so that they have a standard deviation of 1.0 on the training set. In practice, this means that each feature will have 99.7% of the samples’ values (3 standard deviation) ranging from -3 to 3
This might be very simple, but why does the standard deviation of 1, scales the data from -3 to 3? How could we calculate that beforehand?
Thanks
If they have a standard deviation of 1 (and the data distribution is approx gaussian), then around 68.3% of the values would lie in the interval ranging from -1 to 1. Similarly, 95.4% of the values would be in the interval -2 to 2, and finally, as stated in your quot, 99.7% of the samples’ values would lie in the range from -3 to 3.
This is know as the 68–95–99.7 rule.
What a great explanation! Thanks Arturo
