[R] HPD credible sets
Hi R users Is there a function in R that gives HPD credible sets. i googled it but was in vain! - dev __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] gamma distribution
thanks for your response. btw i am calculating the power of the wilcoxon test. i divide the total no. of rejections by the no. of simulations. so for 1000 simulations, at 0.05 level of significance if the no. of rejections are 50 then the power will be 50/1000 = 0.05. thats y im importing in excel the p values. is my approach correct?? thanks n regards -dev Quoting Uwe Ligges [EMAIL PROTECTED]: Answering both messges here: 1. [EMAIL PROTECTED] wrote: Hi I appreciate your response. This is what I observed..taking the log transform of the raw gamma does change the p value of the test. That is what I am importing into excel (the p - values) Well, so you made a mistake! And I still do not know why anybody realy want to import data to Excel, if the data is already in R. For me, the results are identical (and there is no reason why not). and then calculating the power of the test (both raw and transformed). can you tell me what exactly your code is doing? See below. 2. [EMAIL PROTECTED] wrote: Hi I ran your code. I think it should give me the number of p values below 0.05 significance level (thats what i could understand from your code), but after running your code there is neither any error that shows up nor any value that the console displays. You are right in the point what the code I sent does: erg - replicate(1000, { x-rgamma(10, 2.5, scale = 10) y-rgamma(10, 2.5, scale = 10) wilcox.test(x, y, var.equal = FALSE)$p.value }) sum(erg 0.05) # 45 and it works for me. It results in a random number close to 50, hopefully. Since both points above seem to be very strange on your machine: Which version of R are you using? We assume the most recent one which is R-2.1.1. Uwe Ligges __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] gamma distribution
Hi Christopher and Uwe. thanks for your time and guidance. I deeply appreciate it. -dev Quoting Christoph Buser [EMAIL PROTECTED]: Hi As Uwe mentioned be careful about the difference the significance level alpha and the power of a test. To do power calculations you should specify and alternative hypothesis H_A, e.g. if you have two populations you want to compare and we assume that they are normal distributed (equal unknown variance for simplicity). We are interested if there is a difference in the mean and want to use the t.test. Our Null hypothesis H_0: there is no difference in the means To do a power calculation for our test, we first have to specify and alternative H_A: the mean difference is 1 (unit) Now for a fix number of observations we can calculate the power of our test, which is in that case the probability that (if the true unknown difference is 1, meaning that H_A is true) our test is significant, meaning if I repeat the test many times (always taking samples with mean difference of 1), the number of significant test divided by the total number of tests is an estimate for the power. In you case the situation is a little bit more complicated. You need to specify an alternative hypothesis. In one of your first examples you draw samples from two gamma distributions with different shape parameter and the same scale. But by varying the shape parameter the two distributions not only differ in their mean but also in their form. I got an email from Prof. Ripley in which he explained in details and very precise some examples of tests and what they are testing. It was in addition to the first posts about t tests and wilcoxon test. I attached the email below and recommend to read it carefully. It might be helpful for you, too. Regards, Christoph Buser -- Christoph Buser [EMAIL PROTECTED] Seminar fuer Statistik, LEO C13 ETH (Federal Inst. Technology)8092 Zurich SWITZERLAND phone: x-41-44-632-4673 fax: 632-1228 http://stat.ethz.ch/~buser/ -- From: Prof Brian Ripley [EMAIL PROTECTED] To: Christoph Buser [EMAIL PROTECTED] cc: Liaw, Andy [EMAIL PROTECTED] Subject: Re: [R] Alternatives to t-tests (was Code Verification) Date: Thu, 21 Jul 2005 10:33:28 +0100 (BST) I believe there is a rather more to this than Christoph's account. The Wilcoxon test is not testing the same null hypothesis as the t-test, and that may very well matter in practice and it does in the example given. The (default in R) Welch t-test tests a difference in means between two samples, not necessarily of the same variance or shape. A difference in means is simple to understand, and is unambiguously defined at least if the distributions have means, even for real-life long-tailed distributions. Inference from the t-test is quite accurate even a long way from normality and from equality of the shapes of the two distributions, except in very small sample sizes. (I point my beginning students at the simulation study in `The Statistical Sleuth' by Ramsey and Schafer, stressing that the unequal-variance t-test ought to be the default choice as it is in R. So I get them to redo the simulations.) The Wilcoxon test tests a shift in location between two samples from distributions of the same shape differing only by location. Having the same shape is part of the null hypothesis, and so is an assumption that needs to be verified if you want to conclude there is a difference in location (e.g. in means). Even if you assume symmetric distributions (so the location is unambiguously defined) the level of the test depends on the shapes, tending to reject equality of location in the presence of difference of shape. So you really are testing equality of distribution, both location and shape, with power concentrated on location-shift alternatives. Given samples from a gamma(shape=2) and gamma(shape=20) distributions, we know what the t-test is testing (equality of means). What is the Wilcoxon test testing? Something hard to describe and less interesting, I believe. BTW, I don't see the value of the gamma simulation as this simultaneously changes mean and shape between the samples. How about checking holding the mean the same: n - 1000 z1 - z2 - numeric(n) for (i in 1:n) { x - rgamma(40, 2.5, 0.1) y - rgamma(40, 10, 0.1*10/2.5) z1[i] - t.test(x, y)$p.value z2[i] - wilcox.test(x, y)$p.value } ## Level 1 - sum(z10.05)/1000 ## 0.049 1 - sum(z20.05)/1000 ## 0.15 ? -- the Wilcoxon test is shown to be a poor test of equality of means. Christoph's simulation shows that it is able to use difference in shape as well as location in the test of these two distributions, whereas the t-test is designed only to use the
Re: [R] gamma distribution
Hi You are right but here I am taking into account the p values I get from the tests on the raw and the transformed samples. And then I calculate the power of the tests based on the # of rejections of the p values. DO you think its a good way to determine the power of a test? thanks -dev Quoting Christoph Buser [EMAIL PROTECTED]: Hi I am a little bit confused. You create two sample (from a gamma distribution) and you do a wilcoxon test with this two samples. Then you use the same monotone transformation (log) for both samples and redo the wilcoxon test. But since the transformations keeps the order of your samples the second wilcoxon test is identical to the first one: x-rgamma(10, 2.5, scale = 10) y-rgamma(10, 2.5, scale = 10) wilcox.test(x, y, var.equal = FALSE) x1-log(x) y1-log(y) wilcox.test(x1, y1, var.equal = FALSE) Maybe you can give some more details about the hypothesis you'd like to test. Regards, Christoph Buser -- Christoph Buser [EMAIL PROTECTED] Seminar fuer Statistik, LEO C13 ETH (Federal Inst. Technology)8092 Zurich SWITZERLAND phone: x-41-44-632-4673 fax: 632-1228 http://stat.ethz.ch/~buser/ -- [EMAIL PROTECTED] writes: Hi R Users This is a code I wrote and just want to confirm if the first 1000 values are raw gamma (z) and the next 1000 values are transformed gamma (k) or not. As I get 2000 rows once I import into excel, the p - values beyond 1000 dont look that good, they are very high. -- sink(a1.txt); for (i in 1:1000) { x-rgamma(10, 2.5, scale = 10) y-rgamma(10, 2.5, scale = 10) z-wilcox.test(x, y, var.equal = FALSE) print(z) x1-log(x) y1-log(y) k-wilcox.test(x1, y1, var.equal = FALSE) print(k) } --- any suggestions are welcome thanks -devarshi __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Re: [R] gamma distribution
Hi I ran your code. I think it should give me the number of p values below 0.05 significance level (thats what i could understand from your code), but after running your code there is neither any error that shows up nor any value that the console displays. thanks in advance -dev. Quoting Uwe Ligges [EMAIL PROTECTED]: [EMAIL PROTECTED] wrote: Hi R Users This is a code I wrote and just want to confirm if the first 1000 values are raw gamma (z) and the next 1000 values are transformed gamma (k) or not. As I get 2000 rows once I import into excel, the p - values beyond 1000 dont look that good, they are very high. He? - log() transforming the data does not change the Wilcoxon statistics (based on ranks!)! - Why is this related to Excel? - What are you going to show? I get erg - replicate(1000, { x-rgamma(10, 2.5, scale = 10) y-rgamma(10, 2.5, scale = 10) wilcox.test(x, y, var.equal = FALSE)$p.value }) sum(erg 0.05) # 45 which seems plausible to me. Uwe Ligges -- sink(a1.txt); for (i in 1:1000) { x-rgamma(10, 2.5, scale = 10) y-rgamma(10, 2.5, scale = 10) z-wilcox.test(x, y, var.equal = FALSE) print(z) x1-log(x) y1-log(y) k-wilcox.test(x1, y1, var.equal = FALSE) print(k) } --- any suggestions are welcome thanks -devarshi __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] gamma distribution
Hi R Users This is a code I wrote and just want to confirm if the first 1000 values are raw gamma (z) and the next 1000 values are transformed gamma (k) or not. As I get 2000 rows once I import into excel, the p - values beyond 1000 dont look that good, they are very high. -- sink(a1.txt); for (i in 1:1000) { x-rgamma(10, 2.5, scale = 10) y-rgamma(10, 2.5, scale = 10) z-wilcox.test(x, y, var.equal = FALSE) print(z) x1-log(x) y1-log(y) k-wilcox.test(x1, y1, var.equal = FALSE) print(k) } --- any suggestions are welcome thanks -devarshi __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] A Question About Inverse Gamma
Hi R users, I am having a little problem finding the the solution to this problem in R: 1. I need to generate normal distribution of sample size 30, mean = 50, sd = 5. 2. From the statistics obtained in step 1, I need to generate the Inverse Gamma distribution. Your views and help will be appreciated. __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] Code Verification
Hi R Users I have a code which I am running for my thesis work. Just want to make sure that its ok. Its a t test I am conducting between two gamma distributions with different shape parameters. the code looks like: sink(a1.txt); for (i in 1:1000) { x-rgamma(40, 2.5, 10) # n = 40, shape = 2.5, Scale = 10 y-rgamma(40, 2.8, 10) # n = 40, shape = 2.8, Scale = 10 z-t.test(x, y) print(z) } I will appreciate it if someone could tell me if its alrite or not. thanks -dev __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] question from environmental statistics
thanks Fran. that was useful but Im still in a fix. its a real life data which looks like this: 0.9 10.9 24.0 6.7 0.6 1.0 2.4 12.4 7.9 15.8 1.4 7.9 11000.0 (benzene conc. taken after WTC attacks)..its just a small chunk of data i pasted for you to look at. its neither normal nor lognormal. someone told me that qq plot does help in determining the distribution. im not sure how to get it. can someone help me in this. thanks Take a look at this document by Vito Ricci: http://cran.r-project.org/doc/contrib/Ricci-distributions-en.pdf Did you try RSiteSearch(Fit distribution) or a Google search? That will lead you to fit.dist{gnlm} and fitdistr{MASS} Cheers Francisco From: [EMAIL PROTECTED] To: r-help@stat.math.ethz.ch Subject: [R] question from environmental statistics Date: Thu, 14 Jul 2005 14:06:45 -0700 Dear R users I want to knw if there is a way in which a raw dataset can be modelled by some distribution. besides the gof test is there any test involving gamma or lognormal that would fit the data. thank you -dev __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
[R] question from environmental statistics
Dear R users I want to knw if there is a way in which a raw dataset can be modelled by some distribution. besides the gof test is there any test involving gamma or lognormal that would fit the data. thank you -dev __ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html