SAS Tutorial | Machine Learning Fundamentals

[@farahildasaputri]

Wow! Ari is so good at explaining! Thank you. I have a clearer understanding now ❤

[@SASUsers]

He does an amazing job, we couldn't agree more. Thanks for tuning in!

[@danicuenda8324]

That was so good, clear and easy to understand. Even if it's SAS related, this is a great tutorial of Machine Learning in general.

[@SASUsers]

Thanks for tuning in, and we're glad you enjoyed the content!

[@dalton_nbg]

Great overview Ari - very easy to follow. Thanks. To confirm, the neural network results will not enable one to get a good sense of which input variables are most strongly related to the outcome variable, correct? However, if you are just interested in accurate prediction, it doesn't matter, right?

[@SASUsers]

Thanks for your inquiry, Nick! We're checking on this for you!

[@SASUsers]

Thanks for your patience, Nick! We reviewed with Ari and he provided a reply back! See below: The neural network results will not tell you which input variables are most strongly related to the target variable. One drawback of neural networks is they have no built-in variable selection, so the neural network doesn’t know which variables are important and which are unimportant. It takes all the variables you give it and makes a model using those variables. In the video, I used a Variable Selection node above the Neural Network node so that we only allow the neural network model to use a variable that is deemed important for predicting the target based on the Variable Selection node. Neural networks are best when you need accurate predictions more than explanations of what variables are important. Nevertheless, it is a good idea to do variable selection before building neural network models, because otherwise, you will waste computational time generating weights for inputs that don’t help in predicting the target. There is additional information on this in our course: Machine Learning Using SAS Viya, which is available at this link: 2.sas.com/605116jbZ.

[@dalton_nbg]

@@SASUsers Great; that is exactly what I needed to know. Please pass along my thanks to Ari.

[@SASUsers]

We absolutely will share with, Ari! Have a great weekend!

[@joleenksa388]

What is DEROG and CLNO variable meaning ?

[@SASUsers]

Thanks for your question! The DEROG variable is the number of derogatory credit reports on the customer’s history and the CLNO variable is the number of lines of credit associated with the customer. So in both cases the variables contain information about the customer’s past use of credit.

[@istuart0]

I find it? very annoying? when you put? inflections? needlessly? several times? in the middle? of sentences?. And again? when you add an inflection? at the end end? of a sentence? that isn't? a question?