Returns a function that takes x and y and returns a formatted output to describe the correlation of x and y
Usage
FormatCorrelation(
method = "pearson",
n.format = NULL,
coeff.format = "%.2f",
p.format = "%.2g",
slope.format = NULL,
rmsd.format = NULL
)Arguments
- method
how to compute correlation coefficients (can be pearson, spearman or kendall)
- n.format
format string for the number of data points (see sprintf); can be NULL (don't output the number of data points)
- coeff.format
format string for the correlation coefficient (see sprintf); can be NULL (don't output the correlation coefficient)
- p.format
format string for the P value (see sprintf); can be NULL (don't output the P value)
- slope.format
format string for the slope (see sprintf); can be NULL (don't output the slope)
- rmsd.format
format string for the root mean square deviation (see sprintf); can be NULL (don't output the rmsd)
Details
Use this for the correlation parameter of PlotScatter
The slope is computed via a principal component analysis and *not* by linear regression
Examples
set.seed(42)
data <- data.frame(u=runif(500)) # generate some correlated data
data$x <- rnorm(500,mean=data$u)
data$y <- rnorm(500,mean=data$u)
fun <- FormatCorrelation()
fun(data$x,data$y)
#> [1] "R=0.10\np=0.032"
fun <- FormatCorrelation(method="spearman",p.format="%.4g")
fun(data$x,data$y)
#> [1] "ρ=0.09\np=0.05028"