You want to find the Taylor series of
g(x)=xtan−1x.
First question comes to mind, can we reduce this problem to a smaller one? Yes. We want to find the Taylor series of
f(x)=tan−1x, then when we are done we will divide the series by
x.
If you want to find the Taylor series of a function, you have to tell about what point this series will be. Otherwise, these mathematicians will assume a point that you don't like. And because I am not a mathematician, I will assume a peaceful point,
x=0, which happens to be that we are finding the Maclaurin series of
f(x).
Don't worry about the name. The Maclaurin series is just a Taylor series about
x=0.
Now comes the difficult (fun) part. We will introduce the machine that produces the Maclaurin series:
n=0∑∞n!f(n)(0)xn
This notation is a little difficult to understand, so I will simplify it.
n=0∑∞n!f(n)(0)xn=f(0)+1!f′(0)x+2!f′′(0)x2+3!f′′′(0)x3+⋅⋅⋅ ⋅
Basically, you just need to take the derivatives of the function
f(x)=tan−1x and plug them above after substituting
x=0. But this machine produce infinite derivatives! Now is the important question, how many of them do you need? The answer is, until you recognize a pattern. Don't panic, usually
2 to
3 derivatives will be enough to see the pattern. Try it!
After you recognize the pattern, you will create an infinite series.
tan−1x=n=0∑∞anxk(n)
The final step is to divide by
x:
xtan−1x=n=0∑∞xanxk(n)=n=0∑∞anxk(n)−1
At this point, you have successfully found the Taylor series of
g(x) about
x=0.
Bonus question. Does this series converge for all values of
x?
