similar to: Cross Validation

Good afternoon. I would like to know if there is any function in R to do LOOCV with these classifiers: 1)SVM 2)Neural Networks 3)C4.5 ( J48) 4)KNN Thanks a lot! [[alternative HTML version deleted]]

Hi, I am using dlda algorithm from supclust package and I am wondering if the output can be a continuous probability instead of discrete class label (zero or one) since it puts some restriction on convariance matrix, compared with lda, while the latter can. thanks, -- Weiwei Shi, Ph.D Research Scientist GeneGO, Inc. "Did you always know?" "No, I did not. But I believed..."

Hello, I have the following function that receives a "function pointer" formal parameter name "fnc": loocv <- function(data, fnc) { n <- length(data.x) score <- 0 for (i in 1:n) { x_i <- data.x[-i] y_i <- data.y[-i] yhat <- fnc(x=x_i,y=y_i) score <- score + (y_i - yhat)^2 } score <- score/n

I am trying to write a wrapper for the last example in help(cv.glm) that deals with leave-one-out-cross-validation (LOOCV) for a logistic model. This wrapper will be used as part of a bigger program. Here is my wrapper funtion : logistic.LOOCV.err <- function( formu=NULL, data=NULL ){ cost.fn <- function(cl, pred) mean( abs(cl-pred) > 0.5 ) glmfit <- glm(

Hy, I got the following problem when trying to build a rpart model and using everything but LOOCV. Originally, I wanted to used k-fold partitioning, but every partitioning except LOOCV throws the following warning: ---- Warning message: In nominalTrainWorkflow(dat = trainData, info = trainInfo, method = method, : There were missing values in resampled performance measures. ----- Below are some

Dear all, not really a R question but: If I want to check for the classification accuracy of a LDA with previous PCA for dimensionality reduction by means of the LOOCV method: Is it ok to do the PCA on the WHOLE dataset ONCE and then run the LDA with the CV option set to TRUE (runs LOOCV) -- OR-- do I need - to compute for each 'test-bag' (the n-1 observations) a PCA

Hi! I'm using GLM, LDA and NaiveBayes for binomial classification. My training set is 70 rows long with 32 features, and my test set is 30 rows long with 32 features. Using Naive Bayes, I can train a model, and then predict the test set with it like so: ass4q1.dLDA = lda(ass4q1.trainSet[,1]~ass4q1.trainSet[,2:3]) table(predict(ass4q1.dNB, ass4q1.testSetDF[,2:3]), ass4q1.testSetDF[,1])

Hello all, Is there any function in R by which I can calculate PRESS and P2 statistics for linear regression in R? Thanks Alex [[alternative HTML version deleted]]

Using R 2.0.0 of July 20 2004 train, test, and cl as defined in example(knn), we have > search() [1] ".GlobalEnv" "package:methods" "package:stats" "package:graphics" [5] "package:utils" "Autoloads" "package:base" > knn(train, test, cl, k=3) Error: couldn't find function "knn" >

Hello. I found the behavior of knn( http://stat.ethz.ch/R-manual/R-devel/library/class/html/knn.html) function looking very strange. Consider the toy example. > library(class) > train <- matrix(nrow=5000,ncol=2,data=rnorm(10000,0,1)) > test <- matrix(nrow=10,ncol=2,data=rnorm(20,0,1)) > cl <- rep(c(0,1),2500) > knn(train,test,cl,1) [1] 1 1 0 0 1 0 1 1 0 1 Levels: 0 1 It

Hi, I was trying to get the optimal 'k' for the knn. To do this I was using the following function : knn.cvk <- function(datmat, cl, k = 2:9) { datmatT <- (datmat) cv.err <- cl.pred <- c() for (i in k) { newpre <- as.vector(knn.cv(datmatT, cl, k = i)) cl.pred <- cbind(cl.pred, newpre) cv.err <- c(cv.err, sum(cl != newpre)) }

Hello, I want to do regression or missing value imputation by knn. I searched r-help mailing list. This question was asked in 2005. ksmooth and loess were recommended. But my case is different. I have many predictors (p>20) and I really want try knn with a given k. ksmooth and loess use band width to define neighborhood size. This contrasts to knn's variable band width via fixing a

Hi I want to create a nice picture about my result of k -the nearest neighbours algorithm. Here is my easy code: ################################# library(klaR) library(ipred) library(mlbench) data(PimaIndiansDiabetes2) dane=na.omit(PimaIndiansDiabetes2)[,c(2,5,9)] dane[,2]=log(dane[,2]) dane[,1:2]=scale(dane[,1:2]) zbior.uczacy=sample(1:nrow(dane),nrow(dane)/2,F)

hi, r users i have a problem with KNN. i have 2 datasets, X0 and X1. >dim(X0) >1471*13 dim(X1) >5221*13 and for every instances in the dataset X1, i want to find the nearest neighbour(1nn) in the dataset X0. and i dont have the true classifications of dataset X1. but the function knn() need true classifications(cl) to do prediction. i just curious if there are some other function

Hello, I am feeling that this question can have a very simple answer, but I can't find it. I need to use the function knn.dist from knnflex library. Whatever I try, I get the error: Error in as.vector.dist(x, "character") : unused argument(s) ("character") First example: > a<-NULL > a<-rbind(a,c(5.2,-8.1)) > a<-rbind(a,c(8.8,-16.1)) >

ear all, This is my first time using this listserv and I am seeking help from the expert. OK, here is my question, I am trying to use impute.knn function in impute library and when I tested the sample code, I got the error as followingt: Here is the sample code: library(impute) data(khanmiss) khan.expr <- khanmiss[-1, -(1:2)] ## ## First example ## if(exists(".Random.seed"))

The function knn.cv in the class package doesn't have error checking to ensure that the length of the classlabel argument is equal to the number of rows in the test set. If the classlabel is short, the result is often a segfault. > library(class) > dat <- matrix(rnorm(1000), nrow=10) > cl <- c(rep(1,5), rep(2,5)) > cl2 <- c(rep(1,5), rep(2,4)) > knn.cv(dat, cl) [1] 2

I have to remove missing data both in character and numeric datatype.I tried using NA condition but it is not working ,please help me to solve this. -- View this message in context: http://www.nabble.com/Handling-missing-data-tp25530192p25530192.html Sent from the R help mailing list archive at Nabble.com.

hi, i am getting an error when calling the impute.knn function (see the screenshot below). what is the problem here and how can it be solved? screenshot: ################## *** caught segfault *** address 0x513c7b84, cause 'memory not mapped' Traceback: 1: .Fortran("knnimp", x, ximp = x, p, n, imiss = imiss, irmiss, as.integer(k), double(p), double(n), integer(p),

Dear List, I have a question regarding "ipred" package. Under 10-fold cv, for different knn ( = 1,3,...25), I am getting same misclassification errors: ############################################# library(ipred) data(iris) cv.k = 10 ## 10-fold cross-validation bwpredict.knn <- function(object, newdata) predict.ipredknn(object, newdata, type="class") for (i in