problem of imbalnce dataset

hanaabdalrahman

hello i am new in data mining and rapidminer, I have problem of imbalance data set, I wok with decision tree and naïve bayes and random foreset the accuracy of DT, NB is very good but it is not real my question is what is best operators that work with three techniques , my data set contain 1031 sample

 

hana mohamed

student

MartinLiebig

Hi @hanaabdalrahman,

as mentioned earlier: This operator is part of an extension and not of RM core. The extension can be found in our marketplace:

https://marketplace.rapidminer.com/UpdateServer/faces/product_details.xhtml?productId=rmx_operator_toolbox

 

Best,

Martin

Thomas_Ott

@hanaabdalrahman you will need to use the Sample operator and toggle on the 'balance data' option. Then enter the classes and # of samples for each class.

MartinLiebig

Hi Hana,

i recommend to use the SMOTE operator which is part of the operator toolbox extension.

 

Best,

Martin

hanaabdalrahman

thanks for replay...

but how i use it the class false (44) only and class true about (986)

 

 

hanaabdalrahman

thanks...

i work on version 8.0.001 these operator not found in it.. what is best one instead of it and how work?

 

hanaabdalrahman

thanks very match..  upsampling operator solve the problem