# Taylor approximation

A Taylor approximation of a function expresses it in terms of its value at some reference point, and how it changes around that point. For a function of several variables, the rates of change are given by the partial derivatives of the function, and in principle all partial derivatives are needed to perfectly represent the function. This could result in an infinite sum, referred to as a Taylor series, and a finite truncation is then a Taylor approximation.

## Intro

The most fundamental theory of the universe known to man start with a box and a string. Yeah yeah i know it sounds pretty crazy but let me explain.

Imagine that you have a box on a string and that you pull on the box and then release it. Then you will find that the box start to oscillate back and forth.

The mathematics of this phenomena is very easy to understand. Using Taylor series we can find that oscillation forces will behave like this if the region is small enough no matter what the real force is.

This phenomenon is used to explain parts of particle physics and quantum mechanics. These theories are used to explain anything from how light can travel in space, to why atoms don't break. So if you wanna understand why the universe exist you can bet yourself that you wanna understand Taylor series.

Taylor series have many more applications in areas such as finance, biology, medicine and chemistry.

## Concept

Back in single variable calc, we used Taylor's theorem to approximate a function around a given point, using a polynomial. Well, surprise surprise, Taylor's theorem for multiple variables does just the same thing.

Taylor expansions approximate functions as polynomials, in one or many dimensions alike

Pick a point. Compute a bunch of derivatives. Plug it into the Taylor formula. And voilà, that's your polynomial.

The idea here is that the approximation and the original function have the same partial derivatives.

## Math

To see an example, let's approximate the function

using a Taylor expansion of order two around the point .

The formula for around a point is

We proceed by calculating all the required derivatives

Finally, our Taylor approximation is

Setting and , we can take a look on how the approximation looks like below

## Taylor approximation

Sometimes you run into wacko functions, like . I don't know about you, but I don't have an intuitive feel for how this thing behaves. I mean, it contains two separate terms. It's the kind of function which might give you nightmares.

But here's the key: we can easily compute the partial derivatives, right? The partial derivatives help us wrap our heads around , at least if we zoom in on a small area around a given point. Actually, the tangent plane is a way of approximating a function. Think about the differentials! As you move further from the point, the change increases. The tangent plane isn't an accurate approximation though, since it only consists of first-order terms.

But we could also draw on information about the second order partial derivatives, improving our approximation. Then you'd get something like

If we're close to the point, where , the higher order terms will be smaller. They won't matter as much.

We can, as a matter of fact, apply the same reasoning to a function of more than variables, getting

### Example

Find the second order Taylor approximation of about the point

the function has the following partial derivatives

note that all the terms with a will evaluate to at the point , plugging this into the formula yields

which is the function we can see in the image above.