thevalley Posted December 22, 2008 Report Share Posted December 22, 2008 (edited) This IA asks to develop a probability density function (pdf) for the chart described below. I'm stuck at the question which says "Add two other columns to the chart, entering the excpected frequencies for the 1000 saplings for both normal and Poisson distribution. Test each as a possiblemodel for the data"The chart is:[Height of saps in m],[Frequency]0.00-0.25 , 610.25-0.50 , 1600.50-0.75 , 2090.75-1.00 . 2021.00-1.25 , 1581.25-1.50 , 1051.50-1.75 , 581.75-2.00 , 292.00-2.25 , 122.25-2.50 , 42.50-2.75 , 12.75-3.00 , 1I have a problem with the Poisson distribution, what am I exactly supposed to do?I calculated the average height of the trees which is 0.88625, and i labelled the height to X{0.00-0.25=1, 0.25-0.50=2, 0.50-0.75=3,....} and use that in the Poisson distribution formula. The curve corresponds to the chart's curve if I plot it, but the values are very off. Eg.1.00-1.25, X=5, mean=0.88625[(0.88625^5)e^(-0.88625)]/(5!) = 0.001870472 -> E(x) = 0.001870472 * 1000 = 1.87 is not equal or close to 158What am I doing wrong? Edited December 22, 2008 by thevalley Reply Link to post Share on other sites More sharing options...
HelpMeOnTOK Posted December 30, 2008 Report Share Posted December 30, 2008 I am working on Math Portfolio Type II, Modelling the Heights of Saplings in case you don't know.The portfolio seems ok except i got confused on Question #5-65. A useful method for finding a pdf for a given data set is to begin with a function that best describes the data's associated cumulative relative frequency curve. What traits must this function possess in order to be defined as a cumulative density function (cdf) for the pdf you are creating?6. Find a suitable function for this cdf, check it to see if it has the necessary traits, and refine it if necessary.What is a cdf??? and when they meant by function, they mean like statistics kind of function or a function (linear, quadratic, exponential, etc.)?thanks Reply Link to post Share on other sites More sharing options...
Banana Posted December 31, 2008 Report Share Posted December 31, 2008 I am working on the same portfolio but i have problems solving the second question .I am getting the mean by multiplying the height X=0 ,1 , 2, 3, 4, and so on and the corresponding frequency.I got a mean which is 0.88625.I start working with the poisson distribution.so, (0.88625^0 x e^-0.88625) divided by 0! = 0.412 which is by no means near the frequency (61).I think i am doing something wrong so may you please give me any ideas on how to tackle this out.Any help would be appreciated.And for the (cdf), i think they meant if the data can be constructed as a cumulative frequency distribution and from then it would be a cdf but there has to be some characteristics for that cumulativ curve to be a cdf. Reply Link to post Share on other sites More sharing options...
HelpMeOnTOK Posted January 2, 2009 Report Share Posted January 2, 2009 Well,it is a good idea you used a unique kind of interval instead of the actual values of sapling heights (ie. .25m) because decimals nver work in factorials...What i think you should do is to first finish the poisson distribution table of values and then graph it out to see how it looks, should look similar to the actual graph only smaller in relative size because the values are in decimals...Then I think you should try magnifying the graph so that it would fit the actual values approximately. In other wrds, multiply the poisson distribution function by any value to increase its values closer to the actual ones. I don't think it is wrong to do that, though you are changing the values of teh distribution, you still have the correct distribution..Hope that helped.I am trying to work out a function here for the cdf, if i haven't done it wrong it should be a cumulative frequency curve, only using relative values...I can't seem to find a function that i know that can fit the data properly...cuz most of those functions can have negative x-values..do I have to set a domain for it?? and what function works best for this..I tried log..didn't seem to work well.I am working on the same portfolio but i have problems solving the second question .I am getting the mean by multiplying the height X=0 ,1 , 2, 3, 4, and so on and the corresponding frequency.I got a mean which is 0.88625.I start working with the poisson distribution.so, (0.88625^0 x e^-0.88625) divided by 0! = 0.412 which is by no means near the frequency (61).I think i am doing something wrong so may you please give me any ideas on how to tackle this out.Any help would be appreciated.And for the (cdf), i think they meant if the data can be constructed as a cumulative frequency distribution and from then it would be a cdf but there has to be some characteristics for that cumulativ curve to be a cdf. Reply Link to post Share on other sites More sharing options...
Munesh Yagna Posted February 7, 2009 Report Share Posted February 7, 2009 even i am not able to find a function for the cdf (Q6).....please some one help me.... Reply Link to post Share on other sites More sharing options...
moneyfaery Posted February 7, 2009 Report Share Posted February 7, 2009 (edited) cdf stands for cumulative (probability) distribution function, if I'm not mistaken e.g. normalcdf, binomcdf (on GDC)pdf is probability distribution functionWhat is a cdf??? and when they meant by function, they mean like statistics kind of function or a function (linear, quadratic, exponential, etc.)?Well, think about it. You're dealing with a huge amount of data that corresponds to a certain random variable, so the function has to be stats/probability related. Edited February 7, 2009 by Irene Reply Link to post Share on other sites More sharing options...
Munesh Yagna Posted February 8, 2009 Report Share Posted February 8, 2009 thanks for the clue i will try my best to find the function........ Reply Link to post Share on other sites More sharing options...
d00m5day Posted February 12, 2009 Report Share Posted February 12, 2009 If you guys have a graphing calculator, you should have that function... Reply Link to post Share on other sites More sharing options...
athena99 Posted February 25, 2009 Report Share Posted February 25, 2009 how did you get the mean of .88625 using X=0,1,2,...?every time i calculate it using those numbers for the x value i do not come up with this number.I am working on the same portfolio but i have problems solving the second question .I am getting the mean by multiplying the height X=0 ,1 , 2, 3, 4, and so on and the corresponding frequency.I got a mean which is 0.88625. Reply Link to post Share on other sites More sharing options...
FML Posted February 26, 2009 Report Share Posted February 26, 2009 how did you get the mean of .88625 using X=0,1,2,...?every time i calculate it using those numbers for the x value i do not come up with this number.I am working on the same portfolio but i have problems solving the second question .I am getting the mean by multiplying the height X=0 ,1 , 2, 3, 4, and so on and the corresponding frequency.I got a mean which is 0.88625.You get the mean of .88625 when you multiply the frequencies to the midpoints of the intervals, add'em all up and divide by the total frequency.Giving X the values 0-11 and then multiplying to the frequencies and getting .88625 is not right at all since giving the X valus 0-11 is in order to derive the Poisson mean... 1 Reply Link to post Share on other sites More sharing options...
athena99 Posted February 26, 2009 Report Share Posted February 26, 2009 You get the mean of .88625 when you multiply the frequencies to the midpoints of the intervals, add'em all up and divide by the total frequency.Giving X the values 0-11 and then multiplying to the frequencies and getting .88625 is not right at all since giving the X valus 0-11 is in order to derive the Poisson mean... ok. i was just confused or i read it wrong. i'm getting stuck on question 2 and my teacher is not helping. i was wondering what the justification of changing the X values was and did you also change if for the normal dist? Reply Link to post Share on other sites More sharing options...
Munesh Yagna Posted March 8, 2009 Report Share Posted March 8, 2009 Hi everyone......i think i have gone ahead of u guys.....i am in Q7.....if anyone needs help...then message me....i will try to help u out....lol...... Reply Link to post Share on other sites More sharing options...
agruveee Posted March 8, 2009 Report Share Posted March 8, 2009 i have just been issued with this project and i am wondering how to find the expected frequency for normal and poisson distribultion. How are the expected frequency releated to the expected values? And what are the probability density functions?... sorry i got the worst grade on this topic and i am really lost....... plz help Reply Link to post Share on other sites More sharing options...
Ongfufu Posted March 15, 2009 Report Share Posted March 15, 2009 What's a good graphing software to use for this portfolio? Reply Link to post Share on other sites More sharing options...
deissi Posted March 15, 2009 Report Share Posted March 15, 2009 Excel or Graphical Analysis. Reply Link to post Share on other sites More sharing options...
Ongfufu Posted March 17, 2009 Report Share Posted March 17, 2009 Oh God, I'm stuck on #2. How do I find the expected frequencies for normal and poisson distributions? I looked up expected frequency for normal distribution, and I ran in to the Chi-Squared Test. I have no idea what that is. Does it help in this IA? Reply Link to post Share on other sites More sharing options...
Munesh Yagna Posted March 18, 2009 Report Share Posted March 18, 2009 expected frequwncy for normal distribution can be found by using gdc.....normalcdf(lower bound, upper bound, mean, standard deviation) = some answerand now multiply that 'some answer' with 1000.......hope it helps....and graphatica is another softwarewhich can be used.....lol.... Reply Link to post Share on other sites More sharing options...
Ongfufu Posted March 18, 2009 Report Share Posted March 18, 2009 Should there be 12 expected frequencies or just 1? Reply Link to post Share on other sites More sharing options...
Munesh Yagna Posted March 18, 2009 Report Share Posted March 18, 2009 there should be 12....and i don't think we can calculate expected frequencies for poisson distribution b'causeit is only applicable for discrete random variable.............and that software is 'graphmatica' and not 'graphatica'..... Reply Link to post Share on other sites More sharing options...
Mahuta ♥ Posted March 18, 2009 Report Share Posted March 18, 2009 b'causeIsnt this the same as putting an 'e' instead of and apostrophy(sp?)?Please do not do this again, no text speaking is allowed. Reply Link to post Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.