I want to use libsvm file in weka, but I am getting error with .libsvm format.
You will have to read that using the read libsvmread and then save it into the format you need. Matlab would allow you to do this with a script of 3 or 4 lines. Do you use any scripting/programming language?
I have used libsvmread in matlab, but i am not getting idea to convert this data in .csv format.
read the data into a matrix and use dlmwrite to write the matrix to a file. CSV format is simply comma-separated values.
I have read libSVM file using:
[lebel instance]=libsvmread(file)
But here my data is saved in two matrix:lebel, instance.
So how can i use it for dimwrite
because dimwrite syntax is: dlmwrite(filename, M)
when i run:-
dimwrite('filename.csv', instance)
following error occur:
Undefined function 'dimwrite' for input arguments of type 'char'.
In MATLAB, use
data = [ double(label) double(instance) ]
to concatenate two matrices (of equal number of rows ; ) into one.
Dear Matthias
After doing this following error is coming:
>> dlmwrite('FILENAME.csv',data)
error: Function is not defined for sparse inputs.
RTFM: http://www.mathworks.ch/ch/help/matlab/ref/full.html
Full(data) makes a sparse matrix full and sparce(data) does the opposite.
it's done. But when i use this .csv file with weka, then it display error msg. Why?
weka uses ARFF as because it contain defined attributes,.csv does not has this feature,weka needs labels for every nominal attributes,my suggestion u format ur data as in ARFF.
weka uses ARFF as because it contain defined attributes,.csv does not has this feature,weka needs labels for every nominal attributes,my suggestion u format ur data as in ARFF
Try using libsvm directly, if you get the same error let us know what the error is..
Here snap of weka is attached with my problem, In this picture start button is not highlighted with .libsvm file created by me. But when i change attribute evaluator field then start button works.
0.0 2:6.0 3:4.0 4:3.0 5:1169.0 6:4.0 7:4.0 8:4.0 9:2.0 11:4.0 13:67.0 14:2.0 15:1.0 16:2.0 17:2.0 18:1.0 19:1.0
1.0 1:1.0 2:48.0 3:2.0 4:3.0 5:5951.0 7:2.0 8:2.0 9:1.0 11:2.0 13:22.0 14:2.0 15:1.0 16:1.0 17:2.0 18:1.0
0.0 1:3.0 2:12.0 3:4.0 4:6.0 5:2096.0 7:3.0 8:2.0 9:2.0 11:3.0 13:49.0 14:2.0 15:1.0 16:1.0 17:1.0 18:2.0
0.0 2:42.0 3:2.0 4:2.0 5:7882.0 7:3.0 8:2.0 9:2.0 10:2.0 11:4.0 12:1.0 13:45.0 14:2.0 15:2.0 16:1.0 17:2.0 18:2.0
1.0 2:24.0 3:3.0 5:4870.0 7:2.0 8:3.0 9:2.0 11:4.0 12:3.0 13:53.0 14:2.0 15:2.0 16:2.0 17:2.0 18:2.0
0.0 1:3.0 2:36.0 3:2.0 4:6.0 5:9055.0 6:4.0 7:2.0 8:2.0 9:2.0 11:4.0 12:3.0 13:35.0 14:2.0 15:2.0 16:1.0 17:1.0 18:2.0 19:1.0
I believe the labels(first number in each line) have to be integers.
This will not give you a runtime error but do you scale your data?
I converted labels into integer, but it's not working. Please tell me all constrain for .csv file.
How you have convert your data into arff ?? (my data is in .mat file)
you should convert the class to nominal using the NumericToNominal filter.
- Badges
- Science topic
- Similar topics
- Data Mining
- Classification