Approximating sum of series
have done binomial expansion before?
taylor?
taylor series yes
how isit related?
calculate the taylor series ... integrate term by term of the taylor series.
hm.. which formula is that?
that is taylor series of the sqrt(1+x^4) expanded at zero.
I am guessing that this integral is elliptik
The question is, can you find how many terms are necessary for a given number of decimal places? How can you predict how fast it will converge.
the upper limit is 1, the lower limit is 0 ... so just use first few terms. looks like this is alternating. there is one way of approximating for alternating series but forgot it. most likely it's something like squeezing.
so means i have to integrate it then?
You are integrating an approximation of the Taylor series expansion for the function. So you are integrating a few terms of the series.
\[\sum_{n=0}^{\infty}\left(\begin{matrix}k \\ n\end{matrix}\right)x^n=1+kx+\frac{ k(k-1)x^2 }{ 2! }+\frac{ k(k-1)(k-2)x^3 }{ 3! }+...\] can i use this formula?
i dont really get it,
i actually got the series 1 + x^4/2 - x^8/8 +... but i dont know to proceed after that
integrate term by term
but up to how much term do you need to evaluate? http://tutorial.math.lamar.edu/Classes/CalcII/EstimatingSeries.aspx
okay, so it will be x + x^5/10 - x^9/72+x^13/208+...
it should be correct to 2 decimal places..
hmm ... you get alternating series.
yeah i think i shd make it into a geometric series form?
do you know how to estimate the value of alternating series?
no ... this is not geometric. if you want general term ... you should use generalized binomial theorem ... but last time you rejected the idea when I suggested.
well, i have no idea how to do it with an integral haha can u teach me? :3
Damn stupid wolf ... do you have Mathematica?
no, what is that? haha, anw if u dont want to we could just proceed to where i am now
http://www.wolframalpha.com/input/?i=Series%5BSqrt%5B1%2Bx%5E4%5D%2C+%7Bx%2C+0%2C+20%7D%5D all right ... first find Taylor expansion of this.
then integrate it term by term.
then find the term \( t_n \) such than \( t_n < 0. 009\)
so means my integration is wrong just now?
where is your integration?
its x+ x^5/10 - x^9/72 + x^13/208+..
yes integrate it ... term by term.
huh, i already did, thats the answer, isnt it that we integrate 1+x^4/2-x^8/8+...?
you mean integrate again? why so?
of course ... your integral is definite integral, you need to put the values of 'x from 0 to 1
http://www.wolframalpha.com/input/?i=Integrate%5Bx%2B+x%5E5%2F10+-+x%5E9%2F72+%2B+x%5E13%2F208%2C+%7Bx%2C+0%2C+1%7D%5D I was trying to tell you that you only need to integrate the first two terms.
woops!! did something go wrong?
i have no idea haha, im just puzzled coz we had to integrate it 2 times.. i got the series expansion of (1+x^4)^1/2 then from then i dont know already
you said integrate t so i did, and i got it, but u said u need to integrate it again and i dont know why it is related from having a definite integral from 0 to 1
http://www.wolframalpha.com/input/?i=Integrate%5BSqrt%5B1%2Bx%5E4%5D%2C+%7Bx%2C+0%2C+1%7D%5D
Evaluate up to this much 1+x^4/2-x^8/8+x^12/16-(5 x^16)/128
the problem is how to get how many terms to get an answer to 2 dec places?
yes ... that is the problem ... when you integrate x^n/k from 0 to 1, you get 1/(n*k)
hmm, so means i just have to like try out and estimate and see which will give me <0.009?
i mean for decimals.
no ... find the term, whose integral from 0 to 1 ... is less than 0.0001
or 0.001
just look at the series http://www.wolframalpha.com/input/?i=Series%5BSqrt%5B1%2Bx%5E4%5D%2C+%7Bx%2C+0%2C+20%7D%5D see if this is less than 0.001 or not ... or just you can take less than 0.009
http://www.wolframalpha.com/input/?i=Integrate%5B%285+x%5E16%29%2F128%2C+%7Bx%2C+0%2C+1%7D%5D
this term is not significant ... so just leave it, and all terms after it. only evaluate the terms before it.
read this article if you find tyme ... http://tutorial.math.lamar.edu/Classes/CalcII/EstimatingSeries.aspx
i got 1.4375
add more terms then
you should get 1.09
okay... btw, i've been taught o take the taylo series but not the expansion but because in this case \[(1+k)^k=\sum_{n=0}^{\infty}\left(\begin{matrix}k \\ n\end{matrix}\right)x^n\] i dont know how to interpret the matrix so i dont know how to find the general term , that would be easier though
yeah i did i got 1.09 thanks!(:
that is not matrix ... it's k choose n. although you hardly use it ... instead you use generalized binomial theorem.
hm, how to use it? can u make this an example?
\[ (1+x)^r = 1 + r x + r(r-1)/2! x^2 + r(r-1)(r-2) x^3/3! + ... \]
haha okay xD thats what i did just now, thank you so much for your help! :D
put r=1/2 and x-> x^4 ... and see what you get, you will get your Taylor series.
yup! figured that out just now, but you were saying that there was some other method u were suggesting but i didnt follow
that is for estimating the value of your integral ...
okay! :D thanks anyways!
sorry for going offline suddenly ... please read this article you will find it very helpful http://tutorial.math.lamar.edu/Classes/CalcII/EstimatingSeries.aspx
thank you! :D
Join our real-time social learning platform and learn together with your friends!