is there are any more efficient/faster way to compare two matrices (column by columns) and to compute p-values using t-test for no difference in means (eventually switching to the chisq.test when necessary)?
Here is my solution:
## generate fake data (e.g., from treatment and control data)
z0 <- matrix(rnorm(100),10,10)
z1 <- matrix(rnorm(100, mean=1.1, sd=2),10,10)
## function to compare columns (bloody for loop)
compare.matrix <- function(z0, z1){
pval <- numeric(ncol(z0)) ## initialize
for(i in 1:ncol(z0)){ ## compare columns
pval[i] <- t.test(z1[, i], z0[, i])$p.value
## if var is categorical, switch test type
if ( length(unique(z1[,i]))==2){
index <- c(rep(0, nrow(z0)), rep(1, nrow(z1)))
xx <- c(z0[,i], z1[,i])
pval[i] <- chisq.test(table(xx, index), simulate.p.value=TRUE)$p.value
}
}
return(pval)
}
compare.matrix(z0, z1)
Here's one way using dplyr. It would probably be better to combine the first three lines into a single step if you've got large matrices, but I separated them for clarity. I think the chi-squared case would be a fairly simple extension.
z0_melt = melt(z0, value.name='z0')[,c('Var2','z0')]
z1_melt = melt(z1, value.name='z1')[,c('Var2','z1')]
all_df = merge(z0_melt, z1_melt)
library(dplyr)
all_df %>%
group_by(Var2) %>%
summarize(p = t.test(z0, z1)$p.value)
Related
I've recently been interested in trying to develop a for-loop that would be able to run multiple generalized additive models and then produce results in a table that ranks them based on AIC, p-value of each smooth in the model, deviance explained of the overall model, etc.
I found this related question in stack overflow which is basically what I want and was able to run this well for gam() instead of gamm(), however I want to expand this to include multiple independent variables in the model, not just 1.
Ideally, the models would run all possible combinations of independent variables against the dependent variable, and it would test combinations anywhere from 1 independent variable in the model, up to all of the possible covariates in "d_pred" in the model.
I have attempted to do this so far by starting out small and finding all possible combinations of 2 independent variables (df_combinations2), which results in a list of data frames. Then I adjusted the rest of the code to run the for loop such that each iteration will run a different combination of the two variables:
library(mgcv)
## Example data
set.seed(0)
dat <- gamSim(1,n=200,scale=2)
set.seed(1)
dat2 <- gamSim(1,n=200,scale=2)
names(dat2)[1:5] <- c("y1", paste0("x", 4:7))
d <- cbind(dat[, 1:5], dat2[, 1:5])
d_resp <- d[ c("y", "y1")]
d_pred <- d[, !(colnames(d) %in% c("y", "y1"))]
df_combinations2 <- lapply(1:(ncol(combn(1:ncol(d_pred), m = 2))),
function(y) d_pred[, combn(1:ncol(d_pred), m = 2)[,y]])
## create a "matrix" list of dimensions i x j
results_m2 <-lapply(1:length(df_combinations2), matrix, data= NA, nrow=ncol(d_resp), ncol=2)
## for-loop
for(k in 1:length(df_combinations2)){
for(i in 1:ncol(d_resp)){
for(j in 1:ncol(df_combinations2[[k]])){
results_m2[i, j][[1]] <- gam(d_resp[, i] ~ s(df_combinations2[[k]][,1])+s(df_combinations2[[k]][,2]))
}
}}
However, after running the for-loop I get the error "Error in all.vars1(gp$fake.formula[-2]) : can't handle [[ in formula".
Anyone know why I am getting this error/ how to fix it?
Any insight is much appreciated. Thanks!
Personally, I would create a data.table() containing all combinations of target variables and combinations of predictors and loop through all rows. See below.
library(data.table)
library(dplyr)
# Example data
set.seed(0)
dat <- gamSim(1,n=200,scale=2)
set.seed(1)
dat2 <- gamSim(1,n=200,scale=2)
names(dat2)[1:5] <- c("y1", paste0("x", 4:7))
d <- cbind(dat[, 1:5], dat2[, 1:5])
#select names of targets and predictors
targets <- c("y", "y1")
predictors <- colnames(d)[!colnames(d) %in% targets]
#create all combinations of predictors
predictor_combinations <- lapply(1:length(predictors), FUN = function(x){
#create combination
combination <- combn(predictors, m = x) |> as.data.table()
#add s() to all for gam
combination <- sapply(combination, FUN = function(y) paste0("s(", y, ")")) |> as.data.table()
#collapse
combination <- summarize_all(combination, .funs = paste0, collapse = "+")
#unlist
combination <- unlist(combination)
#remove names
names(combination) <- NULL
#return
return(combination)
})
#merge combinations of predictors as vector
predictor_combinations <- do.call(c, predictor_combinations)
#create folder to save results to
if(!dir.exists("dev")){
dir.create("dev")
}
if(!dir.exists("dev/models")){
dir.create("dev/models")
}
#create and save hypergrid (all combinations of targets and predictors combinations)
if(!file.exists("dev/hypergrid.csv")){
#create hypergrid and save to dev
hypergrid <- expand.grid(target = targets, predictors = predictor_combinations) |> as.data.table()
#add identifier
hypergrid[, model := paste0("model", 1:nrow(hypergrid))]
#save to dev
fwrite(hypergrid, file = "dev/hypergrid.csv")
} else{
#if file exists read
hypergrid <- fread("dev/hypergrid.csv")
}
#loop through hypergrid, create GAM models
#progressbar
pb <- txtProgressBar(min = 1, max = nrow(hypergrid), style = 3)
for(i in 1:nrow(hypergrid)){
#update progressbar
setTxtProgressBar(pb, i)
#select target
target <- hypergrid[i,]$target
#select predictors
predictors <- hypergrid[i,]$predictors
#create formula
gam.formula <- as.formula(paste0(target, "~", predictors))
#run gam
gam.model <- gam(gam.formula, data = d)
#save gam model do dev/model
saveRDS(gam.model, file = paste0("dev/models/", hypergrid[i,]$model, ".RDS"))
}
#example where you extract model performances
for(i in 1:nrow(hypergrid)){
#read the right model
rel.model <- readRDS(paste0("dev/models/", hypergrid[i,]$model, ".RDS"))
#extract model performance, add to hypergrid
hypergrid[i, R2 := summary(rel.model)[["r.sq"]]]
}
#arrange hypergrid on target and r2
hypergrid <- dplyr::arrange(hypergrid, hypergrid$target, desc(hypergrid$R2))
Which would give
head(hypergrid)
target predictors model R2
1: y s(x0)+s(x1)+s(x2)+s(x4)+s(x5) model319 0.6957242
2: y s(x0)+s(x1)+s(x2)+s(x3)+s(x4)+s(x5) model423 0.6953753
3: y s(x0)+s(x1)+s(x2)+s(x4)+s(x5)+s(x7) model437 0.6942054
4: y s(x0)+s(x1)+s(x2)+s(x5) model175 0.6941025
5: y s(x0)+s(x1)+s(x2)+s(x4)+s(x5)+s(x6) model435 0.6940569
6: y s(x0)+s(x1)+s(x2)+s(x3)+s(x4)+s(x5)+s(x7) model481 0.6939756
All models are saved to a folder with an identifier (for if you want to use the model or extract more information from the model).
Notably, p-hacking comes to mind using this appraoch and I would be careful by conducting your analysis like this.
I have two datasets with abundance data from groups of different species. Columns are species and rows are sites. The sites (rows) are identical between the two datasets and what i am trying to do is to correlate the columns of the first dataset to the columns of the second dataset in order to see if there is a positive or a negative correlation.
library(Hmisc)
rcorr(otu.table.filter$sp1,new6$spA, type="spearman"))$P
rcorr(otu.table.filter$sp1,new6$spA, type="spearman"))$r
the first will give me the p value of the relation between sp1 and spA and the second the r value
I initially created a loop that allowed me to check all species of the first dataframe with a single column of the second dataframe. Needless to say if I was to make this work I would have to repeat the process a few hundred times.
My simple loop for one column of df1(new6) against all columns of df2(otu.table.filter)
pvalues = list()
for(i in 1:ncol(otu.table.filter)) {
pvalues[[i]] <-(rcorr(otu.table.filter[ , i], new6$Total, type="spearman"))$P
}
rvalues = list()
for(i in 1:ncol(otu.table.filter)) {
rvalues[[i]] <-(rcorr(otu.table.filter[ , i], new6$Total, type="spearman"))$r
}
p<-NULL
for(i in 1:length(pvalues)){
tmp <-print(pvalues[[i]][2])
p <- rbind(p, tmp)
}
r<-NULL
for(i in 1:length(rvalues)){
tmp <-print(rvalues[[i]][2])
r <- rbind(r, tmp)
}
fdr<-as.matrix(p.adjust(p, method = "fdr", n = length(p)))
sprman<-cbind(r,p,fdr)
and using the above as a starting point I tried to create a nested loop that each time would examine a column of df1 vs all columns of df2 and then it would proceed to the second column of df1 against all columns of df2 etc etc
but here i am a bit lost and i could not find an answer for a solution in r
I would assume that the pvalues output should be a list of
pvalues[[i]][[j]]
and similarly the rvalues output
rvalues[[i]][[j]]
but I am a bit lost and I dont know how to do that as I tried
pvalues = list()
rvalues = list()
for (j in 1:7){
for(i in 1:ncol(otu.table.filter)) {
pvalues[[i]][[j]] <-(rcorr(otu.table.filter[ , i], new7[,j], type="spearman"))$P
}
for(i in 1:ncol(otu.table.filter)) {
rvalues[[i]][[j]] <-(rcorr(otu.table.filter[ , i], new7[,j], type="spearman"))$r
}
}
but I cannot make it work cause I am not sure how to direct the output in the lists and then i would also appreciate if someone could help me with the next part which would be to extract for each comparison the p and r value and apply the fdr function (similar to what i did with my simple loop)
here is a subset of my two dataframes
Here a small demo. Let's assume two matrices x and y with a sample size n. Then correlation and approximate p-values can be estimated as:
n <- 100
x <- matrix(rnorm(10 * n), nrow = n)
y <- matrix(rnorm(5 * n), nrow = n)
## correlation matrix
r <- cor(x, y, method = "spearman")
## p-values
pval <- function(r, n) 2 * (1 - pt(abs(r)/sqrt((1 - r^2)/(n - 2)), n - 2))
pval(r, n)
## for comparison
cor.test(x[,1], y[,1], method = "spearman", exact = FALSE)
More details can be found here: https://stats.stackexchange.com/questions/312216/spearman-correlation-significancy-test
Edit
And finally a loop with cor.test:
## for comparison
p <- matrix(NA, nrow = ncol(x), ncol=ncol(y))
for (i in 1:ncol(x)) {
for (j in 1:ncol(y)) {
p[i, j] <- cor.test(x[,i], y[,j], method = "spearman")$p.value
}
}
p
The values differ a somewhat, because the first uses the t-approximation then the second the "exact AS 89 algorithm" of cor.test.
I have a dataset with x number of columns, consisting of groups of test results, for example test1_1, test1_2 etc. Each set of tests has a different number of test results associated with it so the actual numbers aren't the same across each test. The final column is my target variable. I'm looking to establish which tests are correlated with the target variable, but I also want to create datasets for each set of tests. I'm also going to be plotting correlation plots of each test against the target variable. I suspect I could probably achieve all of this in a few lines of code within a for/while loop, however, I'm not sure where to begin.
Using lapply this could be achieved like so:
library(dplyr)
library(corrplot)
set.seed(42)
dataset <- data.frame(
test1_1 = runif(20),
test1_2 = runif(20),
test2_1 = runif(20),
test2_2 = runif(20),
Target = runif(20)
)
test_cols <- gsub("_\\d+$", "", names(dataset))
test_cols <- test_cols[grepl("^test", test_cols)]
test_cols <- unique(test_cols)
test_cols <- setNames(test_cols, test_cols)
test_fun <- function(x, test) {
x <- x %>%
select((starts_with(test)) | matches("Target"))
cor(x)
}
cor_test <- lapply(test_cols, test_fun, x = dataset)
cplot <- lapply(cor_test, corrplot)
This is similar to #stefan's answer using split.default to split the columns by pattern in the column names.
tmp <- dplyr::select(dataset, -Target)
list_plot <- lapply(split.default(tmp, sub('_.*', '', names(tmp))), function(x) {
corrplot::corrplot(cor(cbind(x, Target = dataset$Target)))
})
I intend to find Pearson correlation coefficient from multi-dim data to one numeric vector in R. Basically, I am expecting to get a correlation matrix by using the Pearson method, want to keep the rows (a.k.a, features for each column) in multi-dim data by using certain correlation coefficient as threshold.However, I tentatively tried some R implementation to do that but didn't get correct correlation matrix though. How can I get this one? can anyone point me out how to make this happen easily in R? any thought?
reproducible example
persons_df <- data.frame(person1=sample(1:20,10, replace = FALSE),
person2=as.factor(sample(10)),
person3=sample(1:25,10, replace = FALSE),
person4=sample(1:30,10, replace = FALSE),
person5=as.factor(sample(10)),
person6=as.factor(sample(10)))
row.names(persons_df) <-letters[1:10]
in persons_df, different features in row-wise and different persons in column-wise are given.
I have also age_df which has age of each person.
age_df <- data.frame(personID= colnames(persons_df),
age=sample(1:50, 6 , replace = FALSE))
my initial attempt:
pearson_corr <- function(df1, df2, verbose=FALSE){
stopifnot(ncol(df1)==nrow(df2))
res <- as.data.frame()
lapply(colnames(df1), function(x){
lapply(x, rownames(y){
if(colnames(x) %in% rownames(df2)){
cor_mat <- stats::cor(y, df2$age, method = "pearson")
ncor <- ncol(cor_mat)
cmatt <- col(cor_mat)
ord <- order(-cmat, cor_mat, decreasing = TRUE)- (ncor*cmatt - ncor)
colnames(ord) <- colnames(cor_mat)
res <- cbind(ID=c(cold(ord), ID2=c(ord)))
res <- as.data.frame(cbind(out, cor=cor_mat[res]))
res <- cbind(res, cor=cor_mat[out])
}
})
})
return(final_df)
}
but above code didn't return correct correlation matrix. what I want to do how each features of the certain person is correlated with his age. Is there any efficient way to make this happen? any idea?
goal:
basically, I want to keep the features which show a high correlation with age. I don't have a better idea to do this in R. Can anyone point me out how to get his done easily and efficiently in R? thanks
mylist = do.call(rbind,
apply(persons_df, 1, function(x){
temp = cor.test(age_df$age, as.numeric(x))
data.frame(t = temp$statistic, p = temp$p.value)
}))
mylist
# t p
#a -1.060264 3.488012e-01
#b -2.292612 8.361623e-02
#c -16.785311 7.382895e-05
#d -1.362776 2.446304e-01
#e -1.922296 1.269356e-01
#f -4.671259 9.509393e-03
#g -3.719296 2.048710e-02
#h -2.684663 5.496171e-02
#i -15.814635 9.341701e-05
#j -2.423014 7.252635e-02
Then use mylist to filter out what values you don't want.
I have a customized function (psup2) that gets data from a data frame and returns a result. The problem is that it takes a while since I am using a "for" loop that runs for every row and column.
Input:
I have a table that contains the ages (table_costumers), an n*m matrix of different terms, and two different mortality tables (for males and females).
The mortality tables i´m using contains one column for ages and another one for its corresponding survival probabilities.
Output:
I want to create a separate dataframe with the same size as that of the term table. The function will take the data from the different mortality tables (depending on the gender) and then apply the function above (psup2) taking the ages from the table X and the terms from the matrix terms.
Up to now I managed to create a very inefficient way to do this...but hopefully by using one of the functions from the apply family this could get faster.
The following code shows the idea of what I am trying to do:
#Function
psup2 <- function(x, age, term) {
P1 = 1
for (i in 1:term) {
P <- x[age + i, 2]
P1 <- P1*P
}
return(P1)
}
#Inputs
terms <- data.frame(V1 = c(1,2,3), V2 = c(1,3,4), V2 = c(2,3,4))
male<- data.frame(age = c(0,1,2,3,4,5), probability = c(0.9981,0.9979,0.9978,.994,.992,.99))
female <- data.frame(age = c(0,1,2,3,4,5), probability = c(0.9983,0.998,0.9979,.9970,.9964,.9950))
table_customers <- data.frame(id = c(1,2,3), age = c(0,0,0), gender = c(1,2,1))
#Loop
output <- data.frame(matrix(NA, nrow = 3, ncol = 0))
for (i in 1:3) {
for (j in 1:3) {
prob <- ifelse(table_customers[j, 3] == 1,
psup2(male, as.numeric(table_customers[j, 2]), as.numeric(terms[j,i])),
psup2(female, as.numeric(table_customers[j, 2]), as.numeric(terms[j,i])))
output[j, i] <- prob
}
}
your psup function can be simplified into:
psup2 <- function(x, age, term) { prod(x$probability[age+(1:term)]) }
So actually, we won't use it, we'll use the formula directly.
We'll put your male and female df next to each other, so we can use the value of the gender column to choose one or another.
mf <- merge(male,female,by="age") # assuming you have the same ages on both sides
input_df <- cbind(table_customers,terms)
output <- t(apply(input_df,1,function(x){sapply(1:3,function(i){prod(mf[x["age"]+(1:x[3+i]),x["gender"]+1])})}))
And that's it :)
The sapply function is used to loop on the columns of terms.
x["age"]+(1:x[3+i]) are the indices of the rows you want to multiply
x["gender"]+1 is the relevant column of the mf data.frame