search for: parsimony

Hi all, I know this a general question, not specific for any R package, even so I hope someone may give me his/her opinion on this. I have a set of 20 candidate models in a binomial GLM. The global model has 52 estimable parameters and sample size is made of about 1500 observations. The global model seems not to have problems of parameters estimability nor get troubles with the convergence of

Hi All, I am trying to use a neural network for my work, but I am not sure about my approach to select a parsimonious model. In R with nnet, the IAC has not been defined for a feed-forward neural network with a single hidden layer. Is this because it does not make sens mathematically in this case? For example, is this pseudo code sensible? Thanks in advance for your help. I am sorry if this

If AIC is giving you a model that is too large, then use BIC (log(n) as the penalty for adding a term in the model). This will yield a more parsimonious model. Now, if you ask me which is the better option, I have to refer you to the huge literature on model selection. Best, Ravi [[alternative HTML version deleted]]

Hi, I am looking at the effect of allelochemicals produced by two freshwater macrophyte species on two different algal species at different days. I am comparing the effect of each macrophyte on each algae at each day. I received help from someone doing the LMEM (Linear mixed-effect models) and he told me to do ANOVA to analyse the LMEM. However, I received these feedback from my examinor; 1. An

...AICc deltAICc w_ic 2 log(1e-04 + transat) 44.63758 4 7.5 -81.27516 0.000000 0.65 -79.67516 0.000000 0.57 1 1 43.02205 3 10.0 -80.04410 1.231069 0.35 -79.12102 0.554146 0.43 The usual conclusion would be that the two models are equivalent and to keep the null model for parsimony (!). However, an anova shows that the variable 'log(1e-04 + transat)' is significantly different from 0 in model 2 (lmmedt9) > anova(lmmedt9) numDF denDF F-value p-value (Intercept) 1 20 289.43109 <.0001 log(1e-04 + transat) 1 20 31....

...icanda praeter necessitatem" > > The principle states that "Entities should not be multiplied > unnecessarily." Sometimes it is quoted in one of its original Latin > forms to give it an air of authenticity. When I was a student this was referred to as the `principle of parsimony', which has the advantage of having the essence in the name. My (Concise Oxford) dictionary has this under `law of parsimony' that no more causes or forces should be assumed than are necessary to account for the facts. It seems that William of Ockham has become as generously (mis-)acknowl...

I have a simulation program that generates a data frame for each run. I aggregate the data.frames into a list (df.list). The structure of all data frames is the same, only the values are different. I then want to aggregate the various runs. Currently I use the following method (for three runs): means = (df.list[[1]]$variable + df.list[[2]]$variable + df.list[[3]]$variable)/3 I would like

Is there any really easy way to truncate integers with several consecutive digits without rounding and without converting from numeric to character (using strsplit, etc.)?? Something along these lines: e.g. = 456 truncfun(e.g., location=1) = 4 truncfun(e.g., location=1:2) = 45 truncfun(e.g., location=2:3) = 56 truncfun(e.g., location=3) = 6 It's one thing using floor(x/100) to get 4 or

There is both a technical and a theoretical element to my question... Should I be able to use the outputs which arise from the fastbw function as inputs to nomogram(). I seem to be failing at this, -- I obtain a subscript out of range error. That I can't do this may speak to technical failings, but I suspect it is because Prof Harrell thinks/knows it injudicious. However, I can't

More principled would be to use a lasso-type approach, which combines selection and estimation in one fell swoop! Ravi ________________________________ From: Ravi Varadhan Sent: Tuesday, June 6, 2017 10:16 AM To: r-help at r-project.org Subject: Subject: [R] glm and stepAIC selects too many effects If AIC is giving you a model that is too large, then use BIC (log(n) as the penalty for adding

Hi All, I have to say upfront that I am a complete neophyte when it comes to programming. Nevertheless I enjoy the challenge of using R because of its incredible statistical resources. My problem is this .........I am running a regression tree analysis using "rpart" and I need to run the calculation repeatedly (say n=50 times) to obtain a distribution of results from which I will pick

>>>> Fine. Use reiserfs and don't worry about ctime. >>>> >>> But reiserfs doesn´t support ACLs. Does it? >> >> Oh yes, it does. Big way. >> > ?? > > I was under the impression that if i wanted acls, i > should use xfs, ext3 (or jsf i believe) but NOT > reisersf. > > Am I wrong? Does (for example) SuSE 8.2 with >

.... To compute the R2, I used the following formula r2=mss/(mss+rss) where mss=sum((fitted(model)-mean(fitted(model)))^2) and rss=sum(resid(model)^2) I think that the R2 is good enough for the model selection knowing the candidate models both have two parameters (so no to care about the principle of parsimony) and my guess is that the models needs to have the same form (which is not the case here: linear form vs exponential form) ) or nested to be compared with frequentist or Bayesian approaches such as the AIC and BIC criterion . Thank you very much in advance Armel [[alternative HTML version delet...

Dear all, Please excuse my ignorance, but I am having difficulty with this, and am unable to find help on the website/Google. I have a series of explanatory variables that I am aiming to get parsimony out of. For example, if I have 10 variables, a-j, I am initially looking at the linear relationships amongst them: my.lm1 <- lm(a ~ b+c+d+e+f+g+h+i+j, data=my.data) summary(my.lm1) my.lm2 <- lm(b ~ a+c+d+e+f+g+h+i+j, data=my.data) etc Instead of repeatedly typing this in, is there a way t...

Hi, I am using the packages tree and rpart to build a classification tree to predict a 0/1 outcome. The package rpart has the advantage that the function plotcp gives a visual representation of the cross-validation results with a horizontal line indicating the 1 standard error rule, i.e. the recommendation to select the most parsimonious model (the smallest tree) whose error is not more than one

Dear members, I would like to switch from nlme to lme4 and try to translate some of my models that worked fine with lme. I have problems with the pdMat classes. Below a toy dataset with a fixed effect F and a random effect R. I gave also 2 similar lme models. The one containing pdLogChol (lme1) is easy to translate (as it is an explicit notation of the default model) The more parsimonious

Dear R-users, dear Ron I use pls from the pls.pcr package for classification. Since I need to know which variables are most influential onto the classification performance, what criteria shall I look at: a) B, the array of regression coefficients for a certain model (means a certain number of latent variables) (and: squared or absolute values?) OR b) the weight matrix RR (or R in the De

Dear expeRts, there is obviously a general trend to use model comparisons, LRT and AIC instead of Wald-test-based significance, at least in the R community. I personally like this approach. And, when using LME's, it seems to be the preferred way (concluded from postings of Brian Ripley and Douglas Bates' article in R-News 5(2005)1), esp. because of problems with the d.f. approximation.

Dear List, I would greatly appreciate help on the following matter: The RuleFit program of Professor Friedman uses partial dependence plots to explore the effect of an explanatory variable on the response variable, after accounting for the average effects of the other variables. The plot method [plot(summary(rq(y ~ x1 + x2, t=seq(.1,.9,.05))))] of Professor Koenker's quantreg program

?On 30/08/2018, 20:15, "R-devel on behalf of Hadley Wickham" <r-devel-bounces at r-project.org on behalf of h.wickham at gmail.com> wrote: On Thu, Aug 30, 2018 at 10:58 AM Martin Maechler <maechler at stat.math.ethz.ch> wrote: > > >>>>> Joris Meys > >>>>> on Thu, 30 Aug 2018 14:48:01 +0200 writes: >