similar to: FW: distance between distributions

Hi, This is more of a general stat question. I am looking for a easily computable measure of a distance between two empirical distributions. Say I have two samples x and y drawn from X and Y. I want to compute a statistics rho(x,y) which is zero if X = Y and grows as X and Y become less similar. Kullback-Leibler distance is the most "official" choice, however it needs estimation of

Hello, pardon my ingorance, but what distance metric is used in this function in the nnclust package? The manual only says: "Find the nearest neighbours of points in one data set from another data set. Useful for Mallows-type distance metrics." BR, Jay

Hi I am looking for an implementation (or alternative to) Mallow's distance or the Earth Mover's distance to compare distributions or unnormalized distributions (signatures). Is there an implementation in R or can somebody recommend an alternative? Thanks Rainer -- Rainer M. Krug, PhD (Conservation Ecology, SUN), MSc (Conservation Biology, UCT), Dipl. Phys. (Germany) Centre of

Not precisely an answer to your question but here are some OO R links that I have collected over time. Not sure if all these links still work. <a href="http://www.stat.wisc.edu/~st771-1/slides/wk2-4.pdf">Bates</a> | <a href="http://www.maths.lth.se/help/R/">Bengtsson</a> | <a

Dear R users, It is my pleasure to announce the availability of package stepR (1.0-2) on CRAN. The main purpose of the package is to fit piecewise constant functions (a.k.a. step-functions or block signals) to serial data in a fully data-driven manner under certain (Gaussian or non-Gaussian) distributional assumptions. It mainly implements the algorithms described in the references below -

Dear R users, It is my pleasure to announce the availability of package stepR (1.0-2) on CRAN. The main purpose of the package is to fit piecewise constant functions (a.k.a. step-functions or block signals) to serial data in a fully data-driven manner under certain (Gaussian or non-Gaussian) distributional assumptions. It mainly implements the algorithms described in the references below -

>?step .... 'step' uses 'add1' and 'drop1' repeatedly; it will work for any method for which they work, and that is determined by having a valid method for 'extractAIC'. When the additive constant can be chosen so that AIC is equal to Mallows' Cp, this is done and the tables are labelled appropriately. so my question is :what constant

Original incorrect subject line was Subject: [Rd] multi.line=FALSE does not work in scan() (PR#1210) It *does* work quite correctly: this was an erroneous file nd you got an error, just one you didn't understand. Please don't use grossly misleading subject lines: they are all that appear in the BUGS file. On Tue, 11 Dec 2001 vograno@arbitrade.com wrote: > The following applies to

Dear all, Does R or S-plus or any of their packages provide any command to form any of the following procedures to find Best Regression Equation - 1. 'All Possible Regressions Procedures' (is there any automated command to perform 2^p regressions and ordering according to criteria R2(adj), mallows Cp, s2- by not setting all the regression models manually), 2. 'Backward

Dear List: I have uploaded version 1.0-4 of DPpackage on CRAN. Since the first version (1.0-0), I have not communicated the improvements of the package. I'll use this email to summarize its current status. The name of the package is motivated by the Dirichlet process. However, DPpackage tries to be a general package for Bayesian nonparametric and semi-parametric data analysis. So

Dear List: I have uploaded version 1.0-4 of DPpackage on CRAN. Since the first version (1.0-0), I have not communicated the improvements of the package. I'll use this email to summarize its current status. The name of the package is motivated by the Dirichlet process. However, DPpackage tries to be a general package for Bayesian nonparametric and semi-parametric data analysis. So

Dear R-Users, I have the CDF of a discrete probability distribution. I now observe a change in this CDF at one point. I would like to find a new CDF such that it has the shortest Kullback-Leibler Distance to the original CDF and respects my new observation. Is there an existing package in R which will let me do this ? Google searches based on entropy revealed nothing. Kind regards, Tolga

So I recognize that: 1. many people hate forward stepwise regression (i've read the archives)--but I need it 2. step() or stepAIC are two ways to get a stepwise regression in R But here's the thing: I can't seem to figure out how to specify that I want the criteria to be Mallow's Cp (and then to subsequently tell me what the Cp stat is). I know it has something to do with

To everyone, I am currently doing analysis through Splus but am having severe problems. I am trying to fit all subset weighted regression to a multivariate data set and get PRESS statistics, along with Rsquare, Rsquare adjusted and Mallows Cp for all possible models. Splus, will give me these statistics but the leaps procedure is highly sensitive to nearly singular data matrices (which mine is)

I have just uploaded the package "stable-0.1.tgz" to the contrib section of CRAN. It enables to compute the density ('dstable'), the distribution ('pstable'), the quantile ('qstable') and the hazard ('hstable') functions of a stable variate. 'stable.mode' computes the mode of a stable distribution. The procedure 'stableglm' also enables

Using a p-value to make any kind of decision is questionable to begin with, and especially unreliable in choosing covariates in regression. Old studies, e.g. by Walls and Weeks and by Bendel and Afifi, have shown that if predictive ability is the criterion of interest and one wishes to use p-values for deciding whether to include a covariate, one should set the p-value bar very large, at 0.25 and

Hi, I am building r-patch from the sources (rsync-ed today). make check produced the following message: running tests of Internet and socket functions expect some differences make[3]: Entering directory `/usr/evahome/vograno/R/tests' running code in 'internet.R' ... OK comparing 'internet.Rout' to './internet.Rout.save' ...18c18 < Content type `text/plain;

Hello dear R-help members, I recently became interested in using biglm with leaps, and found myself somewhat confused as to how to use the two together, in different settings. I couldn't find any example codes for the leaps() package (except for in the help file, and the examples there are not as rich as they could be). That is why I turn to you in case you could share some good tips and

Hello devs. I would like to propose how I plan to go about improving and getting a system that can be integrated into Xapian in this GSoC for the clustering branch. I have identified three areas of work which were not touched last time. 1) Automated Performance Analysis I had roughly implemented 2 evaluation techniques previously (Distance b/w document and centroids within clusters and

I would like to figure out the equations for calculating "AIC" in both "step() function" and "AIC () function". They are different. Then I just type "step" in the R console, and found the "AIC" used in "step() function" is "extractAIC". I went to the R help, and found: "The criterion used is AIC = - 2*log L + k *