Hello everyone,
First of all, it's important to say that I've been following this forum for some time now, and it helped me a lot – so thank you!
Now it's finally my turn to ask for help, and I really hope you could help me out
I'm working on a project requiring machine-learning, currently using a SVM model in Weka, while the data preparation is done by code.
Now I am tasked with transferring all the coded data preparations into RM, but I'm having difficulties with it.
I'll try to simplify the problem.
Let's say we are trying to predict which students will be suitable for the high school basketball team, using the age and height as attributes.
Basically I'm creating features for SVM using every combination of the attributes, in this case using two (in reality I'm currently up to four attributes, possibly more to come…)
foreach student : exampleSet // from repository
foreach age : constAgeArray // [8, 9, 10]
foreach height : constHeightArray // [130, 135, 140]
if (student.age < age && student.height > height)
// set feature BASKETBALL_POTENTIAL_{age}_{height} = 1
else
// set feature BASKETBALL_POTENTIAL_{age}_{height} = 0
1. I've tried all of the Loop operators to create nested loops, but the process became extremely cumbersome and eventually did not work.
2. Is it possible to define const-arrays in RM? Now I'm using additional exampleSets as arrays…
3. Should I even use RM for this kind of data preparation? Or the best practice is to do it by other means, and import the result into RM for further use (i.e. classification and regression)
4. I would be really grateful if someone could give a RM example for the above basketball data preparation
Thanks in advance!!
