similar to: safe prediction from lm

Bug in predict.lm & poly The predict function doesn't work when used with poly and newdata. For example, I'd expect the following code to work, and plot a fitted cubic to the nearly straight line: x <- 1:10 y <- x + rnorm(10)/100 plot(x,y) fit <- lm(y ~ poly(x,3)) newx <- seq(1,10,len=100) lines(newx,predict(fit,newdata=data.frame(x=newx))) However, the plotted

Hi, I was going through the examples in smartpred. It seems there's an unstated dependency on the fitted() function in package flexmix. n = 20 set.seed(86) x = sort(runif(n)) y = sort(runif(n)) library(splines) fit = lm(y ~ ns(x, df=5)) plot(x, y) lines(x, fitted(fit)) # won't work w/o prior loading of the flexmix package. newx = seq(0, 1, len=n) points(newx, predict(fit,

Dear Everybody, Version 0.5 of the smart prediction library smartpred is now available. The smartpred library consists of some modified existing and new functions to fix a problem relating to the modelling functions lm and glm. The modified versions now remember the "parameters" of special smartened functions (bs, ns, scale and poly are supplied). We call this "smart"

I've got a gls formula that's a mix of continuous and ordered variables. I wanted to use gls because I wanted to use the varIdent structure. Anyway, attempts to use "predict.gls" choke with the error that the levels I use are not allowed for one of them -- the first one alphabetically, so I'd guess the second would have the same problem. So I have three linked questions --

I have a classification tree model similar to the following (slightly simplified here): > treemod<-tree(y~x) where y is a factor and x is a matrix of numeric predictors. They have dimensions: > length(y) [1] 1163 > dim(x) [1] 1163 75 I?ve evaluated the tree model and am happy with the fit. I also have a matrix of cases that I want to use the tree model to classify. Call it

Hi, I would like to fit my data with a 4th order polynomial. Now I have only 5 data point, I should have a polynomial that exactly pass the five point Then I would like to compute the "fitted" or "predict" value with a relatively large x dataset. How can I do it? BTW, I thought the model "prodfn" should pass by (0,0), but I just wonder why the const is

Are there any general guidelines for writing functions that appear in model formulas for functions like lm()/glm()? The ones I can think of are ns()/bs() in `splines' and s() in `mgcv'. It seems that ns()/bs() basically create matrix with a number of attributes and a class, while s() returns a list of parameters. It seems that gam() in `mgcv' has its own way of dealing with

Hi there, Could somebody help me disect this legacy R script I inherited at work, I have two questions: 1. I've tried to upgrade our R version from 1.6.2 (yeah, I know), to R 2.0, but some of the lines in this script are not compatible with R 2.0, could someone help me figure out where the problem is? 2. the jpeg generated (attached) seems to be off on some of the data, is there a better way

Hello everyone I'm a beginner in Stats and R, I'm using R 2.10.1. I need to create a multivariate qq plot, there is 8 variable group with each has 55 number of input. An example of what I did so far, just to get my point out: > data=read.csv(file.choose(),header=T) > data country village group av_expen P2ary_ed no_fisher 1 Cook Islands Aitutaki D

I'm trying to do the following out of sample regression with autoregressive terms and additional x variables: y(t+1)=const+B(L)*y(t)+C(1)*x_1(t)...+C(K)*x_K(t) where: B(L) = lag polynom. for AR terms C(1..K) = are the coeffs. on K exogenous variables that have only 1 lag Question 1: ----------- Suppose I use arima to fit the model:

I'd like to compare two models which were fitted using gls, however I'm having trouble interpreting the results of gls. If any of you could offer me some advice, I'd greatly appreciate it. Short explanation of models: These two models have the same fixed-effects structure (two independent, linear effects), and differ only in that the second model includes a corExp structure for

I think that the "dimname names" of tables and arrays could make aperm() and apply() (and probably some other functions) easier to use. (dimname names are, for example, created by table() ) The use would be something like: -- x <-table( from=sample(3,100,rep=T), to=sample(5,100,rep=T)) trans <- x / apply(x,"from",sum) y <- aperm( trans,

Hello, How can I label a secondary axis in R? At the moment it's labelled as c(-100,200). Obviously I would like it to be more sensible. Here is the code I am using newx = -100+37.5*((1:9)-1) axis(4,at=newx,labels=(newx+100)/3750) Thanks, Rob -- View this message in context: http://www.nabble.com/Labelling-a-secondary-axis-in-R-tp16807708p16807708.html Sent from the R help mailing list

Hi, I am segfaulting when using predict.loess() (checked with r62092). I've traced the source with the help of valgrind (output pasted below) and it appears that this is due to int overflow when allocating an int work array in loess_workspace(): liv = 50 + ((int)pow((double)2, (double)D) + 4) * nvmax + 2 * N; where liv is an (global) int. For D=1 (one x variable), this overflows at

Dear R- helpers Following a draft structure of the R script for which I am facing problem Step 1 x <- of type array with original values y <- of type array with original values Step 2 for (ctr in 1:10) { # my problem here the both x and y still show the original values from step 1 # in spite of making changes to the old values of the arrays x and y in the function function

Thanks for the reproducable example. I can confirm that it fails on my machine using survival 2-37.5, the next soon-to-be-released version, The issue is with NextMethod, and my assumption that the called routine inherited everything from the parent, including the environment chain. A simple test this AM showed me that the assumption is false. It might have been true for Splus. Working this

Hello R Gurus, I am perplexed by the different results I obtained when I ran code like this: set.seed(100) test1<-randomForest(BinaryY~., data=Xvars, trees=51, mtry=5, seed=200) predict(test1, newdata=cbind(NewBinaryY, NewXs), type="response") and this code: set.seed(100) test2<-randomForest(BinaryY~., data=Xvars, trees=51, mtry=5, seed=200, xtest=NewXs, ytest=NewBinarY) The

A couple weeks ago I posted a message on this topic to r-help, the response was that this seemed like odd behavior, and that I ought to post it to one of the developer lists. I posted to r-sig-mixed-models, but didn't get any response. So, with good intentions, I decided to try posting once more, but to this more general list. The goal is (1) FYI, to make you aware of this issue, in case it

Colleagues, I am using "nls" successfully (2.11.1, OS X) but I am having difficulties retrieving part of the output - residual sum of squares. I have assigned the output to FIT: > > FIT > Nonlinear regression model > model: NEWY ~ PMESOR + PAMPLITUDE * cos(2 * pi * (NEWX - POFFSET)/PERIOD) > data: parent.frame() > PMESOR PAMPLITUDE POFFSET >

I'm making some functions to illustrate regressions and I have been staring at termplot and predict.lm and residuals.lm to see how this is done. I've wondered who wrote predict.lm originally, because I think it is very clever. I got interested because termplot doesn't work with interactive models: > m1 <- lm(y ~ x1*x2) > termplot(m1) Error in `[.data.frame`(mf, , i) :