Adam optimization algorithm in Alteryx Machine Learning

[hdesai7]

I am using Alteryx Analytics Cloud - Machine Learning for a fraud detection dataset. I am trying to use the Adam Optimization Algorithm in the machine learning model. I am not able to find a setting in Alteryx that let’s me set the optimizer.

I also exported the python code. But, I can’t find the part of the code that determines the optimizer.  If I had to type out the code in Python, it would look something like this:

model.compile(optimizer='adam',

Am I missing something? Is the Adam optimization algorithm a part of one of the models already? I couldn’t find anything in the documentation. If they are a part of the model, which model should I select?

[fesrimo]

Hey I thinks this article could be interesting for your purpose

https://knowledge.alteryx.com/index/s/article/Deploying-a-Keras-TensorFlow-Model-to-Promote-15834614...

[Deano478]

Hey @hdesai7 honestly while Alteryx is a great software it doesn't excel when it comes to ML. For Advanced ML and ML use cases in general I would recommend you explore the Knime Analytics Platform as it offers a wide range of nodes for different ML Models.

[c-lopez]

Hi @hdesai7,

In order to customize a specific portion of the pipeline that AML generates you will have more luck with the more advanced python export option that is presented when you finish building a model on AML.

 

This will give you a Python pipeline for the model you've built on AACP and this is where you can make specific changes, here is an excellent blog post depicting what I am referring to.

Under the hood AML uses Eval ML so documentation to accomplish what you need may be here. 

I hope this helps!