I would like to use kennard stone algorithm for splitting a dataset into training and test set. Does Weka /Rapidminer/any free software with GUI has a feature to do this? If so, would like to know the steps to perform this function.
.
Rapid Miner has it :
http://docs.rapidminer.com/studio/operators/data_transformation/filtering/sampling/sample_kennard_stone.html
R has it :
http://soilspec.r-forge.r-project.org/ken.sto.html
Weka i dunno ...
.
Just a word of caution : with a simple KS procedure, your training and test sets will not be independent (which may be of very little significance in practice but will ruin many theoretical approaches) ; the KS procedure can be modified so as to build independent samples while keeping the uniformity property ; see DUPLEX
http://www.mff.cuni.cz/veda/konference/wds/proc/pdf10/WDS10_105_i1_Reitermanova.pdf
(the above link is a short introduction to data-splitting)
.R.D. Snee, Validation of regression models: methods and examples, Technometrics 19 (1977) 415-428
(the original paper)
R has it :
http://finzi.psych.upenn.edu/library/prospectr/html/duplex.html
.
Weka has the requested feature.
Please see http://weka.sourceforge.net/doc.dev/weka/classifiers/Evaluation.html
- Badges
- Science topic