Combining two or more algorithms together

Question

Hi

I was wondering if there was a possibility to combine two algorithms together, would we need to use a python script or is an inbuilt way of doing is?

Thanks in advance for your help!

Answer 1

Could you please provide an example of your use case, which algorithms you would want to combine, and what you want to accomplish (e.g. higher accuracy, etc.)?

---

Ted Way [MSFT], Program Manager, Microsoft Azure Machine Learning

Answer 2

I would to achieve higher accuracy, for classification algorithms

Answer 3

here is a good example using Ensemble with Stacking:

https://gallery.cortanaintelligence.com/Experiment/Building-Ensemble-of-Classifiers-using-Stacking-2

Yes, you generally will resort to R/Python to produce the winning prediction based whatever criteria you decide on.

Answer 4

There is not a built-in method to do ensembles in Azure ML. But there is a few things you can do.

As mentioned in another reply; you can do Stacking: Basically you select the "Scored Probabilities" of your base models (the ones you want to combine) and use them as inputs for another Model (called a super-learner) 

https://gallery.cortanaintelligence.com/Experiment/Building-Ensemble-of-Classifiers-using-Stacking-2

For another ensembles methods such as Voting, Bagging, weighted sum, etc. You can do them in a Python/R script.

Answer 5

You can also use Apply Math for voting, Bagging and weighted sum