diff --git a/Census2022-CER_regP_48_CLUSTER_std-1-5-15.R b/Census2022-CER_regP_48_CLUSTER_std-1-5-15.R index 5a2fa6f4eb475880c19ccdfa3206c70748341275..dc9fd6859eade3d2ba7af366a45c813333ad5d56 100755 --- a/Census2022-CER_regP_48_CLUSTER_std-1-5-15.R +++ b/Census2022-CER_regP_48_CLUSTER_std-1-5-15.R @@ -125,6 +125,7 @@ stddata = mydatam[,177:224] stddata = na.omit(stddata) stddata = scale(stddata) ## make mean 0, stdev 1 +## k means clustering ## partitioning by a plot of the within groups sum of squares (WSS) by number of clusters extracted (looking for a bend in the plot, http://www.statmethods.net/advstats/cluster.html) ## wss <- (nrow(mydata)-1)*sum(apply(mydata,2,var)) wss <- (nrow(stddata)-1)*sum(apply(stddata,2,var))