Doubt in definition of Machine learning

machine_learning

#1

In the glossary, ML is defined as
“In these techniques, we expect the algorithms to learn by itself without being explicitly programmed.”

When I have an logistic regression algorithm, do I need to always recalculate the parameters with new data to call it a machine learning model? If yes, how often?

How does it “learn by itself”?

Best regards!


#2

@jpzaidan,

When you say “learn by itself” you mean that the algorithm decides what should be the “parameters” for logistic regression for a data set instead of you “manually” solving some equations and going through data.

Every dataset has different characteristics so parameters for predicting would be different with each new data so you will have to run logistic regression on the new data every time.

Hope this helps :slight_smile:


#3

Thank you @mohdsanadzakirizvi

So, if I run a logistic regression in a historic dataset to predict something, and use the same parameters per 6 months, for example. Can i say this is machine learning?


#4

Yes, you can. But the results may or may not be suitable