Re: [R] about a p-value < 2.2e-16
thanks a lot, Jiefei ! and thanks to all for your time and comments ! have a good weekend ! On Fri, Mar 19, 2021 at 10:01 PM Jiefei Wang wrote: > Hi Bogdan, > > I think the journal is asking about the exact value of the pvalue, it > doesn't matter if it is from the exact distribution or normal > approximation. However, it does not make any sense to report such a small > pvlaue. If I was you, I would show the reviewers the exact pvalue they want > and gently explain why you did not put it into your paper. If they insist > that the number must be on the paper, then go ahead and do it. > > Best, > Jiefei > > > > Bogdan Tanasa 于 2021年3月20日周六 上午2:39写道: > >> Thank you Kevin, their wording is "Please note that the exact p value >> should be provided, when possible, etc" >> >> by "exact p-value" i believe that they do mean indeed the actual number, >> and not to specify "exact=TRUE" ; >> >> as we are working with 1000 genes, shall i specify "exact=TRUE" on my PC, >> it runs out of memory ... >> >> wilcox.test(rnorm(1000), rnorm(1000, 2), exact=TRUE)$p.value >> >> On Fri, Mar 19, 2021 at 11:10 AM Kevin Thorpe >> wrote: >> >> > I have to ask since. Are you sure the journal simply means by exact >> > p-value that they don’t want to see a p-value given as < 0.0001, for >> > example, and simply want the actual number? >> > >> > I cannot imagine they really meant exact as in the p-value from some >> exact >> > distribution. >> > >> > -- >> > Kevin E. Thorpe >> > Head of Biostatistics, Applied Health Research Centre (AHRC) >> > Li Ka Shing Knowledge Institute of St. Michael's >> > Assistant Professor, Dalla Lana School of Public Health >> > University of Toronto >> > email: kevin.tho...@utoronto.ca Tel: 416.864.5776 Fax: 416.864.3016 >> > >> > > On Mar 19, 2021, at 1:22 PM, Bogdan Tanasa wrote: >> > > >> > > EXTERNAL EMAIL: >> > > >> > > Dear all, thank you all for comments and help. >> > > >> > > as far as i can see, shall we have samples of 1000 records, only >> > > "exact=FALSE" allows the code to run: >> > > >> > > wilcox.test(rnorm(1000), rnorm(1000, 2), exact=FALSE)$p.value >> > > [1] 7.304863e-231 >> > > >> > > shall i use "exact=TRUE", it runs out of memory on my 64GB RAM PC : >> > > >> > > wilcox.test(rnorm(1000), rnorm(1000, 2), exact=TRUE)$p.value >> > > (the job is terminated by OS) >> > > >> > > shall you have any other suggestions, please let me know. thanks a >> lot ! >> > > >> > > On Fri, Mar 19, 2021 at 9:05 AM Bert Gunter >> > wrote: >> > > >> > >> I **believe** -- if my old memory still serves-- that the "exact" >> > >> specification uses a home grown version of the algorithm to calculate >> > >> exact, or close approximations to the exact, permutation >> distribution >> > >> originally developed by Cyrus Mehta, founder of StatXact software. >> Of >> > >> course, examining the C code source would determine this, but I don't >> > care >> > >> to attempt this. >> > >> >> > >> If this is (no longer?) correct, please point this out. >> > >> >> > >> Best, >> > >> >> > >> Bert Gunter >> > >> >> > >> "The trouble with having an open mind is that people keep coming >> along >> > and >> > >> sticking things into it." >> > >> -- Opus (aka Berkeley Breathed in his "Bloom County" comic strip ) >> > >> >> > >> >> > >> On Fri, Mar 19, 2021 at 8:42 AM Jiefei Wang >> wrote: >> > >> >> > >>> Hi Spencer, >> > >>> >> > >>> Thanks for your test results, I do not know the answer as I haven't >> > >>> used wilcox.test for many years. I do not know if it is possible to >> > >>> compute >> > >>> the exact distribution of the Wilcoxon rank sum statistic, but I >> think >> > it >> > >>> is very likely, as the document of `Wilcoxon` says: >> > >>> >> > >>> This distribution is obtained as follows. Let x and y be two random, >> > >>> independent samples of size m and n. Then the Wilcoxon rank sum >> > statistic >> > >>> is the number of all pairs (x[i], y[j]) for which y[j] is not >> greater >> > than >> > >>> x[i]. This statistic takes values between 0 and m * n, and its mean >> and >> > >>> variance are m * n / 2 and m * n * (m + n + 1) / 12, respectively. >> > >>> >> > >>> As a nice feature of the non-parametric statistic, it is usually >> > >>> distribution-free so you can pick any distribution you like to >> compute >> > the >> > >>> same statistic. I wonder if this is the case, but I might be wrong. >> > >>> >> > >>> Cheers, >> > >>> Jiefei >> > >>> >> > >>> >> > >>> On Fri, Mar 19, 2021 at 10:57 PM Spencer Graves < >> > >>> spencer.gra...@effectivedefense.org> wrote: >> > >>> >> > >> > >> > On 2021-3-19 9:52 AM, Jiefei Wang wrote: >> > > After digging into the R source, it turns out that the argument >> > >>> `exact` >> > has >> > > nothing to do with the numeric precision. It only affects the >> > >>> statistic >> > > model used to compute the p-value. When `exact=TRUE` the true >> > distribution >> > > of the statistic will be used. Otherwise, a normal
Re: [R] about a p-value < 2.2e-16
Hi Bogdan, I think the journal is asking about the exact value of the pvalue, it doesn't matter if it is from the exact distribution or normal approximation. However, it does not make any sense to report such a small pvlaue. If I was you, I would show the reviewers the exact pvalue they want and gently explain why you did not put it into your paper. If they insist that the number must be on the paper, then go ahead and do it. Best, Jiefei Bogdan Tanasa 于 2021年3月20日周六 上午2:39写道: > Thank you Kevin, their wording is "Please note that the exact p value > should be provided, when possible, etc" > > by "exact p-value" i believe that they do mean indeed the actual number, > and not to specify "exact=TRUE" ; > > as we are working with 1000 genes, shall i specify "exact=TRUE" on my PC, > it runs out of memory ... > > wilcox.test(rnorm(1000), rnorm(1000, 2), exact=TRUE)$p.value > > On Fri, Mar 19, 2021 at 11:10 AM Kevin Thorpe > wrote: > > > I have to ask since. Are you sure the journal simply means by exact > > p-value that they don’t want to see a p-value given as < 0.0001, for > > example, and simply want the actual number? > > > > I cannot imagine they really meant exact as in the p-value from some > exact > > distribution. > > > > -- > > Kevin E. Thorpe > > Head of Biostatistics, Applied Health Research Centre (AHRC) > > Li Ka Shing Knowledge Institute of St. Michael's > > Assistant Professor, Dalla Lana School of Public Health > > University of Toronto > > email: kevin.tho...@utoronto.ca Tel: 416.864.5776 Fax: 416.864.3016 > > > > > On Mar 19, 2021, at 1:22 PM, Bogdan Tanasa wrote: > > > > > > EXTERNAL EMAIL: > > > > > > Dear all, thank you all for comments and help. > > > > > > as far as i can see, shall we have samples of 1000 records, only > > > "exact=FALSE" allows the code to run: > > > > > > wilcox.test(rnorm(1000), rnorm(1000, 2), exact=FALSE)$p.value > > > [1] 7.304863e-231 > > > > > > shall i use "exact=TRUE", it runs out of memory on my 64GB RAM PC : > > > > > > wilcox.test(rnorm(1000), rnorm(1000, 2), exact=TRUE)$p.value > > > (the job is terminated by OS) > > > > > > shall you have any other suggestions, please let me know. thanks a lot > ! > > > > > > On Fri, Mar 19, 2021 at 9:05 AM Bert Gunter > > wrote: > > > > > >> I **believe** -- if my old memory still serves-- that the "exact" > > >> specification uses a home grown version of the algorithm to calculate > > >> exact, or close approximations to the exact, permutation distribution > > >> originally developed by Cyrus Mehta, founder of StatXact software. Of > > >> course, examining the C code source would determine this, but I don't > > care > > >> to attempt this. > > >> > > >> If this is (no longer?) correct, please point this out. > > >> > > >> Best, > > >> > > >> Bert Gunter > > >> > > >> "The trouble with having an open mind is that people keep coming along > > and > > >> sticking things into it." > > >> -- Opus (aka Berkeley Breathed in his "Bloom County" comic strip ) > > >> > > >> > > >> On Fri, Mar 19, 2021 at 8:42 AM Jiefei Wang > wrote: > > >> > > >>> Hi Spencer, > > >>> > > >>> Thanks for your test results, I do not know the answer as I haven't > > >>> used wilcox.test for many years. I do not know if it is possible to > > >>> compute > > >>> the exact distribution of the Wilcoxon rank sum statistic, but I > think > > it > > >>> is very likely, as the document of `Wilcoxon` says: > > >>> > > >>> This distribution is obtained as follows. Let x and y be two random, > > >>> independent samples of size m and n. Then the Wilcoxon rank sum > > statistic > > >>> is the number of all pairs (x[i], y[j]) for which y[j] is not greater > > than > > >>> x[i]. This statistic takes values between 0 and m * n, and its mean > and > > >>> variance are m * n / 2 and m * n * (m + n + 1) / 12, respectively. > > >>> > > >>> As a nice feature of the non-parametric statistic, it is usually > > >>> distribution-free so you can pick any distribution you like to > compute > > the > > >>> same statistic. I wonder if this is the case, but I might be wrong. > > >>> > > >>> Cheers, > > >>> Jiefei > > >>> > > >>> > > >>> On Fri, Mar 19, 2021 at 10:57 PM Spencer Graves < > > >>> spencer.gra...@effectivedefense.org> wrote: > > >>> > > > > > > On 2021-3-19 9:52 AM, Jiefei Wang wrote: > > > After digging into the R source, it turns out that the argument > > >>> `exact` > > has > > > nothing to do with the numeric precision. It only affects the > > >>> statistic > > > model used to compute the p-value. When `exact=TRUE` the true > > distribution > > > of the statistic will be used. Otherwise, a normal approximation > will > > >>> be > > > used. > > > > > > I think the documentation needs to be improved here, you can > compute > > >>> the > > > exact p-value *only* when you do not have any ties in your data. If > > >>> you > > > have ties in your data you will get the p-value from the normal > >
Re: [R] about a p-value < 2.2e-16
Yes, Bogdan, that sounds *exactly* right. ;-) -- it runs out of memory trying to calculate the exact permutation distribution. What you apparently get with exact = FALSE is the exact answer( to within floating point arithmetic's approximation) to a normal approximation. ... and furthermore... I would imagine any random number below, say, 1e-100 would serve equally well and would be equally correct/incorrect. I also imagine that a sensible display of the paired differences or even just a count of how many of the thousand are, say, >0, would make even more sense than an overwrought and unnecessary p-value. But that is just my personal opinion of senseless standard scientific practice, and if anyone want to dispute it, please reply OFFLIST, though I would probably not disagree with any such criticism of my cynicism. Bert Gunter "The trouble with having an open mind is that people keep coming along and sticking things into it." -- Opus (aka Berkeley Breathed in his "Bloom County" comic strip ) On Fri, Mar 19, 2021 at 10:22 AM Bogdan Tanasa wrote: > Dear all, thank you all for comments and help. > > as far as i can see, shall we have samples of 1000 records, only > "exact=FALSE" allows the code to run: > > wilcox.test(rnorm(1000), rnorm(1000, 2), exact=FALSE)$p.value > [1] 7.304863e-231 > > shall i use "exact=TRUE", it runs out of memory on my 64GB RAM PC : > > wilcox.test(rnorm(1000), rnorm(1000, 2), exact=TRUE)$p.value > (the job is terminated by OS) > > shall you have any other suggestions, please let me know. thanks a lot ! > > On Fri, Mar 19, 2021 at 9:05 AM Bert Gunter > wrote: > >> I **believe** -- if my old memory still serves-- that the "exact" >> specification uses a home grown version of the algorithm to calculate >> exact, or close approximations to the exact, permutation distribution >> originally developed by Cyrus Mehta, founder of StatXact software. Of >> course, examining the C code source would determine this, but I don't care >> to attempt this. >> >> If this is (no longer?) correct, please point this out. >> >> Best, >> >> Bert Gunter >> >> "The trouble with having an open mind is that people keep coming along >> and sticking things into it." >> -- Opus (aka Berkeley Breathed in his "Bloom County" comic strip ) >> >> >> On Fri, Mar 19, 2021 at 8:42 AM Jiefei Wang wrote: >> >>> Hi Spencer, >>> >>> Thanks for your test results, I do not know the answer as I haven't >>> used wilcox.test for many years. I do not know if it is possible to >>> compute >>> the exact distribution of the Wilcoxon rank sum statistic, but I think it >>> is very likely, as the document of `Wilcoxon` says: >>> >>> This distribution is obtained as follows. Let x and y be two random, >>> independent samples of size m and n. Then the Wilcoxon rank sum statistic >>> is the number of all pairs (x[i], y[j]) for which y[j] is not greater >>> than >>> x[i]. This statistic takes values between 0 and m * n, and its mean and >>> variance are m * n / 2 and m * n * (m + n + 1) / 12, respectively. >>> >>> As a nice feature of the non-parametric statistic, it is usually >>> distribution-free so you can pick any distribution you like to compute >>> the >>> same statistic. I wonder if this is the case, but I might be wrong. >>> >>> Cheers, >>> Jiefei >>> >>> >>> On Fri, Mar 19, 2021 at 10:57 PM Spencer Graves < >>> spencer.gra...@effectivedefense.org> wrote: >>> >>> > >>> > >>> > On 2021-3-19 9:52 AM, Jiefei Wang wrote: >>> > > After digging into the R source, it turns out that the argument >>> `exact` >>> > has >>> > > nothing to do with the numeric precision. It only affects the >>> statistic >>> > > model used to compute the p-value. When `exact=TRUE` the true >>> > distribution >>> > > of the statistic will be used. Otherwise, a normal approximation >>> will be >>> > > used. >>> > > >>> > > I think the documentation needs to be improved here, you can compute >>> the >>> > > exact p-value *only* when you do not have any ties in your data. If >>> you >>> > > have ties in your data you will get the p-value from the normal >>> > > approximation no matter what value you put in `exact`. This behavior >>> > should >>> > > be documented or a warning should be given when `exact=TRUE` and ties >>> > > present. >>> > > >>> > > FYI, if the exact p-value is required, `pwilcox` function will be >>> used to >>> > > compute the p-value. There are no details on how it computes the >>> pvalue >>> > but >>> > > its C code seems to compute the probability table, so I assume it >>> > computes >>> > > the exact p-value from the true distribution of the statistic, not a >>> > > permutation or MC p-value. >>> > >>> > >>> >My example shows that it does NOT use Monte Carlo, because >>> > otherwise it uses some distribution. I believe the term "exact" means >>> > that it uses the permutation distribution, though I could be mistaken. >>> > If it's NOT a permutation distribution, I don't know what it is. >>> > >>> > >>> >Spencer
Re: [R] about a p-value < 2.2e-16
Thank you Kevin, their wording is "Please note that the exact p value should be provided, when possible, etc" by "exact p-value" i believe that they do mean indeed the actual number, and not to specify "exact=TRUE" ; as we are working with 1000 genes, shall i specify "exact=TRUE" on my PC, it runs out of memory ... wilcox.test(rnorm(1000), rnorm(1000, 2), exact=TRUE)$p.value On Fri, Mar 19, 2021 at 11:10 AM Kevin Thorpe wrote: > I have to ask since. Are you sure the journal simply means by exact > p-value that they don’t want to see a p-value given as < 0.0001, for > example, and simply want the actual number? > > I cannot imagine they really meant exact as in the p-value from some exact > distribution. > > -- > Kevin E. Thorpe > Head of Biostatistics, Applied Health Research Centre (AHRC) > Li Ka Shing Knowledge Institute of St. Michael's > Assistant Professor, Dalla Lana School of Public Health > University of Toronto > email: kevin.tho...@utoronto.ca Tel: 416.864.5776 Fax: 416.864.3016 > > > On Mar 19, 2021, at 1:22 PM, Bogdan Tanasa wrote: > > > > EXTERNAL EMAIL: > > > > Dear all, thank you all for comments and help. > > > > as far as i can see, shall we have samples of 1000 records, only > > "exact=FALSE" allows the code to run: > > > > wilcox.test(rnorm(1000), rnorm(1000, 2), exact=FALSE)$p.value > > [1] 7.304863e-231 > > > > shall i use "exact=TRUE", it runs out of memory on my 64GB RAM PC : > > > > wilcox.test(rnorm(1000), rnorm(1000, 2), exact=TRUE)$p.value > > (the job is terminated by OS) > > > > shall you have any other suggestions, please let me know. thanks a lot ! > > > > On Fri, Mar 19, 2021 at 9:05 AM Bert Gunter > wrote: > > > >> I **believe** -- if my old memory still serves-- that the "exact" > >> specification uses a home grown version of the algorithm to calculate > >> exact, or close approximations to the exact, permutation distribution > >> originally developed by Cyrus Mehta, founder of StatXact software. Of > >> course, examining the C code source would determine this, but I don't > care > >> to attempt this. > >> > >> If this is (no longer?) correct, please point this out. > >> > >> Best, > >> > >> Bert Gunter > >> > >> "The trouble with having an open mind is that people keep coming along > and > >> sticking things into it." > >> -- Opus (aka Berkeley Breathed in his "Bloom County" comic strip ) > >> > >> > >> On Fri, Mar 19, 2021 at 8:42 AM Jiefei Wang wrote: > >> > >>> Hi Spencer, > >>> > >>> Thanks for your test results, I do not know the answer as I haven't > >>> used wilcox.test for many years. I do not know if it is possible to > >>> compute > >>> the exact distribution of the Wilcoxon rank sum statistic, but I think > it > >>> is very likely, as the document of `Wilcoxon` says: > >>> > >>> This distribution is obtained as follows. Let x and y be two random, > >>> independent samples of size m and n. Then the Wilcoxon rank sum > statistic > >>> is the number of all pairs (x[i], y[j]) for which y[j] is not greater > than > >>> x[i]. This statistic takes values between 0 and m * n, and its mean and > >>> variance are m * n / 2 and m * n * (m + n + 1) / 12, respectively. > >>> > >>> As a nice feature of the non-parametric statistic, it is usually > >>> distribution-free so you can pick any distribution you like to compute > the > >>> same statistic. I wonder if this is the case, but I might be wrong. > >>> > >>> Cheers, > >>> Jiefei > >>> > >>> > >>> On Fri, Mar 19, 2021 at 10:57 PM Spencer Graves < > >>> spencer.gra...@effectivedefense.org> wrote: > >>> > > > On 2021-3-19 9:52 AM, Jiefei Wang wrote: > > After digging into the R source, it turns out that the argument > >>> `exact` > has > > nothing to do with the numeric precision. It only affects the > >>> statistic > > model used to compute the p-value. When `exact=TRUE` the true > distribution > > of the statistic will be used. Otherwise, a normal approximation will > >>> be > > used. > > > > I think the documentation needs to be improved here, you can compute > >>> the > > exact p-value *only* when you do not have any ties in your data. If > >>> you > > have ties in your data you will get the p-value from the normal > > approximation no matter what value you put in `exact`. This behavior > should > > be documented or a warning should be given when `exact=TRUE` and ties > > present. > > > > FYI, if the exact p-value is required, `pwilcox` function will be > >>> used to > > compute the p-value. There are no details on how it computes the > >>> pvalue > but > > its C code seems to compute the probability table, so I assume it > computes > > the exact p-value from the true distribution of the statistic, not a > > permutation or MC p-value. > > > My example shows that it does NOT use Monte Carlo, because > otherwise it uses some distribution. I believe the term "exact" means
Re: [R] about a p-value < 2.2e-16
I have to ask since. Are you sure the journal simply means by exact p-value that they don’t want to see a p-value given as < 0.0001, for example, and simply want the actual number? I cannot imagine they really meant exact as in the p-value from some exact distribution. -- Kevin E. Thorpe Head of Biostatistics, Applied Health Research Centre (AHRC) Li Ka Shing Knowledge Institute of St. Michael's Assistant Professor, Dalla Lana School of Public Health University of Toronto email: kevin.tho...@utoronto.ca Tel: 416.864.5776 Fax: 416.864.3016 > On Mar 19, 2021, at 1:22 PM, Bogdan Tanasa wrote: > > EXTERNAL EMAIL: > > Dear all, thank you all for comments and help. > > as far as i can see, shall we have samples of 1000 records, only > "exact=FALSE" allows the code to run: > > wilcox.test(rnorm(1000), rnorm(1000, 2), exact=FALSE)$p.value > [1] 7.304863e-231 > > shall i use "exact=TRUE", it runs out of memory on my 64GB RAM PC : > > wilcox.test(rnorm(1000), rnorm(1000, 2), exact=TRUE)$p.value > (the job is terminated by OS) > > shall you have any other suggestions, please let me know. thanks a lot ! > > On Fri, Mar 19, 2021 at 9:05 AM Bert Gunter wrote: > >> I **believe** -- if my old memory still serves-- that the "exact" >> specification uses a home grown version of the algorithm to calculate >> exact, or close approximations to the exact, permutation distribution >> originally developed by Cyrus Mehta, founder of StatXact software. Of >> course, examining the C code source would determine this, but I don't care >> to attempt this. >> >> If this is (no longer?) correct, please point this out. >> >> Best, >> >> Bert Gunter >> >> "The trouble with having an open mind is that people keep coming along and >> sticking things into it." >> -- Opus (aka Berkeley Breathed in his "Bloom County" comic strip ) >> >> >> On Fri, Mar 19, 2021 at 8:42 AM Jiefei Wang wrote: >> >>> Hi Spencer, >>> >>> Thanks for your test results, I do not know the answer as I haven't >>> used wilcox.test for many years. I do not know if it is possible to >>> compute >>> the exact distribution of the Wilcoxon rank sum statistic, but I think it >>> is very likely, as the document of `Wilcoxon` says: >>> >>> This distribution is obtained as follows. Let x and y be two random, >>> independent samples of size m and n. Then the Wilcoxon rank sum statistic >>> is the number of all pairs (x[i], y[j]) for which y[j] is not greater than >>> x[i]. This statistic takes values between 0 and m * n, and its mean and >>> variance are m * n / 2 and m * n * (m + n + 1) / 12, respectively. >>> >>> As a nice feature of the non-parametric statistic, it is usually >>> distribution-free so you can pick any distribution you like to compute the >>> same statistic. I wonder if this is the case, but I might be wrong. >>> >>> Cheers, >>> Jiefei >>> >>> >>> On Fri, Mar 19, 2021 at 10:57 PM Spencer Graves < >>> spencer.gra...@effectivedefense.org> wrote: >>> On 2021-3-19 9:52 AM, Jiefei Wang wrote: > After digging into the R source, it turns out that the argument >>> `exact` has > nothing to do with the numeric precision. It only affects the >>> statistic > model used to compute the p-value. When `exact=TRUE` the true distribution > of the statistic will be used. Otherwise, a normal approximation will >>> be > used. > > I think the documentation needs to be improved here, you can compute >>> the > exact p-value *only* when you do not have any ties in your data. If >>> you > have ties in your data you will get the p-value from the normal > approximation no matter what value you put in `exact`. This behavior should > be documented or a warning should be given when `exact=TRUE` and ties > present. > > FYI, if the exact p-value is required, `pwilcox` function will be >>> used to > compute the p-value. There are no details on how it computes the >>> pvalue but > its C code seems to compute the probability table, so I assume it computes > the exact p-value from the true distribution of the statistic, not a > permutation or MC p-value. My example shows that it does NOT use Monte Carlo, because otherwise it uses some distribution. I believe the term "exact" means that it uses the permutation distribution, though I could be mistaken. If it's NOT a permutation distribution, I don't know what it is. Spencer > > Best, > Jiefei > > > > On Fri, Mar 19, 2021 at 10:01 PM Jiefei Wang >>> wrote: > >> Hey, >> >> I just want to point out that the word "exact" has two meanings. It >>> can >> mean the numerically accurate p-value as Bogdan asked in his first email, >> or it could mean the p-value calculated from the exact distribution >>> of the >> statistic(In this case, U stat). These two are actually not
Re: [R] about a p-value < 2.2e-16
Dear all, thank you all for comments and help. as far as i can see, shall we have samples of 1000 records, only "exact=FALSE" allows the code to run: wilcox.test(rnorm(1000), rnorm(1000, 2), exact=FALSE)$p.value [1] 7.304863e-231 shall i use "exact=TRUE", it runs out of memory on my 64GB RAM PC : wilcox.test(rnorm(1000), rnorm(1000, 2), exact=TRUE)$p.value (the job is terminated by OS) shall you have any other suggestions, please let me know. thanks a lot ! On Fri, Mar 19, 2021 at 9:05 AM Bert Gunter wrote: > I **believe** -- if my old memory still serves-- that the "exact" > specification uses a home grown version of the algorithm to calculate > exact, or close approximations to the exact, permutation distribution > originally developed by Cyrus Mehta, founder of StatXact software. Of > course, examining the C code source would determine this, but I don't care > to attempt this. > > If this is (no longer?) correct, please point this out. > > Best, > > Bert Gunter > > "The trouble with having an open mind is that people keep coming along and > sticking things into it." > -- Opus (aka Berkeley Breathed in his "Bloom County" comic strip ) > > > On Fri, Mar 19, 2021 at 8:42 AM Jiefei Wang wrote: > >> Hi Spencer, >> >> Thanks for your test results, I do not know the answer as I haven't >> used wilcox.test for many years. I do not know if it is possible to >> compute >> the exact distribution of the Wilcoxon rank sum statistic, but I think it >> is very likely, as the document of `Wilcoxon` says: >> >> This distribution is obtained as follows. Let x and y be two random, >> independent samples of size m and n. Then the Wilcoxon rank sum statistic >> is the number of all pairs (x[i], y[j]) for which y[j] is not greater than >> x[i]. This statistic takes values between 0 and m * n, and its mean and >> variance are m * n / 2 and m * n * (m + n + 1) / 12, respectively. >> >> As a nice feature of the non-parametric statistic, it is usually >> distribution-free so you can pick any distribution you like to compute the >> same statistic. I wonder if this is the case, but I might be wrong. >> >> Cheers, >> Jiefei >> >> >> On Fri, Mar 19, 2021 at 10:57 PM Spencer Graves < >> spencer.gra...@effectivedefense.org> wrote: >> >> > >> > >> > On 2021-3-19 9:52 AM, Jiefei Wang wrote: >> > > After digging into the R source, it turns out that the argument >> `exact` >> > has >> > > nothing to do with the numeric precision. It only affects the >> statistic >> > > model used to compute the p-value. When `exact=TRUE` the true >> > distribution >> > > of the statistic will be used. Otherwise, a normal approximation will >> be >> > > used. >> > > >> > > I think the documentation needs to be improved here, you can compute >> the >> > > exact p-value *only* when you do not have any ties in your data. If >> you >> > > have ties in your data you will get the p-value from the normal >> > > approximation no matter what value you put in `exact`. This behavior >> > should >> > > be documented or a warning should be given when `exact=TRUE` and ties >> > > present. >> > > >> > > FYI, if the exact p-value is required, `pwilcox` function will be >> used to >> > > compute the p-value. There are no details on how it computes the >> pvalue >> > but >> > > its C code seems to compute the probability table, so I assume it >> > computes >> > > the exact p-value from the true distribution of the statistic, not a >> > > permutation or MC p-value. >> > >> > >> >My example shows that it does NOT use Monte Carlo, because >> > otherwise it uses some distribution. I believe the term "exact" means >> > that it uses the permutation distribution, though I could be mistaken. >> > If it's NOT a permutation distribution, I don't know what it is. >> > >> > >> >Spencer >> > > >> > > Best, >> > > Jiefei >> > > >> > > >> > > >> > > On Fri, Mar 19, 2021 at 10:01 PM Jiefei Wang >> wrote: >> > > >> > >> Hey, >> > >> >> > >> I just want to point out that the word "exact" has two meanings. It >> can >> > >> mean the numerically accurate p-value as Bogdan asked in his first >> > email, >> > >> or it could mean the p-value calculated from the exact distribution >> of >> > the >> > >> statistic(In this case, U stat). These two are actually not related, >> > even >> > >> though they all called "exact". >> > >> >> > >> Best, >> > >> Jiefei >> > >> >> > >> On Fri, Mar 19, 2021 at 9:31 PM Spencer Graves < >> > >> spencer.gra...@effectivedefense.org> wrote: >> > >> >> > >>> >> > >>> On 2021-3-19 12:54 AM, Bogdan Tanasa wrote: >> > thanks a lot, Vivek ! in other words, assuming that we work with >> 1000 >> > >>> data >> > points, >> > >> > shall we use EXACT = TRUE, it uses the normal approximation, >> > >> > while if EXACT=FALSE (for these large samples), it does not ? >> > >>> >> > >>> As David Winsemius noted, the documentation is not clear. >> > >>> Consider the following: >> > >>> >> > set.seed(1) > x <-
Re: [R] about a p-value < 2.2e-16
For me, it was always clear based on the documentation that if there are ties, then the normal approximation is used (irrespective of what 'exact' is set to). In fact, if there are ties, the output even tells you that this is happening: wilcox.test(c(1,3,2,2,4), exact=TRUE) [...] Warning message: In wilcox.test.default(c(1, 3, 2, 2, 4), exact = TRUE) : cannot compute exact p-value with ties Best, Wolfgang >-Original Message- >From: Jiefei Wang [mailto:szwj...@gmail.com] >Sent: Friday, 19 March, 2021 16:32 >To: Viechtbauer, Wolfgang (SP) >Cc: r-help >Subject: Re: [R] about a p-value < 2.2e-16 > >Dear Wolfgang, > >Thanks for the documentation, but the document only states the default >behavior, >it does not mention what would happen if we tell it to compute the exact >p-value >but the data has ties. I think this would be misleading as people might think >their result is exact by specifying `exact=TRUE` but the truth is that their >data >contains ties and the result is from the normal approximation. > >Best, >Jiefei > >On Fri, Mar 19, 2021 at 11:18 PM Viechtbauer, Wolfgang (SP) > wrote: >Dear Jiefei, > >This behavior is documented. From help(wilcox.test): > >"By default (if exact is not specified), an exact p-value is computed if the >samples contain less than 50 finite values and there are no ties. Otherwise, a >normal approximation is used." > >Best, >Wolfgang > >>-Original Message- >>From: R-help [mailto:r-help-boun...@r-project.org] On Behalf Of Jiefei Wang >>Sent: Friday, 19 March, 2021 15:52 >>To: Spencer Graves >>Cc: r-help; Bogdan Tanasa >>Subject: Re: [R] about a p-value < 2.2e-16 >> >>After digging into the R source, it turns out that the argument `exact` has >>nothing to do with the numeric precision. It only affects the statistic >>model used to compute the p-value. When `exact=TRUE` the true distribution >>of the statistic will be used. Otherwise, a normal approximation will be >>used. >> >>I think the documentation needs to be improved here, you can compute the >>exact p-value *only* when you do not have any ties in your data. If you >>have ties in your data you will get the p-value from the normal >>approximation no matter what value you put in `exact`. This behavior should >>be documented or a warning should be given when `exact=TRUE` and ties >>present. >> >>FYI, if the exact p-value is required, `pwilcox` function will be used to >>compute the p-value. There are no details on how it computes the pvalue but >>its C code seems to compute the probability table, so I assume it computes >>the exact p-value from the true distribution of the statistic, not a >>permutation or MC p-value. >> >>Best, >>Jiefei __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] about a p-value < 2.2e-16
I **believe** -- if my old memory still serves-- that the "exact" specification uses a home grown version of the algorithm to calculate exact, or close approximations to the exact, permutation distribution originally developed by Cyrus Mehta, founder of StatXact software. Of course, examining the C code source would determine this, but I don't care to attempt this. If this is (no longer?) correct, please point this out. Best, Bert Gunter "The trouble with having an open mind is that people keep coming along and sticking things into it." -- Opus (aka Berkeley Breathed in his "Bloom County" comic strip ) On Fri, Mar 19, 2021 at 8:42 AM Jiefei Wang wrote: > Hi Spencer, > > Thanks for your test results, I do not know the answer as I haven't > used wilcox.test for many years. I do not know if it is possible to compute > the exact distribution of the Wilcoxon rank sum statistic, but I think it > is very likely, as the document of `Wilcoxon` says: > > This distribution is obtained as follows. Let x and y be two random, > independent samples of size m and n. Then the Wilcoxon rank sum statistic > is the number of all pairs (x[i], y[j]) for which y[j] is not greater than > x[i]. This statistic takes values between 0 and m * n, and its mean and > variance are m * n / 2 and m * n * (m + n + 1) / 12, respectively. > > As a nice feature of the non-parametric statistic, it is usually > distribution-free so you can pick any distribution you like to compute the > same statistic. I wonder if this is the case, but I might be wrong. > > Cheers, > Jiefei > > > On Fri, Mar 19, 2021 at 10:57 PM Spencer Graves < > spencer.gra...@effectivedefense.org> wrote: > > > > > > > On 2021-3-19 9:52 AM, Jiefei Wang wrote: > > > After digging into the R source, it turns out that the argument `exact` > > has > > > nothing to do with the numeric precision. It only affects the statistic > > > model used to compute the p-value. When `exact=TRUE` the true > > distribution > > > of the statistic will be used. Otherwise, a normal approximation will > be > > > used. > > > > > > I think the documentation needs to be improved here, you can compute > the > > > exact p-value *only* when you do not have any ties in your data. If you > > > have ties in your data you will get the p-value from the normal > > > approximation no matter what value you put in `exact`. This behavior > > should > > > be documented or a warning should be given when `exact=TRUE` and ties > > > present. > > > > > > FYI, if the exact p-value is required, `pwilcox` function will be used > to > > > compute the p-value. There are no details on how it computes the pvalue > > but > > > its C code seems to compute the probability table, so I assume it > > computes > > > the exact p-value from the true distribution of the statistic, not a > > > permutation or MC p-value. > > > > > >My example shows that it does NOT use Monte Carlo, because > > otherwise it uses some distribution. I believe the term "exact" means > > that it uses the permutation distribution, though I could be mistaken. > > If it's NOT a permutation distribution, I don't know what it is. > > > > > >Spencer > > > > > > Best, > > > Jiefei > > > > > > > > > > > > On Fri, Mar 19, 2021 at 10:01 PM Jiefei Wang > wrote: > > > > > >> Hey, > > >> > > >> I just want to point out that the word "exact" has two meanings. It > can > > >> mean the numerically accurate p-value as Bogdan asked in his first > > email, > > >> or it could mean the p-value calculated from the exact distribution of > > the > > >> statistic(In this case, U stat). These two are actually not related, > > even > > >> though they all called "exact". > > >> > > >> Best, > > >> Jiefei > > >> > > >> On Fri, Mar 19, 2021 at 9:31 PM Spencer Graves < > > >> spencer.gra...@effectivedefense.org> wrote: > > >> > > >>> > > >>> On 2021-3-19 12:54 AM, Bogdan Tanasa wrote: > > thanks a lot, Vivek ! in other words, assuming that we work with > 1000 > > >>> data > > points, > > > > shall we use EXACT = TRUE, it uses the normal approximation, > > > > while if EXACT=FALSE (for these large samples), it does not ? > > >>> > > >>> As David Winsemius noted, the documentation is not clear. > > >>> Consider the following: > > >>> > > set.seed(1) > x <- rnorm(100) > y <- rnorm(100, 2) > > > wilcox.test(x, > > >>> y)$p.value > > >>> [1] 1.172189e-25 > wilcox.test(x, y)$p.value [1] 1.172189e-25 > > > > >>> wilcox.test(x, y, EXACT=TRUE)$p.value [1] 1.172189e-25 > > wilcox.test(x, > > >>> y, EXACT=TRUE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, > > >>> exact=TRUE)$p.value [1] 4.123875e-32 > wilcox.test(x, y, > > >>> exact=TRUE)$p.value [1] 4.123875e-32 > > wilcox.test(x, y, > > >>> EXACT=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, > > >>> EXACT=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, > > >>> exact=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, > > >>> exact=FALSE)$p.value [1] 1.172189e-25 > We
Re: [R] about a p-value < 2.2e-16
Hi Spencer, Thanks for your test results, I do not know the answer as I haven't used wilcox.test for many years. I do not know if it is possible to compute the exact distribution of the Wilcoxon rank sum statistic, but I think it is very likely, as the document of `Wilcoxon` says: This distribution is obtained as follows. Let x and y be two random, independent samples of size m and n. Then the Wilcoxon rank sum statistic is the number of all pairs (x[i], y[j]) for which y[j] is not greater than x[i]. This statistic takes values between 0 and m * n, and its mean and variance are m * n / 2 and m * n * (m + n + 1) / 12, respectively. As a nice feature of the non-parametric statistic, it is usually distribution-free so you can pick any distribution you like to compute the same statistic. I wonder if this is the case, but I might be wrong. Cheers, Jiefei On Fri, Mar 19, 2021 at 10:57 PM Spencer Graves < spencer.gra...@effectivedefense.org> wrote: > > > On 2021-3-19 9:52 AM, Jiefei Wang wrote: > > After digging into the R source, it turns out that the argument `exact` > has > > nothing to do with the numeric precision. It only affects the statistic > > model used to compute the p-value. When `exact=TRUE` the true > distribution > > of the statistic will be used. Otherwise, a normal approximation will be > > used. > > > > I think the documentation needs to be improved here, you can compute the > > exact p-value *only* when you do not have any ties in your data. If you > > have ties in your data you will get the p-value from the normal > > approximation no matter what value you put in `exact`. This behavior > should > > be documented or a warning should be given when `exact=TRUE` and ties > > present. > > > > FYI, if the exact p-value is required, `pwilcox` function will be used to > > compute the p-value. There are no details on how it computes the pvalue > but > > its C code seems to compute the probability table, so I assume it > computes > > the exact p-value from the true distribution of the statistic, not a > > permutation or MC p-value. > > >My example shows that it does NOT use Monte Carlo, because > otherwise it uses some distribution. I believe the term "exact" means > that it uses the permutation distribution, though I could be mistaken. > If it's NOT a permutation distribution, I don't know what it is. > > >Spencer > > > > Best, > > Jiefei > > > > > > > > On Fri, Mar 19, 2021 at 10:01 PM Jiefei Wang wrote: > > > >> Hey, > >> > >> I just want to point out that the word "exact" has two meanings. It can > >> mean the numerically accurate p-value as Bogdan asked in his first > email, > >> or it could mean the p-value calculated from the exact distribution of > the > >> statistic(In this case, U stat). These two are actually not related, > even > >> though they all called "exact". > >> > >> Best, > >> Jiefei > >> > >> On Fri, Mar 19, 2021 at 9:31 PM Spencer Graves < > >> spencer.gra...@effectivedefense.org> wrote: > >> > >>> > >>> On 2021-3-19 12:54 AM, Bogdan Tanasa wrote: > thanks a lot, Vivek ! in other words, assuming that we work with 1000 > >>> data > points, > > shall we use EXACT = TRUE, it uses the normal approximation, > > while if EXACT=FALSE (for these large samples), it does not ? > >>> > >>> As David Winsemius noted, the documentation is not clear. > >>> Consider the following: > >>> > set.seed(1) > x <- rnorm(100) > y <- rnorm(100, 2) > > wilcox.test(x, > >>> y)$p.value > >>> [1] 1.172189e-25 > wilcox.test(x, y)$p.value [1] 1.172189e-25 > > > >>> wilcox.test(x, y, EXACT=TRUE)$p.value [1] 1.172189e-25 > wilcox.test(x, > >>> y, EXACT=TRUE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, > >>> exact=TRUE)$p.value [1] 4.123875e-32 > wilcox.test(x, y, > >>> exact=TRUE)$p.value [1] 4.123875e-32 > > wilcox.test(x, y, > >>> EXACT=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, > >>> EXACT=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, > >>> exact=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, > >>> exact=FALSE)$p.value [1] 1.172189e-25 > We get two values here: > >>> 1.172189e-25 and 4.123875e-32. The first one, I think, is the normal > >>> approximation, which is the same as exact=FALSE. I think that with > >>> exact=FALSE, you get a permutation distribution, though I'm not sure. > >>> You might try looking at "wilcox_test in package coin for exact, > >>> asymptotic and Monte Carlo conditional p-values, including in the > >>> presence of ties" to see if it is clearer. NOTE: R is case sensitive, > so > >>> "EXACT" is a different variable from "exact". It is interpreted as an > >>> optional argument, which is not recognized and therefore ignored in > this > >>> context. > >>>Hope this helps. > >>>Spencer > >>> > >>> > On Thu, Mar 18, 2021 at 10:47 PM Vivek Das > wrote: > > > Hi Bogdan, > > > > You can also get the information from the link of the Wilcox.test > >>> function >
Re: [R] about a p-value < 2.2e-16
Dear Jiefei, and all, many thanks for your time and comments, suggestions, insights. -- bogdan On Fri, Mar 19, 2021 at 7:52 AM Jiefei Wang wrote: > After digging into the R source, it turns out that the argument `exact` > has nothing to do with the numeric precision. It only affects the statistic > model used to compute the p-value. When `exact=TRUE` the true distribution > of the statistic will be used. Otherwise, a normal approximation will be > used. > > I think the documentation needs to be improved here, you can compute the > exact p-value *only* when you do not have any ties in your data. If you > have ties in your data you will get the p-value from the normal > approximation no matter what value you put in `exact`. This behavior should > be documented or a warning should be given when `exact=TRUE` and ties > present. > > FYI, if the exact p-value is required, `pwilcox` function will be used to > compute the p-value. There are no details on how it computes the pvalue but > its C code seems to compute the probability table, so I assume it computes > the exact p-value from the true distribution of the statistic, not a > permutation or MC p-value. > > Best, > Jiefei > > > > On Fri, Mar 19, 2021 at 10:01 PM Jiefei Wang wrote: > >> Hey, >> >> I just want to point out that the word "exact" has two meanings. It can >> mean the numerically accurate p-value as Bogdan asked in his first email, >> or it could mean the p-value calculated from the exact distribution of the >> statistic(In this case, U stat). These two are actually not related, even >> though they all called "exact". >> >> Best, >> Jiefei >> >> On Fri, Mar 19, 2021 at 9:31 PM Spencer Graves < >> spencer.gra...@effectivedefense.org> wrote: >> >>> >>> >>> On 2021-3-19 12:54 AM, Bogdan Tanasa wrote: >>> > thanks a lot, Vivek ! in other words, assuming that we work with 1000 >>> data >>> > points, >>> > >>> > shall we use EXACT = TRUE, it uses the normal approximation, >>> > >>> > while if EXACT=FALSE (for these large samples), it does not ? >>> >>> >>>As David Winsemius noted, the documentation is not clear. >>> Consider the following: >>> >>> > set.seed(1) > x <- rnorm(100) > y <- rnorm(100, 2) > > wilcox.test(x, >>> y)$p.value >>> [1] 1.172189e-25 > wilcox.test(x, y)$p.value [1] 1.172189e-25 > > >>> wilcox.test(x, y, EXACT=TRUE)$p.value [1] 1.172189e-25 > wilcox.test(x, >>> y, EXACT=TRUE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, >>> exact=TRUE)$p.value [1] 4.123875e-32 > wilcox.test(x, y, >>> exact=TRUE)$p.value [1] 4.123875e-32 > > wilcox.test(x, y, >>> EXACT=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, >>> EXACT=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, >>> exact=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, >>> exact=FALSE)$p.value [1] 1.172189e-25 > We get two values here: >>> 1.172189e-25 and 4.123875e-32. The first one, I think, is the normal >>> approximation, which is the same as exact=FALSE. I think that with >>> exact=FALSE, you get a permutation distribution, though I'm not sure. >>> You might try looking at "wilcox_test in package coin for exact, >>> asymptotic and Monte Carlo conditional p-values, including in the >>> presence of ties" to see if it is clearer. NOTE: R is case sensitive, so >>> "EXACT" is a different variable from "exact". It is interpreted as an >>> optional argument, which is not recognized and therefore ignored in this >>> context. >>> Hope this helps. >>> Spencer >>> >>> >>> > On Thu, Mar 18, 2021 at 10:47 PM Vivek Das wrote: >>> > >>> >> Hi Bogdan, >>> >> >>> >> You can also get the information from the link of the Wilcox.test >>> function >>> >> page. >>> >> >>> >> “By default (if exact is not specified), an exact p-value is computed >>> if >>> >> the samples contain less than 50 finite values and there are no ties. >>> >> Otherwise, a normal approximation is used.” >>> >> >>> >> For more: >>> >> >>> >> >>> https://stat.ethz.ch/R-manual/R-devel/library/stats/html/wilcox.test.html >>> >> >>> >> Hope this helps! >>> >> >>> >> Best, >>> >> >>> >> VD >>> >> >>> >> >>> >> On Thu, Mar 18, 2021 at 10:36 PM Bogdan Tanasa >>> wrote: >>> >> >>> >>> Dear Peter, thanks a lot. yes, we can see a very precise p-value, >>> and that >>> >>> was the request from the journal. >>> >>> >>> >>> if I may ask another question please : what is the meaning of >>> "exact=TRUE" >>> >>> or "exact=FALSE" in wilcox.test ? >>> >>> >>> >>> i can see that the "numerically precise" p-values are different. >>> thanks a >>> >>> lot ! >>> >>> >>> >>> tst = wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) >>> >>> tst$p.value >>> >>> [1] 8.535524e-25 >>> >>> >>> >>> tst = wilcox.test(rnorm(100), rnorm(100, 2), exact=FALSE) >>> >>> tst$p.value >>> >>> [1] 3.448211e-25 >>> >>> >>> >>> On Thu, Mar 18, 2021 at 10:15 PM Peter Langfelder < >>> >>> peter.langfel...@gmail.com> wrote: >>> >>> >>> I thinnk the answer is much simpler. The print method for hypothesis >>> tests (class
Re: [R] about a p-value < 2.2e-16
Dear Wolfgang, Thanks for the documentation, but the document only states the default behavior, it does not mention what would happen if we tell it to compute the exact p-value but the data has ties. I think this would be misleading as people might think their result is exact by specifying `exact=TRUE` but the truth is that their data contains ties and the result is from the normal approximation. Best, Jiefei On Fri, Mar 19, 2021 at 11:18 PM Viechtbauer, Wolfgang (SP) < wolfgang.viechtba...@maastrichtuniversity.nl> wrote: > Dear Jiefei, > > This behavior is documented. From help(wilcox.test): > > "By default (if exact is not specified), an exact p-value is computed if > the samples contain less than 50 finite values and there are no ties. > Otherwise, a normal approximation is used." > > Best, > Wolfgang > > >-Original Message- > >From: R-help [mailto:r-help-boun...@r-project.org] On Behalf Of Jiefei > Wang > >Sent: Friday, 19 March, 2021 15:52 > >To: Spencer Graves > >Cc: r-help; Bogdan Tanasa > >Subject: Re: [R] about a p-value < 2.2e-16 > > > >After digging into the R source, it turns out that the argument `exact` > has > >nothing to do with the numeric precision. It only affects the statistic > >model used to compute the p-value. When `exact=TRUE` the true distribution > >of the statistic will be used. Otherwise, a normal approximation will be > >used. > > > >I think the documentation needs to be improved here, you can compute the > >exact p-value *only* when you do not have any ties in your data. If you > >have ties in your data you will get the p-value from the normal > >approximation no matter what value you put in `exact`. This behavior > should > >be documented or a warning should be given when `exact=TRUE` and ties > >present. > > > >FYI, if the exact p-value is required, `pwilcox` function will be used to > >compute the p-value. There are no details on how it computes the pvalue > but > >its C code seems to compute the probability table, so I assume it computes > >the exact p-value from the true distribution of the statistic, not a > >permutation or MC p-value. > > > >Best, > >Jiefei > [[alternative HTML version deleted]] __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] about a p-value < 2.2e-16
Dear Jiefei, This behavior is documented. From help(wilcox.test): "By default (if exact is not specified), an exact p-value is computed if the samples contain less than 50 finite values and there are no ties. Otherwise, a normal approximation is used." Best, Wolfgang >-Original Message- >From: R-help [mailto:r-help-boun...@r-project.org] On Behalf Of Jiefei Wang >Sent: Friday, 19 March, 2021 15:52 >To: Spencer Graves >Cc: r-help; Bogdan Tanasa >Subject: Re: [R] about a p-value < 2.2e-16 > >After digging into the R source, it turns out that the argument `exact` has >nothing to do with the numeric precision. It only affects the statistic >model used to compute the p-value. When `exact=TRUE` the true distribution >of the statistic will be used. Otherwise, a normal approximation will be >used. > >I think the documentation needs to be improved here, you can compute the >exact p-value *only* when you do not have any ties in your data. If you >have ties in your data you will get the p-value from the normal >approximation no matter what value you put in `exact`. This behavior should >be documented or a warning should be given when `exact=TRUE` and ties >present. > >FYI, if the exact p-value is required, `pwilcox` function will be used to >compute the p-value. There are no details on how it computes the pvalue but >its C code seems to compute the probability table, so I assume it computes >the exact p-value from the true distribution of the statistic, not a >permutation or MC p-value. > >Best, >Jiefei __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] about a p-value < 2.2e-16
On 2021-3-19 9:52 AM, Jiefei Wang wrote: After digging into the R source, it turns out that the argument `exact` has nothing to do with the numeric precision. It only affects the statistic model used to compute the p-value. When `exact=TRUE` the true distribution of the statistic will be used. Otherwise, a normal approximation will be used. I think the documentation needs to be improved here, you can compute the exact p-value *only* when you do not have any ties in your data. If you have ties in your data you will get the p-value from the normal approximation no matter what value you put in `exact`. This behavior should be documented or a warning should be given when `exact=TRUE` and ties present. FYI, if the exact p-value is required, `pwilcox` function will be used to compute the p-value. There are no details on how it computes the pvalue but its C code seems to compute the probability table, so I assume it computes the exact p-value from the true distribution of the statistic, not a permutation or MC p-value. My example shows that it does NOT use Monte Carlo, because otherwise it uses some distribution. I believe the term "exact" means that it uses the permutation distribution, though I could be mistaken. If it's NOT a permutation distribution, I don't know what it is. Spencer Best, Jiefei On Fri, Mar 19, 2021 at 10:01 PM Jiefei Wang wrote: Hey, I just want to point out that the word "exact" has two meanings. It can mean the numerically accurate p-value as Bogdan asked in his first email, or it could mean the p-value calculated from the exact distribution of the statistic(In this case, U stat). These two are actually not related, even though they all called "exact". Best, Jiefei On Fri, Mar 19, 2021 at 9:31 PM Spencer Graves < spencer.gra...@effectivedefense.org> wrote: On 2021-3-19 12:54 AM, Bogdan Tanasa wrote: thanks a lot, Vivek ! in other words, assuming that we work with 1000 data points, shall we use EXACT = TRUE, it uses the normal approximation, while if EXACT=FALSE (for these large samples), it does not ? As David Winsemius noted, the documentation is not clear. Consider the following: set.seed(1) > x <- rnorm(100) > y <- rnorm(100, 2) > > wilcox.test(x, y)$p.value [1] 1.172189e-25 > wilcox.test(x, y)$p.value [1] 1.172189e-25 > > wilcox.test(x, y, EXACT=TRUE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, EXACT=TRUE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, exact=TRUE)$p.value [1] 4.123875e-32 > wilcox.test(x, y, exact=TRUE)$p.value [1] 4.123875e-32 > > wilcox.test(x, y, EXACT=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, EXACT=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, exact=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, exact=FALSE)$p.value [1] 1.172189e-25 > We get two values here: 1.172189e-25 and 4.123875e-32. The first one, I think, is the normal approximation, which is the same as exact=FALSE. I think that with exact=FALSE, you get a permutation distribution, though I'm not sure. You might try looking at "wilcox_test in package coin for exact, asymptotic and Monte Carlo conditional p-values, including in the presence of ties" to see if it is clearer. NOTE: R is case sensitive, so "EXACT" is a different variable from "exact". It is interpreted as an optional argument, which is not recognized and therefore ignored in this context. Hope this helps. Spencer On Thu, Mar 18, 2021 at 10:47 PM Vivek Das wrote: Hi Bogdan, You can also get the information from the link of the Wilcox.test function page. “By default (if exact is not specified), an exact p-value is computed if the samples contain less than 50 finite values and there are no ties. Otherwise, a normal approximation is used.” For more: https://stat.ethz.ch/R-manual/R-devel/library/stats/html/wilcox.test.html Hope this helps! Best, VD On Thu, Mar 18, 2021 at 10:36 PM Bogdan Tanasa wrote: Dear Peter, thanks a lot. yes, we can see a very precise p-value, and that was the request from the journal. if I may ask another question please : what is the meaning of "exact=TRUE" or "exact=FALSE" in wilcox.test ? i can see that the "numerically precise" p-values are different. thanks a lot ! tst = wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) tst$p.value [1] 8.535524e-25 tst = wilcox.test(rnorm(100), rnorm(100, 2), exact=FALSE) tst$p.value [1] 3.448211e-25 On Thu, Mar 18, 2021 at 10:15 PM Peter Langfelder < peter.langfel...@gmail.com> wrote: I thinnk the answer is much simpler. The print method for hypothesis tests (class htest) truncates the p-values. In the above example, instead of using wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) and copying the output, just print the p-value: tst = wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) tst$p.value [1] 2.988368e-32 I think this value is what the journal asks for. HTH, Peter On Thu, Mar 18, 2021 at 10:05 PM Spencer Graves wrote:
Re: [R] about a p-value < 2.2e-16
After digging into the R source, it turns out that the argument `exact` has nothing to do with the numeric precision. It only affects the statistic model used to compute the p-value. When `exact=TRUE` the true distribution of the statistic will be used. Otherwise, a normal approximation will be used. I think the documentation needs to be improved here, you can compute the exact p-value *only* when you do not have any ties in your data. If you have ties in your data you will get the p-value from the normal approximation no matter what value you put in `exact`. This behavior should be documented or a warning should be given when `exact=TRUE` and ties present. FYI, if the exact p-value is required, `pwilcox` function will be used to compute the p-value. There are no details on how it computes the pvalue but its C code seems to compute the probability table, so I assume it computes the exact p-value from the true distribution of the statistic, not a permutation or MC p-value. Best, Jiefei On Fri, Mar 19, 2021 at 10:01 PM Jiefei Wang wrote: > Hey, > > I just want to point out that the word "exact" has two meanings. It can > mean the numerically accurate p-value as Bogdan asked in his first email, > or it could mean the p-value calculated from the exact distribution of the > statistic(In this case, U stat). These two are actually not related, even > though they all called "exact". > > Best, > Jiefei > > On Fri, Mar 19, 2021 at 9:31 PM Spencer Graves < > spencer.gra...@effectivedefense.org> wrote: > >> >> >> On 2021-3-19 12:54 AM, Bogdan Tanasa wrote: >> > thanks a lot, Vivek ! in other words, assuming that we work with 1000 >> data >> > points, >> > >> > shall we use EXACT = TRUE, it uses the normal approximation, >> > >> > while if EXACT=FALSE (for these large samples), it does not ? >> >> >>As David Winsemius noted, the documentation is not clear. >> Consider the following: >> >> > set.seed(1) > x <- rnorm(100) > y <- rnorm(100, 2) > > wilcox.test(x, >> y)$p.value >> [1] 1.172189e-25 > wilcox.test(x, y)$p.value [1] 1.172189e-25 > > >> wilcox.test(x, y, EXACT=TRUE)$p.value [1] 1.172189e-25 > wilcox.test(x, >> y, EXACT=TRUE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, >> exact=TRUE)$p.value [1] 4.123875e-32 > wilcox.test(x, y, >> exact=TRUE)$p.value [1] 4.123875e-32 > > wilcox.test(x, y, >> EXACT=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, >> EXACT=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, >> exact=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, >> exact=FALSE)$p.value [1] 1.172189e-25 > We get two values here: >> 1.172189e-25 and 4.123875e-32. The first one, I think, is the normal >> approximation, which is the same as exact=FALSE. I think that with >> exact=FALSE, you get a permutation distribution, though I'm not sure. >> You might try looking at "wilcox_test in package coin for exact, >> asymptotic and Monte Carlo conditional p-values, including in the >> presence of ties" to see if it is clearer. NOTE: R is case sensitive, so >> "EXACT" is a different variable from "exact". It is interpreted as an >> optional argument, which is not recognized and therefore ignored in this >> context. >> Hope this helps. >> Spencer >> >> >> > On Thu, Mar 18, 2021 at 10:47 PM Vivek Das wrote: >> > >> >> Hi Bogdan, >> >> >> >> You can also get the information from the link of the Wilcox.test >> function >> >> page. >> >> >> >> “By default (if exact is not specified), an exact p-value is computed >> if >> >> the samples contain less than 50 finite values and there are no ties. >> >> Otherwise, a normal approximation is used.” >> >> >> >> For more: >> >> >> >> >> https://stat.ethz.ch/R-manual/R-devel/library/stats/html/wilcox.test.html >> >> >> >> Hope this helps! >> >> >> >> Best, >> >> >> >> VD >> >> >> >> >> >> On Thu, Mar 18, 2021 at 10:36 PM Bogdan Tanasa >> wrote: >> >> >> >>> Dear Peter, thanks a lot. yes, we can see a very precise p-value, and >> that >> >>> was the request from the journal. >> >>> >> >>> if I may ask another question please : what is the meaning of >> "exact=TRUE" >> >>> or "exact=FALSE" in wilcox.test ? >> >>> >> >>> i can see that the "numerically precise" p-values are different. >> thanks a >> >>> lot ! >> >>> >> >>> tst = wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) >> >>> tst$p.value >> >>> [1] 8.535524e-25 >> >>> >> >>> tst = wilcox.test(rnorm(100), rnorm(100, 2), exact=FALSE) >> >>> tst$p.value >> >>> [1] 3.448211e-25 >> >>> >> >>> On Thu, Mar 18, 2021 at 10:15 PM Peter Langfelder < >> >>> peter.langfel...@gmail.com> wrote: >> >>> >> I thinnk the answer is much simpler. The print method for hypothesis >> tests (class htest) truncates the p-values. In the above example, >> instead of using >> >> wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) >> >> and copying the output, just print the p-value: >> >> tst = wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) >> tst$p.value >>
Re: [R] about a p-value < 2.2e-16
Hey, I just want to point out that the word "exact" has two meanings. It can mean the numerically accurate p-value as Bogdan asked in his first email, or it could mean the p-value calculated from the exact distribution of the statistic(In this case, U stat). These two are actually not related, even though they all called "exact". Best, Jiefei On Fri, Mar 19, 2021 at 9:31 PM Spencer Graves < spencer.gra...@effectivedefense.org> wrote: > > > On 2021-3-19 12:54 AM, Bogdan Tanasa wrote: > > thanks a lot, Vivek ! in other words, assuming that we work with 1000 > data > > points, > > > > shall we use EXACT = TRUE, it uses the normal approximation, > > > > while if EXACT=FALSE (for these large samples), it does not ? > > >As David Winsemius noted, the documentation is not clear. > Consider the following: > > > set.seed(1) > x <- rnorm(100) > y <- rnorm(100, 2) > > wilcox.test(x, > y)$p.value > [1] 1.172189e-25 > wilcox.test(x, y)$p.value [1] 1.172189e-25 > > > wilcox.test(x, y, EXACT=TRUE)$p.value [1] 1.172189e-25 > wilcox.test(x, > y, EXACT=TRUE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, > exact=TRUE)$p.value [1] 4.123875e-32 > wilcox.test(x, y, > exact=TRUE)$p.value [1] 4.123875e-32 > > wilcox.test(x, y, > EXACT=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, > EXACT=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, > exact=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, > exact=FALSE)$p.value [1] 1.172189e-25 > We get two values here: > 1.172189e-25 and 4.123875e-32. The first one, I think, is the normal > approximation, which is the same as exact=FALSE. I think that with > exact=FALSE, you get a permutation distribution, though I'm not sure. > You might try looking at "wilcox_test in package coin for exact, > asymptotic and Monte Carlo conditional p-values, including in the > presence of ties" to see if it is clearer. NOTE: R is case sensitive, so > "EXACT" is a different variable from "exact". It is interpreted as an > optional argument, which is not recognized and therefore ignored in this > context. > Hope this helps. > Spencer > > > > On Thu, Mar 18, 2021 at 10:47 PM Vivek Das wrote: > > > >> Hi Bogdan, > >> > >> You can also get the information from the link of the Wilcox.test > function > >> page. > >> > >> “By default (if exact is not specified), an exact p-value is computed if > >> the samples contain less than 50 finite values and there are no ties. > >> Otherwise, a normal approximation is used.” > >> > >> For more: > >> > >> > https://stat.ethz.ch/R-manual/R-devel/library/stats/html/wilcox.test.html > >> > >> Hope this helps! > >> > >> Best, > >> > >> VD > >> > >> > >> On Thu, Mar 18, 2021 at 10:36 PM Bogdan Tanasa > wrote: > >> > >>> Dear Peter, thanks a lot. yes, we can see a very precise p-value, and > that > >>> was the request from the journal. > >>> > >>> if I may ask another question please : what is the meaning of > "exact=TRUE" > >>> or "exact=FALSE" in wilcox.test ? > >>> > >>> i can see that the "numerically precise" p-values are different. > thanks a > >>> lot ! > >>> > >>> tst = wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) > >>> tst$p.value > >>> [1] 8.535524e-25 > >>> > >>> tst = wilcox.test(rnorm(100), rnorm(100, 2), exact=FALSE) > >>> tst$p.value > >>> [1] 3.448211e-25 > >>> > >>> On Thu, Mar 18, 2021 at 10:15 PM Peter Langfelder < > >>> peter.langfel...@gmail.com> wrote: > >>> > I thinnk the answer is much simpler. The print method for hypothesis > tests (class htest) truncates the p-values. In the above example, > instead of using > > wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) > > and copying the output, just print the p-value: > > tst = wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) > tst$p.value > > [1] 2.988368e-32 > > > I think this value is what the journal asks for. > > HTH, > > Peter > > On Thu, Mar 18, 2021 at 10:05 PM Spencer Graves > wrote: > > I would push back on that from two perspectives: > > > > > > 1. I would study exactly what the journal said very > > carefully. If they mandated "wilcox.test", that function has an > > argument called "exact". If that's what they are asking, then using > > that argument gives the exact p-value, e.g.: > > > > > > > wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) > > > > Wilcoxon rank sum exact test > > > > data: rnorm(100) and rnorm(100, 2) > > W = 691, p-value < 2.2e-16 > > > > > > 2. If that's NOT what they are asking, then I'm not > > convinced what they are asking makes sense: There is is no such > thing > > as an "exact p value" except to the extent that certain assumptions > > hold, and all models are wrong (but some are useful), as George Box > > famously said years ago.[1] Truth only exists in mathematics,
Re: [R] about a p-value < 2.2e-16
On 2021-3-19 12:54 AM, Bogdan Tanasa wrote: > thanks a lot, Vivek ! in other words, assuming that we work with 1000 data > points, > > shall we use EXACT = TRUE, it uses the normal approximation, > > while if EXACT=FALSE (for these large samples), it does not ? As David Winsemius noted, the documentation is not clear. Consider the following: > set.seed(1) > x <- rnorm(100) > y <- rnorm(100, 2) > > wilcox.test(x, > y)$p.value [1] 1.172189e-25 > wilcox.test(x, y)$p.value [1] 1.172189e-25 > > wilcox.test(x, y, EXACT=TRUE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, EXACT=TRUE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, exact=TRUE)$p.value [1] 4.123875e-32 > wilcox.test(x, y, exact=TRUE)$p.value [1] 4.123875e-32 > > wilcox.test(x, y, EXACT=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, EXACT=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, exact=FALSE)$p.value [1] 1.172189e-25 > wilcox.test(x, y, exact=FALSE)$p.value [1] 1.172189e-25 > We get two values here: 1.172189e-25 and 4.123875e-32. The first one, I think, is the normal approximation, which is the same as exact=FALSE. I think that with exact=FALSE, you get a permutation distribution, though I'm not sure. You might try looking at "wilcox_test in package coin for exact, asymptotic and Monte Carlo conditional p-values, including in the presence of ties" to see if it is clearer. NOTE: R is case sensitive, so "EXACT" is a different variable from "exact". It is interpreted as an optional argument, which is not recognized and therefore ignored in this context. Hope this helps. Spencer > On Thu, Mar 18, 2021 at 10:47 PM Vivek Das wrote: > >> Hi Bogdan, >> >> You can also get the information from the link of the Wilcox.test function >> page. >> >> “By default (if exact is not specified), an exact p-value is computed if >> the samples contain less than 50 finite values and there are no ties. >> Otherwise, a normal approximation is used.” >> >> For more: >> >> https://stat.ethz.ch/R-manual/R-devel/library/stats/html/wilcox.test.html >> >> Hope this helps! >> >> Best, >> >> VD >> >> >> On Thu, Mar 18, 2021 at 10:36 PM Bogdan Tanasa wrote: >> >>> Dear Peter, thanks a lot. yes, we can see a very precise p-value, and that >>> was the request from the journal. >>> >>> if I may ask another question please : what is the meaning of "exact=TRUE" >>> or "exact=FALSE" in wilcox.test ? >>> >>> i can see that the "numerically precise" p-values are different. thanks a >>> lot ! >>> >>> tst = wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) >>> tst$p.value >>> [1] 8.535524e-25 >>> >>> tst = wilcox.test(rnorm(100), rnorm(100, 2), exact=FALSE) >>> tst$p.value >>> [1] 3.448211e-25 >>> >>> On Thu, Mar 18, 2021 at 10:15 PM Peter Langfelder < >>> peter.langfel...@gmail.com> wrote: >>> I thinnk the answer is much simpler. The print method for hypothesis tests (class htest) truncates the p-values. In the above example, instead of using wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) and copying the output, just print the p-value: tst = wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) tst$p.value [1] 2.988368e-32 I think this value is what the journal asks for. HTH, Peter On Thu, Mar 18, 2021 at 10:05 PM Spencer Graves wrote: > I would push back on that from two perspectives: > > > 1. I would study exactly what the journal said very > carefully. If they mandated "wilcox.test", that function has an > argument called "exact". If that's what they are asking, then using > that argument gives the exact p-value, e.g.: > > > > wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) > > Wilcoxon rank sum exact test > > data: rnorm(100) and rnorm(100, 2) > W = 691, p-value < 2.2e-16 > > > 2. If that's NOT what they are asking, then I'm not > convinced what they are asking makes sense: There is is no such thing > as an "exact p value" except to the extent that certain assumptions > hold, and all models are wrong (but some are useful), as George Box > famously said years ago.[1] Truth only exists in mathematics, and > that's because it's a fiction to start with ;-) > > > Hope this helps. > Spencer Graves > > > [1] > https://en.wikipedia.org/wiki/All_models_are_wrong > > > On 2021-3-18 11:12 PM, Bogdan Tanasa wrote: >>< https://meta.stackexchange.com/questions/362285/about-a-p-value-2-2e-16 >> Dear all, >> >> i would appreciate having your advice on the following please : >> >> in R, the wilcox.test() provides "a p-value < 2.2e-16", when we >>> compare >> sets of 1000 genes expression (in the genomics field). >> >> however, the journal asks us to provide the exact p
Re: [R] about a p-value < 2.2e-16
Sent from my iPhone > On Mar 18, 2021, at 10:26 PM, Bogdan Tanasa wrote: > > Dear Spencer, thank you very much for your prompt email and help. When > using : > >> wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) > W = 698, p-value < 2.2e-16 > >> wilcox.test(rnorm(100), rnorm(100, 2), exact=FALSE) > W = 1443, p-value < 2.2e-16 > > and in both cases p-value < 2.2e-16. By "exact" p-value, i have meant the > "precise" p-value ; > > If I may ask please, could we write p-value = 0 ? > > i have noted a similar conversation on stackexchange, although the answer > is not very clear (to me). The reason it wasn’t and couldn’t be “clear” was that the underlying scientific question and the statistical methods were not precisely described. The same lack of background information still persists in this discussion. — David > > https://stats.stackexchange.com/questions/78839/how-should-tiny-p-values-be-reported-and-why-does-r-put-a-minimum-on-2-22e-1 > > thanks again, > > bogdan > >> On Thu, Mar 18, 2021 at 10:05 PM Spencer Graves < >> spencer.gra...@effectivedefense.org> wrote: >> I would push back on that from two perspectives: >>1. I would study exactly what the journal said very >> carefully. If they mandated "wilcox.test", that function has an >> argument called "exact". If that's what they are asking, then using >> that argument gives the exact p-value, e.g.: >>> wilcox.test(rnorm(100), rnorm(100, 2), exact=TRUE) >>Wilcoxon rank sum exact test >> data: rnorm(100) and rnorm(100, 2) >> W = 691, p-value < 2.2e-16 >>2. If that's NOT what they are asking, then I'm not >> convinced what they are asking makes sense: There is is no such thing >> as an "exact p value" except to the extent that certain assumptions >> hold, and all models are wrong (but some are useful), as George Box >> famously said years ago.[1] Truth only exists in mathematics, and >> that's because it's a fiction to start with ;-) >> Hope this helps. >> Spencer Graves >> [1] >> https://en.wikipedia.org/wiki/All_models_are_wrong On 2021-3-18 11:12 PM, Bogdan Tanasa wrote: < >> https://meta.stackexchange.com/questions/362285/about-a-p-value-2-2e-16> >>> Dear all, >>> i would appreciate having your advice on the following please : >>> in R, the wilcox.test() provides "a p-value < 2.2e-16", when we compare >>> sets of 1000 genes expression (in the genomics field). >>> however, the journal asks us to provide the exact p value ... >>> would it be legitimate to write : "p-value = 0" ? thanks a lot, >>> -- bogdan >>> [[alternative HTML version deleted]] >>> __ >>> R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see >>> https://stat.ethz.ch/mailman/listinfo/r-help >>> PLEASE do read the posting guide >> http://www.R-project.org/posting-guide.html >>> and provide commented, minimal, self-contained, reproducible code. > > [[alternative HTML version deleted]] > > __ > R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. __ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
Re: [R] Failure in predicting parameters
Thank you, I'll try it! On Thu, Mar 18, 2021 at 9:46 PM Rui Barradas wrote: > > Hello, > > Maybe a bit late but there is a contributed package [1] for quantitative > PCR fitting non-linear models with the Levenberg-Marquardt algorithm. > > estim and vector R below are your model and your fitted values vector. > The RMSE of this fit is smaller than your model's. > > > Isn't this simpler? > > > library(qpcR) > > df1 <- data.frame(Cycles = seq_along(high), high) > > fit <- pcrfit( >data = df1, >cyc = 1, >fluo = 2 > ) > summary(fit) > > coef(estim) > coef(fit) > > > sqrt(sum(resid(estim)^2)) > #[1] 1724.768 > sqrt(sum(resid(fit)^2)) > #[1] 1178.318 > > > highpred <- predict(fit, newdata = df1) > > plot(1:45, high, type = "l", col = "red") > points(1:45, R, col = "blue") > points(1:45, highpred$Prediction, col = "cyan", pch = 3) > > > [1] https://CRAN.R-project.org/package=qpcR > > Hope this helps, > > Rui Barradas > > Às 06:51 de 18/03/21, Luigi Marongiu escreveu: > > It worked. I re-written the equation as: > > ``` > > rutledge_param <- function(p, x, y) ( (p$M / ( 1 + exp(-(x-p$m)/p$s)) > > ) + p$B ) - y > > ``` > > and used Desmos to estimate the slope, so: > > ``` > > estim <- nls.lm(par = list(m = halfCycle, s = 2.77, M = MaxFluo, B = > > high[1]), > > fn = rutledge_param, x = 1:45, y = high) > > summary(estim) > > R <- rutledge(list(half_fluorescence = 27.1102, slope = 2.7680, > > max_fluorescence = 11839.7745, back_fluorescence = > > -138.8615) , 1:45) > > points(1:45, R, type="l", col="red") > > ``` > > > > Thanks > > > > On Tue, Mar 16, 2021 at 8:29 AM Luigi Marongiu > > wrote: > >> > >> Just an update: > >> I tried with desmos and the fitting looks good. Desmos calculated the > >> parameters as: > >> Fmax = 11839.8 > >> Chalf = 27.1102 (with matches with my estimate of 27 cycles) > >> k = 2.76798 > >> Fb = -138.864 > >> I forced R to accept the right parameters using a single named list > >> and re-written the formula (it was a bit unclear in the paper): > >> ``` > >> rutledge <- function(p, x) { > >>m = p$half_fluorescence > >>s = p$slope > >>M = p$max_fluorescence > >>B = p$back_fluorescence > >>y = (M / (1+exp( -((x-m)/s) )) ) + B > >>return(y) > >> } > >> ``` > >> but when I apply it I get a funny graph: > >> ``` > >> desmos <- rutledge(list(half_fluorescence = 27.1102, slope = 2.76798, > >> max_fluorescence = 11839.8, back_fluorescence > >> = -138.864) , high) > >> ``` > >> > >> On Mon, Mar 15, 2021 at 7:39 AM Luigi Marongiu > >> wrote: > >>> > >>> Hello, > >>> the negative data comes from the machine. Probably I should use raw > >>> data directly, although in the paper this requirement is not reported. > >>> The p$x was a typo. Now I corrected it and I got this error: > >>> ``` > >>> > rutledge_param <- function(p, x, y) ((p$M / (1 + exp(-1*(x-p$m)/p$s))) + > p$B) - y > estim <- nls.lm(par = list(m = halfFluo, s = slopes, M = MaxFluo, B = > high[1]), > >>> + fn = rutledge_param, x = 1:45, y = high) > >>> Error in dimnames(x) <- dn : > >>>length of 'dimnames' [2] not equal to array extent > >>> ``` > >>> Probably because 'slopes' is a vector instead of a scalar. Since the > >>> slope is changing, I don't think is right to use a scalar, but I tried > >>> and I got: > >>> ``` > estim <- nls.lm(par = list(m = halfFluo, s = 1, M = MaxFluo, B = > high[1]), > >>> + fn = rutledge_param, x = 1:45, y = high) > estim > >>> Nonlinear regression via the Levenberg-Marquardt algorithm > >>> parameter estimates: 6010.94, 1, 12021.88, 4700.4928889 > >>> residual sum-of-squares: 1.14e+09 > >>> reason terminated: Relative error in the sum of squares is at most `ftol'. > >>> ``` > >>> The values reported are the same I used at the beginning apart from > >>> the last (the background parameter) which is 4700 instead of zero. If > >>> I plug it, I get an L shaped plot that is worse than that at the > >>> beginning: > >>> ``` > >>> after = init = rutledge(halfFluo, 1, MaxFluo, 4700.4928889, high) > >>> points(1:45, after, type="l", col="blue") > >>> ``` > >>> What did I get wrong here? > >>> Thanks > >>> > >>> On Sun, Mar 14, 2021 at 8:05 PM Bill Dunlap > >>> wrote: > > > rutledge_param <- function(p, x, y) ((p$M / (1 + > > exp(-1*(p$x-p$m)/p$s))) + p$B) - y > > Did you mean that p$x to be just x? As is, this returns numeric(0) > for the p that nls.lm gives it because p$x is NULL and NULL-aNumber is > numeric(). > > -Bill > > On Sun, Mar 14, 2021 at 9:46 AM Luigi Marongiu > wrote: > > > > Hello, > > I would like to use the Rutledge equation > > (https://pubmed.ncbi.nlm.nih.gov/15601990/) to model PCR data. The > > equation is: > > Fc = Fmax / (1+exp(-(C-Chalf)/k)) + Fb > > I defined the equation and another that subtracts the values from the > >