How to Calculate Rolling Correlation in R?

In this article, we will discuss Rolling Correlation in R Programming Language.

Correlation is used to get the relationship between two variables.

• It will result in 1 if the correlation is positive.
• It will result in -1 if the correlation is negative.
• it will result in 0 if there is no correlation.

Rolling correlations are used to get the relationship between two-time series on a rolling window. We can calculate by using rollapply() function, This is available in the zoo package, So we have to load this package.

Syntax:

rollapply(data, width, FUN, by.column=TRUE)

where,

• data is the input dataframe.
• width is an integer that specifies the window width for the rolling correlation.
• FUN is the  function to be applied.
• by.column is used to specify whether to apply the function to each column separately.

We can get correlation using cor() function.

Syntax:

cor(column1,column2)

Example 1: R program to calculate rolling correlation for the dataframe.

# load the library
library(zoo)

# create dataframe with 3 columns
data = data.frame(day=1:15,
                   col1=c(35:49),
                   col2=c(33:47))

# display
print(data)

# get rolling correlation for col1 and 
# col2 with width 6
print(rollapply(data, width=6, function(x) cor(x[,2],x[,3]), 
                by.column=FALSE))

Output:

Example 2:

# load the library
library(zoo)

# create dataframe with 3 columns
data = data.frame(
                   col1=c(23,45,23,32,23),
                   col2=c(1,45,67,32,45))

# display
print(data)

# get rolling correlation for col1 and
# col2 with width 2
print(rollapply(data, width=2, function(x) cor(x[,1],x[,2]),
                by.column=FALSE))

Output: