In the second week of my supervised Machine Learning course, I learned about the concept, but I still don’t understand how to apply it to real-world data. For instance, I’m struggling with converting high values into lower ones and comprehending the significance of ranges such as 2 < x < 12.
Normalizing data is importante in order to make the algorithm work fast and give right result , so if have a column of feature that have data range for exemple between 1000 and 14534 (just random numbers) and you have another column of feature that have a data range between 1 and 5, it will look non right to treat the two columns the same way with same formulas , because the won’t give significant result, so we should aim for transforming the first columns data into a data that looks similar to the second one (similar by numbers only) so that we get a range looks closer to the second data column (not necessary the same range) , and the data in both columns will play arround the same numers , and that result a good utlization of the algorithm and give more segnificant result. i think that what’s we mean by feature scaling , and i hope you understant.