Automatic Linear Regression

[gshikha16]

Hi,

I have recently started learning about KNIME functionalities and I need help finding a way of doing the automatic data preparation and automatic variable selection in KNIME.
For the same purpose, I have used SPSS before with the Automatic Linear Modeling feature includes (Data preparation feature - on and best model selections method, Automatic variable selection).
How this can be done in KNIME?

Thanks,
Shikha

[Daniel_Weikert]

Hi Shikha,
in KNIME there are several nodes which might help you. E.g. Feature Selection Node, Parameter Optimization loop are the ones I can currently think of. You might have a look at them
br

[mlauber71]

@gshikha16 welcome to the KNIME forum

Concerning automatic model building I would like to point you to two approaches. One I built a wrapper for H2O.ai AutoML module - and one version with the added benefit of using R’s vtreat to prepare the data.

You would have to install R on you machine and make it work with KNIME. Yes not the most comfortable setup, but hey: you get some very powerful tools that could help you with these tasks.

H2O.ai AutoML in KNIME for regression problems

H2O.ai Automl - a powerful auto-machine-learning framework wrapped with KNIME It features various models like Random Forest or XGBoost along with Deep Learning. It has wrappers for R and Python but also could be used from KNIME. The results will be written to a folder and the models will be stored in MOJO format to be used in KNIME (as well as on a Big Data cluster via Sparkling Water). One major parameter to set is the running time the model has to test various models and do some hyperparam…

H2O.ai AutoML (wrapped with R) with vtreat data preparation in KNIME for...

Then KNIME is working to deploy Guided analytics and automated machine learning. I am not sure if there currently is a component for regression tasks. Maybe @paolotamag can say something about that.

https://www.knime.com/integrated-deployment-knime-blog-series

knime/Examples – Automation

[paolotamag]

Hi there,
KNIME Workflows and Verified Components offer automated machine learning only for classification task at the moment. That is:

Guided Automation – simons

as well as the most recent and simplified one:

AutoML Component via Interactive Views – knime

using:

AutoML – knime

The AutoML component is made to work with other components to cover the many use cases:

eXplainable Artificial Intelligence (XAI) - Simple – knime

Model Monitoring with AutoML and MLOps – corey

Regarding AutoML for Regression Models:

We are still working on it. For now we have an example you can look at to take inspiration that we use for education purposes in our KNIME Pros Learnathon. This is the solution of the exercise we give there to solve AutoML for regression. This is not a Verified Components and it is not recommended to work in a real use case: it does not have parameter optimization for example.

KNIME Pros Learnathon - Group 3 : Automation - Solution – knime

Hopefully we will supply a more refined versions this year.

Cheers
Paolo