What is the purpose of the Operator "Normalize"?

Ellie98

Hello,

my teacher explained today that we should only use the Operator "Normalize Z-Score" if we only have numeric attributes and the "Normalize MinMax" if we have mixed attributes (numeric, nominal and binominal). I don't really understood when to use those different Operators. He mentioned to use it, if the Min or Max value difference is too high, but what exactly is the purpose of the Normalize Operator? When do we actually need it? What exactly is it doing?

Thanks in advance!

Caperez

Hi @Ellie98,

As a short answer, the target of the normalizaiton is to  change the values of the dataset to a common scale conservating the relative differences between the examples.

This can have a drastic importance in some models.

please read this two interesting explanations. 

Understand Data Normalization in Machine Learning | by Zixuan Zhang | Towards Data Science
Why Data Normalization is necessary for Machine Learning models | by Urvashi Jaitley | Medium

regards