relative absolute error                 54.4793 %
Root relative squared error             60.5867 %
Total Number of Instances             3190    
---------------------------------------------------------------------

*Manual parameter entry*

I switch to the Classify tab. I right click on the Algorithm Selection line
and select "Enter Configuration". And I enter the following. (By the way,
Auto-Weka reverses ordering for -S and-O):
-------------------------------------
*weka.classifiers.functions.SMO -C 1.063722598005212 -N 2 -K
"weka.classifiers.functions.supportVector.Puk -O 0.8073168741747109 -S
6.030922204438996"
*

Result:
Correctly Classified Instances        3084               96.6771 %
Incorrectly Classified Instances       106                3.3229 %
Kappa statistic                          0.946
Mean absolute error                      0.2304
Root mean squared error                  0.2869
Relative absolute error                 56.1945 %
Root relative squared error             63.3555 %
Total Number of Instances             3190    
------------------------------------------------------------------
It is very confusing to have two different results. I've been searching for
the reason for this for weeks but can't find it I think eventually I'll have
to get to Auto-Weka's coders.


Cheers,
Erkan.

--
Sent from: https://weka.8497.n7.nabble.com/
_______________________________________________
Wekalist mailing list -- wekalist@list.waikato.ac.nz
Send posts to wekalist@list.waikato.ac.nz
To unsubscribe send an email to wekalist-leave@list.waikato.ac.nz
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html