ICC unused argument - r

I need to calculate the 95% CI of iCC.
I'm using this code:
icc(mydata[,c(1,2)], model = "twoway",type = "agreement", unit = "average")
I obtain this error message.
Error in icc(mydata[, c(1, 1)], model = "twoway", type = "agreement", :
unused arguments (model = "twoway", type = "agreement", unit = "average")
My data:
mydata= data.frame(A=c(0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 4, 0, 0, 0, 0, 3), B=c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 5, 0, 0, 0, 0, 2))
How can I fix it?
Thank you!

You might have competing packages loaded that both use icc(). The psych package and the irr package both have this as a function name. If both are loaded, the psych package is likely masking the irr command. Those are not valid arguments for psych::icc() but they are for irr::icc()
Try:
irr:: icc(mydata[,c(1,2)], model = "twoway",type = "agreement", unit = "average")

Related

Holt Winters Model Failing with Error "Optimization Failure" in R

I was trying to forecast a model using holt winters in R and this failed as showing the below mentioned error. I am not sure what is the issue and how to bypass it? I am actually forecasting this on loop so, because of this error my other time series forecasts with other SKU's are not getting completed.
Can someone help me out with this. Please
Attached is the dataset which is used and the code which I ran for the same.
library(dplyr)
library(tidyverse)
library(tidyr)
library(tidymodels)
library(forecast)
library(prophet)
library(readxl)
pd1 <- readxl::read_excel("C:X/X/X")
colnames(pd1)[1]="X1"
Y1 = ts(data = pd1$X1,
frequency = 12,
start = c(2019,1),
end = c(2022,8))
training_Y1 = window(x=Y1,
end = c(2022,8))
Model_Y1 = HoltWinters(x=training_Y1,
seasonal = "additive")
The data frame pd1 is given below:
structure(list(X1 = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 38, 7, 7,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0)), row.names = c(NA, -44L), class = "data.frame")

Why am I getting this error when using glmnet in R

When trying to predict I am getting this error
error in evaluating the argument 'x' in selecting a method for function 'as.matrix': Cholmod error 'X and/or Y have wrong dimensions' at file ../MatrixOps/cholmod_sdmult.c, line 90
Here is my code so far
library(data.table)
library(caret)
library(Metrics)
library(glmnet)
library(plotmo)
library(lubridate)
#Reading in the necessary data needed for this project
train <- fread("project/volume/data/processed/start_train.csv")
test<-fread("project/volume/data/processed/start_test.csv")
example_sub<-fread("project/volume/data/processed/example_submission.csv")
card_tab <- fread("project/volume/data/processed/card_tab.csv")
#Merging the card_tab dataset with both my train and test datasets to add more variables to each
train = merge(train, card_tab, by = "id")
test = merge(test, card_tab, by = "id")
train$power = as.numeric(train$power)
train$toughness = as.numeric(train$toughness)
test$power = as.numeric(test$power)
test$toughness = as.numeric(test$toughness)
train$powerovertough = train$power/train$toughness
test$powerovertough = test$power/test$toughness
train$current_date<-as_date(train$current_date)
train<-train[order(-current_date)]
test$current_date<-as_date(test$current_date)
test<-test[order(-current_date)]
#Handling NA values in both train and test. The NA values will be set to 0
train[is.na(train)] <- 0
test[is.na(test)] <- 0
# Specifying which columns of our model we will be dropping and applying it to our train and test datasets
drops<- c('id','future_date','current_date','card_name','power','loyalty','cmc','type','colors','mana_cost','subtypes', 'text','set','set_name')
train<-train[, !drops, with = FALSE]
test<-test[, !drops, with = FALSE]
#Saving the response variable in the train dataset
train_y<-train$future_price
test$future_price<-0
#Using dummies to train model
dummies <- dummyVars(future_price ~ ., data = train)
train<-predict(dummies, newdata = train)
test<-predict(dummies, newdata = test)
train<-data.table(train)
test<-data.table(test)
#Cross validating the model to fin the best lamda value
train<-as.matrix(train)
test<-as.matrix(test)
gl_model<-cv.glmnet(train, train_y, alpha = 1,family="gaussian")
bestlam<-gl_model$lambda.min
# Fitting the full model
gl_model<-glmnet(train, train_y, alpha = 1,family="gaussian")
plot_glmnet(gl_model)
saveRDS(gl_model,"./project/volume/models/gl_model.model")
test<-as.matrix(test)
#use the full model
pred<-predict(gl_model,s=bestlam, newx = test)
I am trying to predict future_price for my test set. The error is saying my dimensions are wrong but I don't know what could be causing them to be different. I have tried observing the data sets as it runs through the code and they seem to have the same variables.
Here is the dput
> dput(head(train))
structure(c(0.25, 0.1, 0.1, 0.1, 0.25, 0.25, 1, 0, 0, 1, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 3, 0, 2, 0, 0, 0, 0.333333333333333,
0, 1, 0), .Dim = c(6L, 20L), .Dimnames = list(NULL, c("current_price",
"typesArtifact", "typesArtifact Creature", "typesCreature", "typesEnchantment",
"typesEnchantment Artifact", "typesEnchantment Creature", "typesInstant",
"typesLand", "typesPlaneswalker", "typesSorcery", "supertypes",
"supertypesBasic", "supertypesLegendary", "rarityCommon", "rarityMythic",
"rarityRare", "rarityUncommon", "toughness", "powerovertough"
)))
> dput(head(test))
structure(c(0.15, 0.16, 2, 0.39, 0.16, 0.19, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0,
0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0,
1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0,
1, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0, 0), .Dim = c(6L, 18L), .Dimnames = list(
NULL, c("current_price", "typesArtifact", "typesArtifact Creature",
"typesCreature", "typesEnchantment", "typesInstant", "typesLand",
"typesPlaneswalker", "typesSorcery", "supertypes", "supertypesBasic",
"supertypesLegendary", "rarityCommon", "rarityMythic", "rarityRare",
"rarityUncommon", "toughness", "powerovertough")))

What does this error mean "order(vertex_attr(g, measure), decreasing = TRUE) : argument 1 is not a vector" in R?

I am trying to calculate robustness, a graph theory measure using R (braingraph package).
Robustness = robustness(my_networkgraph, type = c("vertex"), measure = ("btwn.cent"))
I get the following error, when I use the above robustness function:
Error in order(vertex_attr(g, measure), decreasing = TRUE) : argument 1 is not a vector
Any idea, what I am doing wrong here?
My network, which is a matrix has been converted to igraph object and robustness was calculated.
My network as a matrix:
mynetwork <- matrix(c(0, 1, 0, 1, 0, 0, 0, 0,
1, 0, 1, 0, 0, 0, 0, 0,
0, 1, 0, 0, 0, 0, 0, 0,
1, 0, 0, 0, 0, 1, 0, 0,
0, 0, 0, 0, 0, 1, 0, 0,
0, 0, 0, 1, 1, 0, 1, 1,
0, 0, 0, 0, 0, 1, 0, 0,
0, 0, 0, 0, 0, 1, 0, 0), nrow = 8)
This matrix was converted as igraph using the following code:
my_networkgraph <-graph_from_adjacency_matrix(mynetwork, mode = c("undirected"),weighted = NULL, diag = TRUE, add.colnames = NULL, add.rownames = NA)
Please help me to understand the above error
Thanks
Priya
There was a bug in the above function. To run the robustness code, you will need to supply a vertex attribute to your network: V(network)$degree <- degree(network) V(network)$btwn.cent <- centr_betw(network)$res

Planned Contrasts on glmmTMB

Apologies if this is a repeat question. Many have posted looking looking for a way to do post-hoc analyses on the conditional model (fixed factors) in glmmTMB. I want to do plannned contrasts between certain groups, not test every pairwise comparison (e.g. Tukey).
The code below worked well on nlme:lme for a lmm. However, it returns an error on the code below.
Error in modelparm.default(model, ...) :
dimensions of coefficients and covariance matrix don't match
Is there a way to do planned contrasts on a glmmTMB?
#filtdens is a dataframe and TRT,DATE,BURN,VEG are factors
filtdens <- merged %>% filter(!BLOCK %in% c("JB2","JB4","JB5") & MEAS =="DENS" &
group == "TOT" & BURN == "N" & VEG == "C")
filtdens$TD <- interaction(filtdens$TRT, filtdens$DATE)
mod2 <- glmmTMB(count~(TD)+(1|BLOCK),
data=filtdens,
zi=~1,
family=nbinom1(link = "log"))
k1 <- matrix(c(0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, -1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, -1, 1, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, -1, 0, 1, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, -1, 1, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, -1, 1, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, -1, 0, 1, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, -1, 1, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, -1, 1, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, -1, 0, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -1, 1), byrow = T, ncol = 12)
summary(glht(mod2, linfct=k1),test=adjusted("bonferroni"))
A reproducible example would be helpful, but: this vignette in the development version offers code that ought to enable multcomp::linfct, i.e.:
glht_glmmTMB <- function (model, ..., component="cond") {
glht(model, ...,
coef. = function(x) fixef(x)[[component]],
vcov. = function(x) vcov(x)[[component]],
df = NULL)
}
modelparm.glmmTMB <- function (model,
coef. = function(x) fixef(x)[[component]],
vcov. = function(x) vcov(x)[[component]],
df = NULL, component="cond", ...) {
multcomp:::modelparm.default(model, coef. = coef., vcov. = vcov.,
df = df, ...)
}
Test (this example is with Tukey, but I don't see why it shouldn't work more generally ...)
library(glmmTMB)
data("cbpp",package="lme4")
cbpp_b1 <- glmmTMB(incidence/size~period+(1|herd),
weights=size,family=binomial,
data=cbpp)
g1 <- glht(cbpp_b1, linfct = mcp(period = "Tukey"))
summary(g1)
This works with the current CRAN version, but the current development version of glmmTMB offers more options (e.g. emmeans(); see the above-linked vignette). You'll need to install via devtools::install_github("glmmTMB/glmmTMB/glmmTMB") (you'll need compilation tools installed as well).

Using car::Anova package for a doubly-multivariate MANOVA in R

I'm trying to run a repeated-measures MANOVA in R, which also contains a number of dependent variables (key outcome variables of behavioural tasks). The repeated-measures are due to a cross-over design, in which individuals took a drug and placebo (in randomised order).
The code I'm running looks like this:
imatrix <- matrix(c(
1, 0, 0, 0, 0, 0, 1,
1, 0, 0, 0, 0, 0, -1,
0, 1, 0, 0, 0, 0, 1,
0, 1, 0, 0, 0, 0, -1,
0, 0, 1, 0, 0, 0, 1,
0, 0, 1, 0, 0, 0, -1,
0, 0, 0, 1, 0, 0, 1,
0, 0, 0, 1, 0, 0, -1,
0, 0, 0, 0, 1, 0, 1,
0, 0, 0, 0, 1, 0, -1,
0, 0, 0, 0, 0, 1, 1,
0, 0, 0, 0, 0, 1, -1
), 12, 7, byrow=TRUE)
colnames(imatrix) <- c("BCST", "CGT", "AST", "AGN", "DDT", "FERT", "NAC")
(imatrix <- list(measure=imatrix[,1:6], condition=imatrix[,7]))
contrasts(condition_factor) <- matrix(c(-1,1,1, -1), ncol=2)
doubly.mod<-lm(cbind(bcst_nac$totPersErr,bcst_placebo$totPersErr,cantab_nac$CGT.Delay.aversion,cantab_placebo$CGT.Delay.aversion,cantab_nac$AST.Switching.cost..Mean..correct.,cantab_placebo$AST.Switching.cost..Mean..correct.,cantab_nac$AGN.Affective.response.bias..Mean.,cantab_placebo$AGN.Affective.response.bias..Mean.,aucs_NAC,aucs_placebo,fert_nac$FERTACCHA,fert_placebo$FERTACCHA)~1))
Manova(doubly.mod, imatrix=imatrix, type =3)
The result is this error: Error in Anova.III.mlm(mod, SSPE, error.df, idata, idesign, icontrasts, :
(list) object cannot be coerced to type 'double'
However, when I change imatrix back from a list to a matrix, I get this error response:
Error in do.call(cbind, imatrix) : second argument must be a list
I've based this off the example from the car::Anova package about doubly multivariate analyses. Please let me know if you can help, or if I can add anything to make this question clearer.

Resources