GMM GMR implementation efficient
Today I used the C++ implementation of GMR for the first time. I was suprised by the amount of steps one should take to perform GMR.
Steps involved:
- Create datapoint object
- Add variable names to the object
- Add Data to datapoint object
- Perform the GMR operation.
I assume we have to use a different datapoint object because the use of 'variable names'. But why do we need these variable names. Why don't we use the input and output indices similar to our matlab GMR function? It will save the hassle of using variable names which, I believe, is prawn to errors/frustration in programming.
When no objections, I can make an alternative simpler implementation.