Using DATAROBOT services

Ok so Let's try the DATAROBOT service with a simple machine learning exercise, this one just using a dummy data set of 200 rows with 4 columns (radio, tv, ad and revenue)

After which we feed it to data robot to train it with it's Auto ML service, it will do it's cross validation and will also select the best model to use for this data set. all you have to do is set the target field, in this case it's revenue. In step 5, I took a sample data set from it and created a test file on which we can evaluate the model, in this case using RMSE.

Overall, the downside is the speed of the service, considering I only have 200 rows of data and around 4 columns, I had to click on the Refresh button to see if the service is still running and had to convince my colleagues to do some test runs as well. I wish they would put a cluster machine option to speed up or be able to use at least more resources. The same strategy that Databricks uses when you need to spin up a cluster. Overall, it's okay for Beginner users who don't want to tinker with the code and just plain up use the model. (it had 1 layer Neural Network on Keras - probably built on Python code)

What's good about it is you get a Summary Report that displays the features and the model parameters used - how to explain that to your C level executive is another matter (Explainable AI vs. Off the Box AI Solutions are always ground for debate)

Overall I still prefer customizable AI solutions on Python notebooks, here is the model output

If you have used Datarobot for your ML Deployment solutions, give in a comment, always out looking to improve solutions or whatever is available on the market.