Options
Using model and interactively playing with predictions?
Contributor II
Hello there,
Could someone please explain me how to do the following? I would really appreciate it!
When I make a process, I can load an imported data set from an excel file, then make a model out of it. I can store the model.
Now I want to use the model, and preferably interactively how do I do that?
Do I have to make an excel file with the exact same inputs as my training set (but with the values erased), then convert it, then look at the
model prediction for my label variable? Is there another way? And preferably interactive.
So for example my training set is:
var1 var2 label
1 2 a
2 2 b
Then I make some model out of it by training (preferably some neural network).
When I apply the model I want to for example use it on a set like:
var1 var2
1
and the prediction would be "100% chance for a"
but if I would input:
var1 var2
2
it would predict "50% chance for a"
Of course, there should be multiple variables, and I would like to interactively play with fixing one or the other to a certain bin
(so instead of having to make lots of excel files and convert them, I view all the bins that var1 could be in the training set - in this case 1 or 2, and set
it to one of these values), and see what happens to the predictions.
Thank you in advance!
Could someone please explain me how to do the following? I would really appreciate it!
When I make a process, I can load an imported data set from an excel file, then make a model out of it. I can store the model.
Now I want to use the model, and preferably interactively how do I do that?
Do I have to make an excel file with the exact same inputs as my training set (but with the values erased), then convert it, then look at the
model prediction for my label variable? Is there another way? And preferably interactive.
So for example my training set is:
var1 var2 label
1 2 a
2 2 b
Then I make some model out of it by training (preferably some neural network).
When I apply the model I want to for example use it on a set like:
var1 var2
1
and the prediction would be "100% chance for a"
but if I would input:
var1 var2
2
it would predict "50% chance for a"
Of course, there should be multiple variables, and I would like to interactively play with fixing one or the other to a certain bin
(so instead of having to make lots of excel files and convert them, I view all the bins that var1 could be in the training set - in this case 1 or 2, and set
it to one of these values), and see what happens to the predictions.
Thank you in advance!
0
Answers
I would like to answer your first question how to deal with traing a model and applying it to some data. I've created two getting started processes in myExperiment to decribe the basic outline to deal with training and applying a model.
Please search for "getting started" in the myExperiment View or see
http://www.myexperiment.org/workflows/1762.html
http://www.myexperiment.org/workflows/1763.html
See http://rapid-i.com/component/option,com_myblog/show,Video-on-RapidMiner-Community-Extension-myExperiment-.html/Itemid,172/lang,en/ for the myExperiment stuff.
Ciao Sebastian
P.s. sorry the late answer
Thank you very much for the answer!
Best regards, Michiel