Hello,

Machine learning is a subset of AI. AI systems are system that emulate human tasks. Machine learning can make AI better: the AI system learns and its performance improves becoming better than if we explicitly programmed the system....

That said, some believe that ML is glorified statistical analysis since linear regression, logistic regression, clustering, etc. (except for neural nets and deep learning) exist and existed already in statistics...For example, linear regression is certainly well established in statistics but it is also often listed as a ML algorithm....

what is the correct view on this?

thank you

 

Maybe you could explain why it matters.

 

Maybe you could explain why it matters.

Fair

Well, it is more of a conceptual curiosity and about making sure I understand what ML is all about and how it is different from statistical analysis.

ML relies on lots of data. Doing ML with little data does not make sense. But traditional statistics can also work with tons of data. Statistical algorithms can get optimized. In ML, we are also and essentially optimizing an algorithm: humans choose the ML algorithm better suited for the problem and the training data "finds" the best weights/parameters for that algorithm via internal error and performance metrics...

I guess the optimization process in statistical analysis is somewhat different from how it is performed in ML...

 

ML is all about predicting outcomes by analyzing accumulated data.
Data analysis has a much broader perspective, and is less oriented toward prediction.