The Altair Community is migrating to a new platform to provide a better experience for you. In preparation for the migration, the Altair Community is on read-only mode from October 28 - November 6, 2024. Technical support via cases will continue to work as is. For any urgent requests from Students/Faculty members, please submit the form linked here
W-M5P application problems
Contributor II
Hallo,
I got some very good results in predicting numerical label ( with numerical + polynominal attributes ) with Weka's M5P operator.
Unfortunately, if i load the model from a file and want to apply it to a new data set ( with exactly the same attributes and type of data, only other values ) it produces no output ( only "?"/ missing values ). What could be wrong?
I found out, that it is possible to apply this operator in one process to the same data it is learned with. Furthermore it is not possible to apply it on an origin data set with maybe one or two attributes more than used in the training process.
Thx a lot for help!
Ozone
I got some very good results in predicting numerical label ( with numerical + polynominal attributes ) with Weka's M5P operator.
Unfortunately, if i load the model from a file and want to apply it to a new data set ( with exactly the same attributes and type of data, only other values ) it produces no output ( only "?"/ missing values ). What could be wrong?
I found out, that it is possible to apply this operator in one process to the same data it is learned with. Furthermore it is not possible to apply it on an origin data set with maybe one or two attributes more than used in the training process.
Thx a lot for help!
Ozone
0
Answers
could you please file a bug report for that? Could you try different WEKA Models as well? Might be there's a general problem, but I haven't the time to address this right now.
Greetings,
Sebastian
I fixed the problem. For some operators it seems to be important that the test example set has the same number of attributes as the training example set.
In my case, I forgot to declare the label attribute (as it was in the training data set) as label. So it was one attribute to much in the test data set.
Nov 18, 2010 2:24:59 PM WARNING: W-M5P: The number of regular attributes of the given example set does not fit the number of attributes of the training example set, training: 47, application: 48
Nov 18, 2010 2:24:59 PM SEVERE: W-M5P: Exception occured while classifying example:48 [class java.lang.ArrayIndexOutOfBoundsException]
Greetz