RAPIDMINER 9.7 BETA ANNOUNCEMENT
The beta program for the RapidMiner 9.7 release is now available. Lots of amazing new improvements including true version control!
Iterating, Macros, Automation
I've attached a process and excel data file. The process looks like this:
Inside those green boxes are two dates and assume the nonexistent Store procedure has a file name in it.
I need to accomplish a few things:
- I need to find a way to iterate over dates in the green boxes. (E.G. say I had 4 years of timeseries data... How could I automate the process of training the model on 24 months, applying the resulting model to one week for testing, advance one week, train on the next 24 months and then test on the next week, rinse and repeat?)
- What is the best way to store Apply Model output (i.e. predictions with confidences, etc.) and performance metrics (confusion matrix-type stuff) for each iteration such that one can go back and look at them all in aggregate. (E.G. How could I *cumulatively* compare the performance of 24mo training periods/test on one week for a year vs 36 mos or 18mo training periods?)
- I suspect there is a better way to cobble together a year's worth of weekly output than spewing out 52 excel sheets and stitching them together by hand... any suggestions?