A dataset can have different attributes. The attributes can have different magnitudes, variances, standard deviations, mean values, etc. For instance, the salary can be in thousands, whereas age is normally a two-digit number. The difference in the scale or magnitude of attributes can actually affect statistical models. Variables with a bigger range dominate those with a smaller range, for linear models. Similarly, the gradient descent algorithm converges faster when variables have similar scales. Feature magnitudes can also affect Euclidean distances.
In this chapter, you will see different feature scaling techniques. ...