[R] How to create lift chart and ROC curve in R

Hi, I have a logit model built in R which I am trying to validate . Could you please help me with creating lift charts and area under ROC for the same. I am trying to use package ROCR The indicative code for this is : newlogit3<- glm(Attrition_ind ~ Time.in.AXP+ Age +as.factor(Incre

[R] Help with creating conditional categorical variables in R

Hi, I am stuck with creating a conditional categorical variable in R If my dataframe data_2 has 3 variables A,B,CI want to create variable D which would be something like : data_2$D <- ifelse(data_2$A < data_2$B & promotion_ind =="N",1,0), this throws up an error "In Ops.factor(data_2$A,dat

[R] Splitting data into test and train (80:20) kepping attributes similar

Hi, I am trying to do some predictive modeling around attrition and want to split the dataset into test and train (80:20) and keep the ratio of attritees:non attrites same. In my dataset the attrition indicator is coded as 0(for non-attritees) and 1 (for attritees) and I want to keep the ratio o

Re: [R] Splitting data into test and train (80:20) kepping attributes similar

lease help Thanks in advance doy -Original Message- From: r-help-boun...@r-project.org [mailto:r-help-boun...@r-project.org] On Behalf Of Dwaipayan Dasgupta Sent: Tuesday, April 24, 2012 9:08 PM To: r-help@r-project.org Subject: [R] Splitting data into test and train (80:20) kepping

Re: [R] Splitting data into test and train (80:20) kepping attributes similar

ary(caTools) before executing your code Am 25.04.2012 um 12:39 schrieb Dwaipayan Dasgupta: > Hi, > Could someone help me with this please , im trying to use > Y = Attrition_data[,1] # extract labels from the data > msk = sample.split (Y, SplitRatio=3/4) > table(Y,msk) > to

Re: [R] Splitting data into test and train (80:20) kepping attributes similar

tiple of vector length (arg 2) I understand that I am trying to append vectors of unequal vector lengths but don't know how to work around this process. Would you help please Thanks, Dwaipayan From: Jessica Streicher [mailto:j.streic...@micromata.de] Sent: Wednesday, April 25, 2012 9:25 P

Re: [R] Splitting data into test and train (80:20) kepping attributes similar

<- rbind(s1,s2) will not work. you need to take data from AD_0 and AD_0 for that, similarly with what you did with s3 and s4. Am 26.04.2012 um 12:56 schrieb Dwaipayan Dasgupta: > Hi , > Thanks again for helping me out. > Here is the code I am using > Ad_1 <- subse

[R] Validation of logistic models in R 2.12

Hi everyone, I am trying to validate a logistic model built in R. Not my version of R is 2.12 and I cannot install ROCR. I have gone to a point where I have the predicted values using the code: pred1 = predict(trainlogit1,testdata_1, type = "response") How do I proceed from here? Is there anoth

[R] FW: Validation of logistic models in R 2.12

Could someone help me on this please? -Original Message- From: r-help-boun...@r-project.org [mailto:r-help-boun...@r-project.org] On Behalf Of Dwaipayan Dasgupta Sent: Thursday, May 03, 2012 1:43 PM To: r-help@r-project.org Subject: [R] Validation of logistic models in R 2.12 Hi