Re: [R] alternatives to KS test applicable to K-samples
Not sure what the issue is here. If I click on the link I provided, I go right to the USGS page where instructions for downloading the software are provided. Brian Brian S. Cade, PhD U. S. Geological Survey Fort Collins Science Center 2150 Centre Ave., Bldg. C Fort Collins, CO 80526-8818 email: ca...@usgs.gov brian_c...@usgs.gov tel: 970 226-9326 On Sat, May 30, 2015 at 3:11 PM, David Winsemius dwinsem...@comcast.net wrote: On May 29, 2015, at 10:02 AM, Cade, Brian wrote: Wensui: There are the multi-response permutation procedures (MRPP) that readily test the omnibus hypothesis of no distributional differences among multiple samples for univariate or multivariate responses. There also are empirical coverage tests that test a similar hypothesis among multiple samples but only for univariate responses. Both are included in the USGS Blossom package for R linked here: https://www.fort.usgs.gov/products/23735 (not yet distributed via CRAN). I did not find a link to an actual package at that page nor on any of the others to which it linked. The MRPP may also be available in other R packages on CRAN (vegan ?). There is an mrpp function in pkg:vegan although its help page description made me think it depended (at least in its default mode) on the squared-deviations from means. I'd suggest using CRAN Task View: Robust Statistical Methods (Maintainer: Martin Maechler) for searching for alternative methods. -- David. Brian Brian S. Cade, PhD U. S. Geological Survey Fort Collins Science Center 2150 Centre Ave., Bldg. C Fort Collins, CO 80526-8818 email: ca...@usgs.gov brian_c...@usgs.gov tel: 970 226-9326 On Fri, May 29, 2015 at 10:31 AM, Wensui Liu liuwen...@gmail.com wrote: Good morning, All I have a stat question not specifically related to the the programming language. To compare distributional consistency / discrepancy between two samples, we usually use kolmogorov-smirnov test, which is implemented in R with ks.test() or in SAS with pro npar1way edf. I am wondering if there is any alternative to KS test that could be generalized to K-samples. Thanks and have a nice weekend. wensui David Winsemius Alameda, CA, USA [[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] alternatives to KS test applicable to K-samples
On May 30, 2015, at 7:09 AM, Wensui Liu wrote: Thanks for your insight, David But I am not interested in comparing means among multiple groups. Instead, I want to compare empirical distributions. In this case, I am not sure if wilcoxon should be still applicable. still appreciate it. The Wilcoxon Rank Sum is not comparing means (or medians as I mistakenly thought in the past) but is a more general test of location. You are correct in thinking that the KS test is implicitly testing a wider range of hypotheses, although it remains fairly weak against specific tests. I wasn't suggesting the coin package simply because of its capacity to generalize the WRS test but because of its capacity to support permutation tests of many sorts. If you are testing at all, then there would seem to be a likelihood (in the vague sense of consideration of possible goals of your testing proces) that you really would be interested in departures from equality of distribution that might have a more specific description, and might therefore be interested in testing strategies with more power, perhaps a compound test for differences in location and spread. -- David. On Fri, May 29, 2015 at 1:32 PM, David Winsemius dwinsem...@comcast.net wrote: On May 29, 2015, at 9:31 AM, Wensui Liu wrote: Good morning, All I have a stat question not specifically related to the the programming language. To compare distributional consistency / discrepancy between two samples, we usually use kolmogorov-smirnov test, which is implemented in R with ks.test() or in SAS with pro npar1way edf. I am wondering if there is any alternative to KS test that could be generalized to K-samples. The 'coin' package (Hothorn, Hornick, van de Weil, and Zeileis) presents a variety of permutation and rank-based tests that would probably be more powerful than any multi-group variant of the KS test. The multi-group variant of the Wilcoxon Rank Sum Test presented in the examples for the help page: ?wilcox_test is the Nemenyi-Damico-Wolfe-Dunn test. -- David Winsemius Alameda, CA, USA -- == WenSui Liu Credit Risk Manager, 53 Bancorp wensui@53.com 513-295-4370 == David Winsemius Alameda, CA, USA __ 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] alternatives to KS test applicable to K-samples
... or in not testing at all. The distributions are not the same, period. So what. Testing for equality is useless -- the real question is: what issues are you trying to address/ what questions are you trying to answer/ can they be answered with the data you have or plan to get? In any case, this does not seem the proper venue for such matters, as it has nothing to do with R -- for now, anyways (mea culpa). I would suggest you post to a statistics list like stats.stackexchange.com to figure out what you want to do and then maybe come back here as necessary (after searching) to get any help you might need with R tools to do it. Cheers, Bert Bert Gunter Data is not information. Information is not knowledge. And knowledge is certainly not wisdom. -- Clifford Stoll On Sat, May 30, 2015 at 10:42 AM, David Winsemius dwinsem...@comcast.net wrote: On May 30, 2015, at 7:09 AM, Wensui Liu wrote: Thanks for your insight, David But I am not interested in comparing means among multiple groups. Instead, I want to compare empirical distributions. In this case, I am not sure if wilcoxon should be still applicable. still appreciate it. The Wilcoxon Rank Sum is not comparing means (or medians as I mistakenly thought in the past) but is a more general test of location. You are correct in thinking that the KS test is implicitly testing a wider range of hypotheses, although it remains fairly weak against specific tests. I wasn't suggesting the coin package simply because of its capacity to generalize the WRS test but because of its capacity to support permutation tests of many sorts. If you are testing at all, then there would seem to be a likelihood (in the vague sense of consideration of possible goals of your testing proces) that you really would be interested in departures from equality of distribution that might have a more specific description, and might therefore be interested in testing strategies with more power, perhaps a compound test for differences in location and spread. -- David. On Fri, May 29, 2015 at 1:32 PM, David Winsemius dwinsem...@comcast.net wrote: On May 29, 2015, at 9:31 AM, Wensui Liu wrote: Good morning, All I have a stat question not specifically related to the the programming language. To compare distributional consistency / discrepancy between two samples, we usually use kolmogorov-smirnov test, which is implemented in R with ks.test() or in SAS with pro npar1way edf. I am wondering if there is any alternative to KS test that could be generalized to K-samples. The 'coin' package (Hothorn, Hornick, van de Weil, and Zeileis) presents a variety of permutation and rank-based tests that would probably be more powerful than any multi-group variant of the KS test. The multi-group variant of the Wilcoxon Rank Sum Test presented in the examples for the help page: ?wilcox_test is the Nemenyi-Damico-Wolfe-Dunn test. -- David Winsemius Alameda, CA, USA -- == WenSui Liu Credit Risk Manager, 53 Bancorp wensui@53.com 513-295-4370 == David Winsemius Alameda, CA, USA __ 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.
Re: [R] alternatives to KS test applicable to K-samples
On May 29, 2015, at 10:02 AM, Cade, Brian wrote: Wensui: There are the multi-response permutation procedures (MRPP) that readily test the omnibus hypothesis of no distributional differences among multiple samples for univariate or multivariate responses. There also are empirical coverage tests that test a similar hypothesis among multiple samples but only for univariate responses. Both are included in the USGS Blossom package for R linked here: https://www.fort.usgs.gov/products/23735 (not yet distributed via CRAN). I did not find a link to an actual package at that page nor on any of the others to which it linked. The MRPP may also be available in other R packages on CRAN (vegan ?). There is an mrpp function in pkg:vegan although its help page description made me think it depended (at least in its default mode) on the squared-deviations from means. I'd suggest using CRAN Task View: Robust Statistical Methods (Maintainer: Martin Maechler) for searching for alternative methods. -- David. Brian Brian S. Cade, PhD U. S. Geological Survey Fort Collins Science Center 2150 Centre Ave., Bldg. C Fort Collins, CO 80526-8818 email: ca...@usgs.gov brian_c...@usgs.gov tel: 970 226-9326 On Fri, May 29, 2015 at 10:31 AM, Wensui Liu liuwen...@gmail.com wrote: Good morning, All I have a stat question not specifically related to the the programming language. To compare distributional consistency / discrepancy between two samples, we usually use kolmogorov-smirnov test, which is implemented in R with ks.test() or in SAS with pro npar1way edf. I am wondering if there is any alternative to KS test that could be generalized to K-samples. Thanks and have a nice weekend. wensui David Winsemius Alameda, CA, USA __ 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] alternatives to KS test applicable to K-samples
thanks for your comment, Bert as pointed out by Brian, mrpp suits my need. On Sat, May 30, 2015 at 2:02 PM, Bert Gunter bgunter.4...@gmail.com wrote: ... or in not testing at all. The distributions are not the same, period. So what. Testing for equality is useless -- the real question is: what issues are you trying to address/ what questions are you trying to answer/ can they be answered with the data you have or plan to get? In any case, this does not seem the proper venue for such matters, as it has nothing to do with R -- for now, anyways (mea culpa). I would suggest you post to a statistics list like stats.stackexchange.com to figure out what you want to do and then maybe come back here as necessary (after searching) to get any help you might need with R tools to do it. Cheers, Bert Bert Gunter Data is not information. Information is not knowledge. And knowledge is certainly not wisdom. -- Clifford Stoll On Sat, May 30, 2015 at 10:42 AM, David Winsemius dwinsem...@comcast.net wrote: On May 30, 2015, at 7:09 AM, Wensui Liu wrote: Thanks for your insight, David But I am not interested in comparing means among multiple groups. Instead, I want to compare empirical distributions. In this case, I am not sure if wilcoxon should be still applicable. still appreciate it. The Wilcoxon Rank Sum is not comparing means (or medians as I mistakenly thought in the past) but is a more general test of location. You are correct in thinking that the KS test is implicitly testing a wider range of hypotheses, although it remains fairly weak against specific tests. I wasn't suggesting the coin package simply because of its capacity to generalize the WRS test but because of its capacity to support permutation tests of many sorts. If you are testing at all, then there would seem to be a likelihood (in the vague sense of consideration of possible goals of your testing proces) that you really would be interested in departures from equality of distribution that might have a more specific description, and might therefore be interested in testing strategies with more power, perhaps a compound test for differences in location and spread. -- David. On Fri, May 29, 2015 at 1:32 PM, David Winsemius dwinsem...@comcast.net wrote: On May 29, 2015, at 9:31 AM, Wensui Liu wrote: Good morning, All I have a stat question not specifically related to the the programming language. To compare distributional consistency / discrepancy between two samples, we usually use kolmogorov-smirnov test, which is implemented in R with ks.test() or in SAS with pro npar1way edf. I am wondering if there is any alternative to KS test that could be generalized to K-samples. The 'coin' package (Hothorn, Hornick, van de Weil, and Zeileis) presents a variety of permutation and rank-based tests that would probably be more powerful than any multi-group variant of the KS test. The multi-group variant of the Wilcoxon Rank Sum Test presented in the examples for the help page: ?wilcox_test is the Nemenyi-Damico-Wolfe-Dunn test. -- David Winsemius Alameda, CA, USA -- == WenSui Liu Credit Risk Manager, 53 Bancorp wensui@53.com 513-295-4370 == David Winsemius Alameda, CA, USA __ 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. -- == WenSui Liu Credit Risk Manager, 53 Bancorp wensui@53.com 513-295-4370 __ 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] alternatives to KS test applicable to K-samples
Thanks for your insight, David But I am not interested in comparing means among multiple groups. Instead, I want to compare empirical distributions. In this case, I am not sure if wilcoxon should be still applicable. still appreciate it. On Fri, May 29, 2015 at 1:32 PM, David Winsemius dwinsem...@comcast.net wrote: On May 29, 2015, at 9:31 AM, Wensui Liu wrote: Good morning, All I have a stat question not specifically related to the the programming language. To compare distributional consistency / discrepancy between two samples, we usually use kolmogorov-smirnov test, which is implemented in R with ks.test() or in SAS with pro npar1way edf. I am wondering if there is any alternative to KS test that could be generalized to K-samples. The 'coin' package (Hothorn, Hornick, van de Weil, and Zeileis) presents a variety of permutation and rank-based tests that would probably be more powerful than any multi-group variant of the KS test. The multi-group variant of the Wilcoxon Rank Sum Test presented in the examples for the help page: ?wilcox_test is the Nemenyi-Damico-Wolfe-Dunn test. -- David Winsemius Alameda, CA, USA -- == WenSui Liu Credit Risk Manager, 53 Bancorp wensui@53.com 513-295-4370 == __ 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] alternatives to KS test applicable to K-samples
Wensui: There are the multi-response permutation procedures (MRPP) that readily test the omnibus hypothesis of no distributional differences among multiple samples for univariate or multivariate responses. There also are empirical coverage tests that test a similar hypothesis among multiple samples but only for univariate responses. Both are included in the USGS Blossom package for R linked here: https://www.fort.usgs.gov/products/23735 (not yet distributed via CRAN). The MRPP may also be available in other R packages on CRAN (vegan ?). Brian Brian S. Cade, PhD U. S. Geological Survey Fort Collins Science Center 2150 Centre Ave., Bldg. C Fort Collins, CO 80526-8818 email: ca...@usgs.gov brian_c...@usgs.gov tel: 970 226-9326 On Fri, May 29, 2015 at 10:31 AM, Wensui Liu liuwen...@gmail.com wrote: Good morning, All I have a stat question not specifically related to the the programming language. To compare distributional consistency / discrepancy between two samples, we usually use kolmogorov-smirnov test, which is implemented in R with ks.test() or in SAS with pro npar1way edf. I am wondering if there is any alternative to KS test that could be generalized to K-samples. Thanks and have a nice weekend. wensui __ 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.
Re: [R] alternatives to KS test applicable to K-samples
Very nice, Brian Sincerely appreciate your assistance! On Friday, May 29, 2015, Cade, Brian ca...@usgs.gov wrote: Wensui: There are the multi-response permutation procedures (MRPP) that readily test the omnibus hypothesis of no distributional differences among multiple samples for univariate or multivariate responses. There also are empirical coverage tests that test a similar hypothesis among multiple samples but only for univariate responses. Both are included in the USGS Blossom package for R linked here: https://www.fort.usgs.gov/products/23735 (not yet distributed via CRAN). The MRPP may also be available in other R packages on CRAN (vegan ?). Brian Brian S. Cade, PhD U. S. Geological Survey Fort Collins Science Center 2150 Centre Ave., Bldg. C Fort Collins, CO 80526-8818 email: ca...@usgs.gov javascript:_e(%7B%7D,'cvml','brian_c...@usgs.gov'); tel: 970 226-9326 On Fri, May 29, 2015 at 10:31 AM, Wensui Liu liuwen...@gmail.com javascript:_e(%7B%7D,'cvml','liuwen...@gmail.com'); wrote: Good morning, All I have a stat question not specifically related to the the programming language. To compare distributional consistency / discrepancy between two samples, we usually use kolmogorov-smirnov test, which is implemented in R with ks.test() or in SAS with pro npar1way edf. I am wondering if there is any alternative to KS test that could be generalized to K-samples. Thanks and have a nice weekend. wensui __ R-help@r-project.org javascript:_e(%7B%7D,'cvml','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. -- == WenSui Liu Credit Risk Manager, 53 Bancorp wensui@53.com 513-295-4370 == [[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] alternatives to KS test applicable to K-samples
On May 29, 2015, at 9:31 AM, Wensui Liu wrote: Good morning, All I have a stat question not specifically related to the the programming language. To compare distributional consistency / discrepancy between two samples, we usually use kolmogorov-smirnov test, which is implemented in R with ks.test() or in SAS with pro npar1way edf. I am wondering if there is any alternative to KS test that could be generalized to K-samples. The 'coin' package (Hothorn, Hornick, van de Weil, and Zeileis) presents a variety of permutation and rank-based tests that would probably be more powerful than any multi-group variant of the KS test. The multi-group variant of the Wilcoxon Rank Sum Test presented in the examples for the help page: ?wilcox_test is the Nemenyi-Damico-Wolfe-Dunn test. -- David Winsemius Alameda, CA, USA __ 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.