Pairwise-complete correlation considered dangerous
http://bwlewis.github.io › missingThe cov and cor functions in the R programming language include several options ... (x = matrix(c(-2,-1,0,1,2,1.5,2,0,1,2,NA,NA,0,1,2),5))
Correlation in R - DataScience Made Simple
www.datasciencemadesimple.com › correlation-in-rCorrelation of vector in R with NA: Note: Correlation in R cannot be calculated if values has NA. So use = “complete.obs” neglects NAs while calculating correlation coefficient in R # correlation in R : handling NA x <- c(0,1,1,2,3,5,8,13,21,NA) y <- log(x+1) cor(x,y,use = "complete.obs") so the output will be
cor function - RDocumentation
www.rdocumentation.org › versions › 3Note that "spearman" basically computes cor (R (x), R (y)) (or cov (., .)) where R (u) := rank (u, na.last = "keep"). In the case of missing values, the ranks are calculated depending on the value of use, either based on complete observations, or based on pairwise completeness with reranking for each pair.