I have a mixed data set where I want to compute distances between observations with Gower in R (daisy function). When I compute distances between different number of observations, the distances seem to change, and I don't fully understand why that is. I assume it's some weights for the observations being used, but I can't find anything explaining this from the documentation

library(cluster) daisy(iris[1:3,], metric = "gower") #Distance between object 1&2 is 0.30 daisy(iris[1:4,], metric = "gower") #Now the distance between object 1&2 is 0.28

Does anyone know why this happens?




Browse other questions tagged or ask your own question.