R - Append Item Iteratively To Matrix?
Entering edit mode
11.8 years ago


I have a table and a vector, e.g.,

Table =

    A    B    C
s1  0    1    2
s2  1    2    1
s3  2    0    1


vec <- c(0.01, 0.087, 0.34)

I wish to iterate over the table to isolate pairs of rows and then use the vector to fit a 1st order linear regression model (Y~A*B). My problem is that I wish to initiate a matrix outside the main for loop and append the variable 'p_value' to that matrix after each iteration.

The code to isolate the p-values is given below:

for (rows in 1:nrow(Table)) { 
    li_row <- Table[rows,] # Extract a single row
    row <- unlist(li_row) # Convert to vector

    for (other in 1:nrow(Table[2, ])) {
        li_row_2 <- Table[other,] # Extract a different single row
        row_2 <- unlist(li_row_2)
        lin_reg <- lm(vec ~ row*row2) # Call 1st order linear model
        summ <- summary(lin_reg) # Store results summary
        f_stat <- summ$fstatistic # Extract F-statistic data
        p_value <- fstat[1] # Extract p-value [**The value to append to matrix**]

In short, upon each iteration, I wish to add the 'p-value' to a matrix. I know a priori how big the matrix will be and also know what labels to attach to rows and columns.

S ;-)

r matrix • 32k views
Entering edit mode

What are you expecting 1:nrow(Table[2, ]) to evaluate to? This is the same as asking for 1:NULL because Table[2, ] is a vector and has nrow of NULL

Entering edit mode
11.8 years ago

Quick and dirty answer; concatenate all your p-values into a single vector as you progress

vec <- c(vec, p_value)

and at the end, convert to a matrix

matrix(c(0, 1, 2, 1, 2, 1, 2, 0, 1), nrow = 3)

Better answer; don't iteratively append, but rather map a function down the rows of your table and use R's built-in vectorisation to create a matrix.

A toy example below applies an anonymous function to each row y of matrix x, returning a new matrix. The function adds the row y to another row, also from x, but chosen randomly (without replacement).

x <- matrix(c(1, 2, 3, 10, 10, 10, 4, 5, 6), nrow = 3)
apply(x, 2, function(y) y + x[sample(1:nrow(x), 1), TRUE])

Login before adding your answer.

Traffic: 746 users visited in the last hour
Help About
Access RSS

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6