Options
merging of two ARFF file for attribute selecting
Contributor I
Hi,
I am new member of Rapidminer. I want to use Rapidminer for attribute selection.
I have 3 file training.arff, test.arff and independent.arff, so now I want to perform attribute selection. I want to use same process for all .arff file and select same attribute.
One way is to merge all .arff and apply feature selection.
otherway to apply feature selection on train.arff and select these attributes in test.arff and blind.arff.
As I new person and I have no idea about these think. So how can solve our problems?
Please reply me.
Thanking you.
I am new member of Rapidminer. I want to use Rapidminer for attribute selection.
I have 3 file training.arff, test.arff and independent.arff, so now I want to perform attribute selection. I want to use same process for all .arff file and select same attribute.
One way is to merge all .arff and apply feature selection.
otherway to apply feature selection on train.arff and select these attributes in test.arff and blind.arff.
As I new person and I have no idea about these think. So how can solve our problems?
Please reply me.
Thanking you.
0
Answers
as a start, have a look at the following operators:
- ExampleSetMerge to merge your three example sets
- AttributeFilter to select attributes manually
- FeatureSelection for forward/backward selection
After loading the example sets you can, if you want that, use an AttributeConstruction to add a column specifying which file the data came from. You can automate this using a FileIterator.
Also, have a look at the samples that come with RapidMiner.
Cheers,
Simon
Thanks for your reply.
According to your suggestion I have used ExampleSetMerge this merge the all thre .arff file.
but when I want to save this file it did,t give option for .arff file. Because now I want to save merged file as ARFF file, its give .dat/.aml file.
Is it possible that I can save merged file as ARFF file.
Thanking you.
quote author=Simon Fischer link=topic=1129.msg4301#msg4301 date=1250067392]
Hi Nitish,
as a start, have a look at the following operators:
- ExampleSetMerge to merge your three example sets
- AttributeFilter to select attributes manually
- FeatureSelection for forward/backward selection
After loading the example sets you can, if you want that, use an AttributeConstruction to add a column specifying which file the data came from. You can automate this using a FileIterator.
Also, have a look at the samples that come with RapidMiner.
Cheers,
Simon
[/quote]
you are close :-) Look into the operator tree again. There also is an ArffExampleSetWriter.
Apart from that: Why would you want to do that? You can as well keep on working on this example set inside RapidMiner. You don't have to save all intermediate results.
Best,
Simon