Original subtitles

Instructor: We will start with one of the

simplest transformations.

Often we are not interested

in an absolute value, but a relative value.

That's usually the case when working with stock prices.

If you open Google and type Apple stock price,

what you'll get is Apple stock price,

but with red or green numbers,

the relative change in Apple stock price.

This is an example of pre-processing that is so common

we don't consider it as such.

Relative metrics are especially useful

when we have a time series data, like stock prices,

Forex exchange rates, and so on.

Still in the world of finance,

we can further transform these relative changes

into logarithms.

Many statistical and mathematical methods

take advantage of logarithms

as they facilitate faster computation.

In machine learning,

log transformations are not as common

but can increase the speed of learning.

Okay, great.

So this is one type of pre-processing we wanted to give

as an example.

Let's continue to our next lesson,

where we will deal with pre-processing transformations

that are more typical for the world of machine learning.

Stick around, and thanks for watching.