Options
help about rapidminer!
coscacho17
Member Posts: 1 
Contributor I
Hi people. Me and a partner are begginers in this area, and we´ve serious problems with this tool. We are from Argentina, we don´t speek english veri well, so try to be patiente please!.The problem is the following:
We are in a research project on datamining in which we use this tool (rapidminer). We must focus on the classification process within the area of datamining, what we need to do is to take a sample of the database, train this sample data and then predict the results of another database with same attributes. The steps we take to do this are:
root
arffexamplesource -> here we include the sample to train the model and indicate the attribute that we predict
Learner -> NaiveBayes or some tree
Model Writer -> here is the path where you save the model
After that, we began the steps to predict the other database:
root
arffexamplesource -> here is the path to the database to predict
ModelLoader -> here is the path of the model saved in the previous steps
ModelAplier
I wonder if these steps are correct, because the results acquired and expected are not equal.
any help is going to help us to keep on working
THANKS!!!
We are in a research project on datamining in which we use this tool (rapidminer). We must focus on the classification process within the area of datamining, what we need to do is to take a sample of the database, train this sample data and then predict the results of another database with same attributes. The steps we take to do this are:
root
arffexamplesource -> here we include the sample to train the model and indicate the attribute that we predict
Learner -> NaiveBayes or some tree
Model Writer -> here is the path where you save the model
After that, we began the steps to predict the other database:
root
arffexamplesource -> here is the path to the database to predict
ModelLoader -> here is the path of the model saved in the previous steps
ModelAplier
I wonder if these steps are correct, because the results acquired and expected are not equal.
any help is going to help us to keep on working
THANKS!!!
0
Answers
your general setup seems to be correct. If anything is wrong, it would be helpful to see, what is wrong in your opinion. If you just think your model does not predict well, you can start by performing a cross validation on the original data to estimate how the performance will be like.
Best,
Simon