# Taylor series sum and factorial

48 views (last 30 days)
Jack on 18 Aug 2014
Edited: Image Analyst on 18 Aug 2014
Hi,
What is the problem with this code:
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function [S] = taylorTerm1(x,n)
x=1.1;
n=(1:1:5);
p=0;
for i=1:n
z = (1:2i+1:1);
iTerm = (-1)^(i)*(x)^2i+1/z;
S = p + iTerm;
end
fprintf('The Sum of the Series is %6.4f.\n',S);
I am trying to find the sum of series (-1)^n * (x^2n+1)/(2n+1)! and the factorial is confusing me.
Thanks

Geoff Hayes on 18 Aug 2014
Jack - part of the problem may have to do with the equation. You state that you are trying to _find the sum of series (-1)^n * (x^2n+1)/(2n+1)! _. Do you really mean this to be
(-1)^n * (x^(2*n+1))/(2*n+1)!
where the exponent of x is 2*n+1? The line of code that is written as
(x)^2i+1/z;
means that the exponent of x is 2*i (I think it more likely that the exponent is considered as 0.0000 + 2.0000i - see below for details), and added to that result is 1/z when what you want instead is x to the power of 2*i+1 and that result divided by z.
The other problem, like you suggest, has to do with the factorial. Your code for this term is z and it is initialized as
z = (1:2i+1:1);
which is a vector of integers from 1 to 1 with a step-size of 2*i+1. So it will always be just one. Use the factorial function instead which will sum the product of the integers from 1,2,...,2*k+1.
This changes your code to
function [S] = taylorTerm1(x,n)
S = 0;
for k=1:n
kTerm = (-1)^(k)*(x^(2*k*+1))/factorial(2*k+1);
S = S + kTerm;
end
Note the differences. I've removed the initialization of
n=(1:1:5);
since this overwrites the input parameter with a vector. This will cause confusion for the line
for i=1:n
and will probably cause the iterations to terminate early.
I've replaced the i with k since MATLAB uses i and j to represent imaginary numbers. In your code, where there was 2i, this was probably being considered as the imaginary number 0.0000 + 2.0000i and not the desired 2*i.
The kterm has been reworked to something more like what you want, and the summation term was changed so that we were always summing S with the values from the previous iterations and the current iteration.
Try the above and see what happens!
##### 2 CommentsShowHide 1 older comment
Geoff Hayes on 18 Aug 2014
Jack - factorial is a standard function within MATLAB so you should have this functionality. What is the error message if you try
factorial(4)
Also, in the Command Window, type
ver
which factorial -all
What are the results? (Please paste them in a comment.)
----------------
As for an alternative, yes, you could replace the factorial(2*k+1) with
prod(1:1:(2*k+1))
Here, prod returns the product of the elements in the array. For example, if
1:1:(2*k+1) = 1 2 3 4 5 ... 2*k 2*k+1
then
prod(1:1:(2*k+1)) = 1*2*3*4*5 ... *(2*k)*(2*k+1) = factorial(2*k+1)
Try the above and see what happens!

Roger Stafford on 18 Aug 2014
Here is a method that doesn't require the 'factorial' function. See if you can figure out why this is equivalent to the formula you are using.
x = 1.1;
n = 10; % <-- Choose the number of terms to add
s = 1;
for k = 2*n:-2:2
s = 1 - x^2/k/(k+1)*s;
end
s = s*x;
You can compare your results to sin(x). Setting n = 10 is adequate for x values this size, but larger values of x would require larger values of n to compare successfully with sin(x).