I have a data set (df) in this format
index <- runif(n = 100,min = 0, max = 1)
type1 <- rep("low", 50)
type2 <- rep("high", 50)
type <- c(type1,type2)
level1 <- rep("single", 25)
level2 <- rep("multiple", 25)
level3 <- rep("single", 25)
level4 <- rep("multiple", 25)
level <- c(level1,level2,level3,level4)
block <- rep(1:5, 10)
set <- rep(1:5, 10)
df <- data.frame("index" = index,"type" = type, "level" = level, "block" = block, "set" = set)
df$block <- as.factor(df$block)
df$set <- as.factor(df$set)
I want to create a model that looks like like this
model <- lmer(index ~ type * level + (1|block) + (1|set), data = df)
However, in my original data the fit is bad because the data is bound between 0 and 1. I want to bootstrap this mixed effects model. Any idea on how to achieve boot-strapping for such a model? I want to compare this this full model with sub-models eg. without interaction, or with level or type alone. I also want with confidence intervals for the final model
The confint() function has a method for merMod objects. The following should work:
confint(model, method = "boot", nsim = 1000)
And with multiple CPUs:
confint(model, method = "boot", nsim = 1000,
parallel = "multicore", ncpus = 8)
Related
For a study I am working on I need to create bootstrapped datasets and inverse probability weights for each dataset and then run a series of models for each of these datasets/weights. I am attempting to do this with a nested for-loop where the first part of the loop creates the weights and the nested loop runs a series of models, each with different outcome variables and/or predictors. I am running about 80 models for each bootstrapped dataset, hence the reason for a more automated way to do this. Below is a example of what I am doing with some mock data:
# Creation of mock data
data <- data.frame("Severity" = as.factor(c(rep("None", 25), rep("Mild", 25), rep("Moderate", 25), rep("Severe", 25))), "Severity2" = as.factor(c(rep("None", 40), rep("Mild", 20), rep("Moderate", 20), rep("Severe", 20))), "Weight" = rnorm(100, mean = 160, sd = 30), "Age" = rnorm(100, mean = 40, sd = 7), "Gender" = as.factor(rbinom(100, size = 1, prob = 0.5)), "Tested" = as.factor(rbinom(100, size = 1, prob = 0.4)))
data$Severity <- ifelse(data$Tested == 0, NA, data$Severity)
data$Severity2 <- ifelse(data$Tested == 0, NA, data$Severity2)
data$Severity <- ordered(data$Severity, levels = c("None", "Mild", "Moderate", "Severe"))
data$Severity2 <- ordered(data$Severity2, levels = c("None", "Mild", "Moderate", "Severe"))
# Creating boostrapped datasets
nboot <- 2
set.seed(10)
boot.samples <- lapply(1:nboot, function(i) {
data[base::sample(1:nrow(data), replace = TRUE),]
})
# Create empty list to store results later
coefs <- list()
# Setting up the outcomes/predictors of each of the models I will run
mod1 <- list("outcome" <- "Severity", "preds" <- c("Weight","Age"))
mod2 <- list("outcome" <- "Severity2", "preds" <- c("Weight", "Age", "Gender"))
models <- list(mod1, mod2)
# Running the for-loop
for(i in 1:length(boot.samples)) {
#Setting up weight creation
null <- glm(formula = Tested ~ 1, family = "binomial", data = boot.samples[[i]])
full <- glm(formula = Tested ~ Age, family = "binomial", data = boot.samples[[i]])
step <- step(null, k = 2, direction = "forward", scope=list(lower = null, upper = full), trace = 0)
pd.combined <- stats::predict(step, type = "response")
numer.combined <- glm(Tested ~ 1, family = "binomial",
data = boot.samples[[i]])
pn.combined <- stats::predict(numer.combined, type = "response")
# Creating stabilized weights
boot.samples[[i]]$ipw <- ifelse(boot.samples[[i]]$Tested==0, ((1-pn.combined)/(1-pd.combined)), (pn.combined)/(pd.combined))
# Now running each model and storing the coefficients
for(j in 1:length(models)) {
outcome <- models[[j]][[1]] # Set the outcome name
predictors <- models[[j]][[2]] # Set the predictor names
model_results <- polr(boot.samples[[i]][,outcome] ~ boot.samples[[i]][, predictors], weights = boot.samples[[i]]$ipw, method = c("logistic"), Hess = TRUE) #Run the model
coefs[[j]] <- model_results$coefficients # Store regression model coefficients in list
}
}
The portion for creating the IPW weights works just fine, but I keep getting an error for the modeling portion that reads:
"Error in model.frame.default(formula = boot.samples[[i]][, outcome] ~ :
invalid type (list) for variable 'boot.samples[[i]][, predictors]'"
Based on the question asked and answered here: Error in model.frame.default ..... : invalid type (list) for variable I know that the issue is with how I'm calling the outcomes and predictors in the model. I've messed around lots of different ways to handle this to no avail, I need to specify the outcome and predictors as I do because in my actual models the outcomes and predictors changes with each model! Any ideas on how to deal with this would be greatly appreciated!
I've tried something like setting outcome <- boot.samples[[i]][,outcome] outside of the model and then just calling outcome in the model, but that gives me the same error.
I am trying to understand why changing the reference level of a factor changes the results of a model. Consider this example:
library(liver)
library(caret)
library(glmnet)
library(dplyr)
data(churn)
head(churn)
# set reference levels
churn$state <- relevel(churn$state, ref = "NE")
churn$area.code <- relevel(churn$area.code, ref = "area_code_408")
churn$intl.plan <- relevel(churn$intl.plan, ref = "yes")
churn$voice.plan <- relevel(churn$voice.plan, ref = "no")
# split into train and test
set.seed(1)
train.index <- createDataPartition(churn$churn, p = 0.8, list = FALSE)
train_churn <- churn[train.index,]
test_churn <- churn[-train.index,]
# add class weights
my_weights = train_churn %>%
select(churn) %>%
group_by(churn) %>%
count()
weight_for_yes = (1 / my_weights$n[1]) * ((my_weights$n[1] + my_weights$n[2]) / 2.0)
weight_for_yes
weight_for_no = (1 / my_weights$n[2]) * ((my_weights$n[1] + my_weights$n[2]) / 2.0)
weight_for_no
model_weights <- ifelse(train_churn$churn == "yes", weight_for_yes, weight_for_no)
# tuning grid
myGrid <- expand.grid(
alpha = 0,
lambda = seq(0,1,0.01)
)
set.seed(1)
mod_1 <- train(churn ~
state +
area.code +
intl.plan +
voice.plan,
data = train_churn,
method = "glmnet",
tuneGrid = myGrid,
weights = model_weights)
mod_1
Tuning parameter 'alpha' was held constant at a value of 0
Accuracy was used to select the optimal model using the largest value.
The final values used for the model were alpha = 0 and lambda = 0.8
prediction <- predict(mod_1, newdata = test_churn)
confusionMatrix(prediction, test_churn$churn)
I also look at a single new prediction
new_data = data.frame(state = c("CA"),
area.code = c("area_code_510"),
intl.plan = c("yes"),
voice.plan = c("no"))
predict(mod_1, newdata = new_data, type = "prob")
Now I restart R, set new reference levels, and rerun all the code. This is the output
# set new reference levels
churn$state <- relevel(churn$state, ref = "OR")
churn$area.code <- relevel(churn$area.code, ref = "area_code_415")
churn$intl.plan <- relevel(churn$intl.plan, ref = "no")
churn$voice.plan <- relevel(churn$voice.plan, ref = "yes")
Tuning parameter 'alpha' was held constant at a value of 0
Accuracy was used to select the optimal model using the largest value.
The final values used for the model were alpha = 0 and lambda = 0.62.
I expected the lambdas to change, but I did not expect the confusion matrix or classification probabilities to changes. Is this normal?
I want to run a between-within design MANCOVA with R, with two dependent variables (Planned and Unplanned), two between-subject variables (Genre [Male, Female] and Urb [Yes, No]), one within-subject variable (Period [Before, During]), and one covariate (BMI).
Here is what I've done (see here for similar calculation: https://stats.stackexchange.com/questions/183441/correct-way-to-perform-a-one-way-within-subjects-manova-in-r):
# Create dummy data
data <- data.frame(Quest_before_planned = sample(1:100, 10),
Quest_during_planned = sample(1:100, 10),
Quest_before_unplanned = sample(1:100, 10),
Quest_during_unplanned = sample(1:100, 10),
Genre = sample(rep(c("Male", "Female"), each = 5)),
Urb = sample(rep(c("Yes", "No"), each = 5)),
BMI = sample(1:100, 10))
# Define the within-subjects factor
period <- as.factor(rep(c('before','during'), each = 2))
idata <- data.frame(period)
# Create the data structure for the linear model
data.model <- with(data, cbind(Quest_before_planned, Quest_during_planned,
Quest_before_unplanned, Quest_during_unplanned))
# Build the multivariate-linear model
mod.mlm <- lm(data.model ~ Genre * Urb, data = data_total)
# Run the MANOVA
mav.blpaq <- Anova(mod.mlm, idata = idata, idesign = ~ period, type = 2)
print(mav.blpaq)
Thus, the between-within design MANOVA here works well. However, I failed to add a covariate (i.e., BMI) to this model. Do you know how can I achieve this?
N.B.: I also tried using the (great) mancova() function , which include a covariate parameter; but with this function, I do not know how to specify that Period is a within-subject variable...
blpaq_macov <- mancova(data_tidy,
deps = c("Quest_planned", "Quest_unplanned"),
factors = c("Genre", "Period", "Urb"),
covs = "BMI",
multivar = "pillai")
I want to model insurance claim count using a Poisson glmnet. The data I have at hand contains the number of claims for each policy (which is the response variable), some features about the policy (gender, region, etc.) as well as the duration of the policy (in years). I want to include the log-duration as an offset term, as we usually do in actuarial science. With the cv.glmnet function of the glmnet package, it is straightforward:
library(tidyverse)
library(glmnet)
n <- 100
dat <- tibble(
nb_claims = rpois(n, lambda = 0.5),
duration = runif(n),
x1 = runif(n),
x2 = runif(n),
x3 = runif(n)
)
fit <- cv.glmnet(
x = dat %>% dplyr::select(x1, x2, x3) %>% as.matrix(),
y = dat %>% pull(nb_claims),
family = "poisson",
offset = dat %>% pull(duration) %>% log()
)
fit
However, my goal is to train this model using the train function of the caret package, because of the many advantages it gives. Indeed, validation, preprocessing as well as feature selection is much better with this package. It is straightforward to train a basic glmnet (without an offset term) with caret:
library(caret)
fit <- caret::train(
x = dat %>% dplyr::select(x1, x2, x3) %>% as.matrix(),
y = dat %>% pull(nb_claims),
method = "glmnet",
family = "poisson"
)
fit
Naively, we could try to add the offset argument in the train function:
fit <- caret::train(
x = dat %>% dplyr::select(x1, x2, x3) %>% as.matrix(),
y = dat %>% pull(nb_claims),
method = "glmnet",
family = "poisson",
offset = dat %>% pull(duration) %>% log()
)
fit
Unfortunately, this code throws the error Error : No newoffset provided for prediction, yet offset used in fit of glmnet. This error occurs because the caret::train function doesn't take care to give a value for the newoffset argument in predict.glmnet function.
In this book, they show how to add an offset term to a GLM model by modifying the source code of the caret::train function. It works perfectly. However, the predict.glm function is quite different from the predict.glmnet function, because it does not have the newoffset argument. I tried to modify the source code of the caret::train function, but I am having some trouble because I do not know well enough how this function works.
A simple way to perform this is pass the offset column as part of x and in each fit and predict call pass as x columns of x which are not the offset. While as offset/newoffset pass the x column corresponding to the offset.
In the following example the offest column of x needs to be named "offset" too. This can be changed relatively easy
To create the function we will just use lots of parts from: https://github.com/topepo/caret/blob/master/models/files/glmnet.R
glmnet is peculiar since it needs a loop, the rest is just rinse and reapeat from https://topepo.github.io/caret/using-your-own-model-in-train.html#illustrative-example-1-svms-with-laplacian-kernels
family = "poisson" will be specified throughout, to change this adopt code from https://github.com/topepo/caret/blob/master/models/files/glmnet.R
glmnet_offset <- list(type = "Regression",
library = c("glmnet", "Matrix"),
loop = function(grid) {
alph <- unique(grid$alpha)
loop <- data.frame(alpha = alph)
loop$lambda <- NA
submodels <- vector(mode = "list", length = length(alph))
for(i in seq(along = alph)) {
np <- grid[grid$alpha == alph[i],"lambda"]
loop$lambda[loop$alpha == alph[i]] <- np[which.max(np)]
submodels[[i]] <- data.frame(lambda = np[-which.max(np)])
}
list(loop = loop, submodels = submodels)
})
glmnet_offset$parameters <- data.frame(parameter = c('alpha', 'lambda'),
class = c("numeric", "numeric"),
label = c('Mixing Percentage', 'Regularization Parameter'))
glmnet_offset$grid <- function(x, y, len = NULL, search = "grid") {
if(search == "grid") {
init <- glmnet::glmnet(Matrix::as.matrix(x[,colnames(x) != "offset"]), y,
family = "poisson",
nlambda = len+2,
alpha = .5,
offset = x[,colnames(x) == "offset"])
lambda <- unique(init$lambda)
lambda <- lambda[-c(1, length(lambda))]
lambda <- lambda[1:min(length(lambda), len)]
out <- expand.grid(alpha = seq(0.1, 1, length = len),
lambda = lambda)
} else {
out <- data.frame(alpha = runif(len, min = 0, 1),
lambda = 2^runif(len, min = -10, 3))
}
out
}
So x[,colnames(x) != "offset"] is x while offset is x[,colnames(x) == "offset"]
glmnet_offset$fit <- function(x, y, wts, param, last, ...) {
theDots <- list(...)
## pass in any model weights
if(!is.null(wts)) theDots$weights <- wts
if(!(class(x)[1] %in% c("matrix", "sparseMatrix")))
x <- Matrix::as.matrix(x)
modelArgs <- c(list(x = x[,colnames(x) != "offset"],
y = y,
alpha = param$alpha,
family = "poisson",
offset = x[,colnames(x) == "offset"]),
theDots)
out <- do.call(glmnet::glmnet, modelArgs)
if(!is.na(param$lambda[1])) out$lambdaOpt <- param$lambda[1]
out
}
glmnet_offset$predict <- function(modelFit, newdata, submodels = NULL) {
if(!is.matrix(newdata)) newdata <- Matrix::as.matrix(newdata)
out <- predict(modelFit,
newdata[,colnames(newdata) != "offset"],
s = modelFit$lambdaOpt,
newoffset = newdata[,colnames(newdata) == "offset"],
type = "response") #important for measures to be appropriate
if(is.matrix(out)) out <- out[,1]
out
if(!is.null(submodels)) {
tmp <- as.list(as.data.frame(predict(modelFit,
newdata[,colnames(newdata) != "offset"],
s = submodels$lambda,
newoffset = newdata[,colnames(newdata) == "offset"],
type = "response"),
stringsAsFactors = TRUE))
out <- c(list(out), tmp)
}
out
}
For some reason which I don't understand yet it does not work without the prob slot
glmnet_offset$prob <- glmnet_offset$predict
glmnet_offset$tags = c("Generalized Linear Model", "Implicit Feature Selection",
"L1 Regularization", "L2 Regularization", "Linear Classifier",
"Linear Regression")
glmnet_offset$sort = function(x) x[order(-x$lambda, x$alpha),]
glmnet_offset$trim = function(x) {
x$call <- NULL
x$df <- NULL
x$dev.ratio <- NULL
x
}
library(tidyverse)
library(caret)
library(glmnet)
n <- 100
set.seed(123)
dat <- tibble(
nb_claims = rpois(n, lambda = 0.5),
duration = runif(n),
x1 = runif(n),
x2 = runif(n),
x3 = runif(n)
)
x = dat %>%
dplyr::select(-nb_claims) %>%
mutate(offset = log(duration)) %>%
dplyr::select(-duration) %>%
as.matrix
fit <- caret::train(
x = x,
y = dat %>% pull(nb_claims),
method = glmnet_offset,
)
fit
100 samples
4 predictor
No pre-processing
Resampling: Bootstrapped (25 reps)
Summary of sample sizes: 100, 100, 100, 100, 100, 100, ...
Resampling results across tuning parameters:
alpha lambda RMSE Rsquared MAE
0.10 0.0001640335 0.7152018 0.01805762 0.5814200
0.10 0.0016403346 0.7152013 0.01805684 0.5814193
0.10 0.0164033456 0.7130390 0.01798125 0.5803747
0.55 0.0001640335 0.7151988 0.01804917 0.5814020
0.55 0.0016403346 0.7150312 0.01802689 0.5812936
0.55 0.0164033456 0.7095996 0.01764947 0.5783706
1.00 0.0001640335 0.7152033 0.01804795 0.5813997
1.00 0.0016403346 0.7146528 0.01798979 0.5810811
1.00 0.0164033456 0.7063482 0.01732168 0.5763653
RMSE was used to select the optimal model using the smallest value.
The final values used for the model were alpha = 1 and lambda = 0.01640335.
predict(fit$finalModel, x[,1:3], newoffset = x[,4]) #works
This will not work with preprocessing in caret since we pass offset as one of the features. However it will work with recipes since you can define columns on which preprocessing functions will be performed via selections. Se article for details: https://tidymodels.github.io/recipes/articles/Selecting_Variables.html
I haven't had time to error check my code. If any problems occur or if there is a mistake somewhere please comment. Thanks.
You can also post an issue in caret github asking this feature (offset/newoffset) to be added to the model
I tried to change the model info a lot of ways, but it was failing miserably. Below I can propose one solution, may not be the best, but will get you somewhere if your data is sensible.
In the poisson / negative binom .. regression, the offset in factor gets introduced into the regression, you can read more here and here:
where tx is the offset. In glmnet, there is a penalty factor you can introduce for each term, and if you let that be 0 for a term, basically you are not penalizing it and it's always included. We can use that for the offset, and you can see this effect only if you use a dataset that makes some sense (note that in your example dataset, the offsets are numbers that make no sense).
Below I use the insurance claims dataset from MASS:
library(tidyverse)
library(glmnet)
library(MASS)
dat <- Insurance
X = model.matrix(Claims ~ District + Group + Age,data=dat)
Y = dat$Claims
OFF = log(dat$Holders)
fit_cv <- cv.glmnet(
x = X,
y = Y,
family = "poisson",
offset = OFF
)
Now using caret, I will fit it without any training, and using the same lambda obtained from the fit in cv.glmnet. One thing you should note too is that cv.glmnet often uses lambda.1se instead of lambda.min:
fit_c <- caret::train(
x = cbind(X,OFF),
y = Y,
method = "glmnet",
family = "poisson",
tuneGrid=data.frame(lambda=fit_cv$lambda.1se,alpha=1),
penalty=c(rep(1,ncol(X)),0),
trControl = trainControl(method="none")
)
We can see how different are the predictions:
p1 = predict(fit_cv,newx=X,newoffset=OFF)
p2 = predict(fit_c,newx=cbind(X,OFF))
plot(p1,p2)
I am using the following code to generate data, and i am estimating regression models across a list of variables (covar1 and covar2). I have also created confidence intervals for the coefficients and merged them together.
I have been examining all sorts of examples here and on other sites, but i can't seem to accomplish what i want. I want to stack the results for each covar into a single data frame, labeling each cluster of results by the covar it is attributable to (i.e., "covar1" and "covar2"). Here is the code for generating data and results using lapply:
##creating a fake dataset (N=1000, 500 at treated, 500 at control group)
#outcome variable
outcome <- c(rnorm(500, mean = 50, sd = 10), rnorm(500, mean = 70, sd = 10))
#running variable
running.var <- seq(0, 1, by = .0001)
running.var <- sample(running.var, size = 1000, replace = T)
##Put negative values for the running variable in the control group
running.var[1:500] <- -running.var[1:500]
#treatment indicator (just a binary variable indicating treated and control groups)
treat.ind <- c(rep(0,500), rep(1,500))
#create covariates
set.seed(123)
covar1 <- c(rnorm(500, mean = 50, sd = 10), rnorm(500, mean = 50, sd = 20))
covar2 <- c(rnorm(500, mean = 10, sd = 20), rnorm(500, mean = 10, sd = 30))
data <- data.frame(cbind(outcome, running.var, treat.ind, covar1, covar2))
data$treat.ind <- as.factor(data$treat.ind)
#Bundle the covariates names together
covars <- c("covar1", "covar2")
#loop over them using a convenient feature of the "as.formula" function
models <- lapply(covars, function(x){
regres <- lm(as.formula(paste(x," ~ running.var + treat.ind",sep = "")), data = d)
ci <-confint(regres, level=0.95)
regres_ci <- cbind(summary(regres)$coefficient, ci)
})
names(models) <- covars
print(models)
Any nudge in the right direction, or link to a post i just haven't come across, is greatly appreciated.
You can use do.call were de second argument is a list (like in here):
do.call(rbind, models)
I made a (possible) improve to your lapply function. This way you can save the estimated parameters and the variables in a data.frame:
models <- lapply(covars, function(x){
regres <- lm(as.formula(paste(x," ~ running.var + treat.ind",sep = "")), data = data)
ci <-confint(regres, level=0.95)
regres_ci <- data.frame(covar=x,param=rownames(summary(regres)$coefficient),
summary(regres)$coefficient, ci)
})
do.call(rbind,models)