Dfyjkt

This is the code of my rolling regression:

# betas from rolling regression will be in betas_rolling_250 matrix
fixed.window<-30
models<-c("A~B+E","A~B+C","A~B+D","A~B+C","A~B+F","A~B+B","A~B+F","A~B+C")
betas_rolling_250<-matrix(,nrow=length(models),ncol=250-fixed.window)
rownames(betas_rolling_250)<-models;

#Sample Matrix

sample_matrix_250<-matrix(rexp(10, rate=.1),nrow=250, ncol=6)
colnames(sample_matrix_250)<-c("A","B","C","D","E","F")

for(i in 1:(length(sample_matrix_250[,1])-fixed.window))
 for(k in 1:length(rownames(betas_rolling_250)))
 coefs<-lm(rownames(betas_rolling_250)[k],data=as.data.frame(sample_matrix_250[(221-i):(251-i),]))$coefficients[3]
 betas_rolling_250[k,i]<-coefs 

First question:

Is it possible to implement this by using lapply function?

I am asking this beacause I will have to do this 4 times. Because I have more 4 differents size dataframes: (sample_matrix_220, sample_matrix_200and sample_matrix_180). I dont want my code to be huge.

Ive been thinking in using list. Thats why I though about lapply function.

Is it possible?

Consider generalizing your setup by passing the only changed variable, matrix column size, as input parameter. Then, past list of numeric values into an lapply call.

Below replaces any 250 to a variable (adjust as needed). The static models and fixed.window should be assigned once outside the function.

get_coeffs <- function(mat_size) 
 betas_rolling <- matrix(,nrow=length(models), ncol=mat_size-fixed.window)
 rownames(betas_rolling) <- models

 sample_matrix <- matrix(rexp(10, rate=.1), nrow=mat_size, ncol=6)
 colnames(sample_matrix) <- LETTERS[1:6]

 for(i in 1:(length(sample_matrix[,1])-fixed.window))
 for(k in 1:length(rownames(betas_rolling)))
 coefs <- lm(rownames(betas_rolling)[k],
 as.data.frame(sample_matrix[(mat_size-fixed.window)-i):(mat_size+1-i),]))$coefficients[3]
 betas_rolling[k,i] <- coefs 
 
 

 return(betas_rolling)


matrix_list <- lapply(c(180, 200, 220, 250), get_coeffs)