I dont know how many mathematicians we have out there, but...ill give it a shot anyway this has to do with statistics and higher moments Is there a maximum or minimum value for the coeeficient of kurtosis? which is- E[(x-u)^3]/sigma^3 Is there a maximum or minimum value for the coefficient of skewness? which is- E[(x-u)^4]/sigma^4 ------------------ "There are some frauds so well conducted that it would be stupidity not to be deceived by them." Charles Caleb Colton (1780-1832)
what the hell is a kurtosis? If u ask the question in english... ------------------ SUCK POLICE!!!!!! To point out individuals or teams that have managed to reach the pinnacle of SUCKINESS!!!!! ----- THE WASHINGTON SUCKSKINS!!!!!!! My Cowboys might SUCK, but its nice to know that we can always rely on the SUCKSKINS to be our B**CH!!!!
You should have asked me that about 18 years ago, when I was a hotshot chemistry major. After that long a time (not to mention how many beers and joints ingested on this brain), it all looks like one of those keyboard stick figures I see on the bottom of some emails. ------------------ Behad Sergeant at Arms of the Clutch BBS
Okay, I know what a "coefficient of variance" is. Is that "E" symbol the "sum" symbol, or like the symbol for Elasticity? The "u" or "mu" (I think), is the symbol for "mean" (I think) The "sigma" is the symbol for standard deviation, no? Give us a little more info / background to the problem. ------------------ "I have a DREAM.........his name's Hakeem." DREAMer's Rocket Page
I bet u guys dont know whats: infinity * infinity ? ------------------ SUCK POLICE!!!!!! To point out individuals or teams that have managed to reach the pinnacle of SUCKINESS!!!!! ----- THE WASHINGTON SUCKSKINS!!!!!!! My Cowboys might SUCK, but its nice to know that we can always rely on the SUCKSKINS to be our B**CH!!!!
U is mu, which is the mean and sigma is the standard deviation basically these coeeficients are the 3rd and 4th moments, with the first and second being the familiar mean and standard deviation respectively. E, then, represents expected value. skewness is a measure describing the lack of normality in distribution...in other words, a heavy tail to one side or the other. kurtosis is a measure describing the amount of weight in the tail of a distribtuion. so basically, the question is asking for a mathematical explanation of wether or not a distribution can be skewed indefinitely in one direction or the other,etc,etc. it may have something to do with chebychev's inequality and the related law of large numbers. that may help, but porbably not ------------------ "There are some frauds so well conducted that it would be stupidity not to be deceived by them." Charles Caleb Colton (1780-1832) [This message has been edited by JayZ750 (edited December 13, 2000).]
Nope, didn't help a whole lot. At least, I knew what you were talking about... sorta. Does the distribution involve a "Z-Table", or if you're talking about larger tails, then I think it's the "T-Table" you use? Anyways, I've only taken STAT 1, and I don't ever want to take another STAT class. I sold that damn STAT text book back as fast as I could. In finance we did use the Z-Table quite a bit, but never as indepth as in STAT. What about the "lamba" variable, or "Poisson"? I wonder how come I like Finance, but hate: Economics, Statistics, and Accounting? Weird, huh? ------------------ "I have a DREAM.........his name's Hakeem." DREAMer's Rocket Page
JayZ, Jiggaman doin' stats... hilarious. I keep picturing Jay Z rappin' about standard normal distributions. I'm definitely no statistician or mathematician, but wouldn't the third and fourth moments be infinite when std. dev. is zero? Of course a std. dev. of zero implies the same data points, too, I think. My only question is whether or not the moments would be called "infinite" or "undefined", because a std. dev. of zero implies (x-u) = 0, too, right? Leaving you with 0/0. Anybody remember L'Hopital's Rule? LOL! Anyway... is that a start at least? The way you're depicting skewness and kurtosis is different from the way I've seen it. I'm not too sure about the "expected value" part you're referring to. I thought that kurtosis and skewness are dependent upon the number of data points in your data set. The way you have skewness represented tells me the only way to arrive at a negative skewness is when your "expected value" is negative because your numerator and denominator are both being raised to a positive power. So how do you get this "expected value"? Anyway... interesting way to start a morning that I'm iced in... weeeee! ------------------ If you like larger booties, Domanique Sachsa would be on that list as well. --Jeff of SaveMeSomeBooty.com fame in the BBS Hangout forum (hope the man doesn't run for a political position - he's dead meat)
They are backwards, yes... We talked in class somewhat baout sigma going to zero and its affects and the teacher made it seem like it was unclear and just because sigma may be zero doesnt mean, for example, that skewness is infinity, etc, etc... but may be on the right track anyway, the help is appreciated... thanks! ------------------ "There are some frauds so well conducted that it would be stupidity not to be deceived by them." Charles Caleb Colton (1780-1832)
It's definitely not determined to be 0 or infinity. Zero divided by zero is what's called an indeterminate form. This is where I was guessing you'd have to use something like L'Hopital's Rule. But in using that rule you have to make sure the numerator and denominator are some function of x and then take the first derivative of each. Then you take the limit of (the derivative of the numerator divided by the derivative of the denominator) as x approaches some value. You then can get what value the function approaches as x approaches that value. Also... you may want to make sure your equations are right. I still say that "E" thingy is really a summation as both kurtosis and skewness are based upon summation of (n-x)/s where n is a data value in a set, x is the mean value of that set, and s is the std. dev. ------------------ If you like larger booties, Domanique Sachsa would be on that list as well. --Jeff of SaveMeSomeBooty.com fame in the BBS Hangout forum (hope the man doesn't run for a political position - he's dead meat)
E is definitely expected value. The first moment, for example, is just E(X) which is just the mean. You are somewhat correct in thinking about summations though as, say , if you have a binomial, where X= X1+X2+X3+...Xn, then the expected value, E(X) = sum [E(Xi)] ,etc,etc I am aware of LHospitals rule and have been looking at it. ------------------ "There are some frauds so well conducted that it would be stupidity not to be deceived by them." Charles Caleb Colton (1780-1832)
Heres a math question that is even more difficult. Shawn Kemps weight times five. ------------------ Ceo of the Walt Williams fan club. Web site coming soon atheistalliance.org
heres one im having trouble with: determine the period: f(x):-2/3cos(x/3-1/2) ------------------ The next time I have meat and mashed potatoes, I think I'll put a very large blob of potatoes on my plate with just a little piece of meat. And if someone asks me why i didn't get more meat, ill just say, "Oh, you mean this?" and pull out a big piece of meat from inside the blob of potatoes, where ive hidden it. Good magic trick, huh?
Add a new Forum in the BBS Hangout category: Hangout, Fantasy League, and Homework. Everyone can post his homework in the new homework forum.
