Generalizing resemblance coefficients to accommodate incomplete data

Large ecological data matrices may be incomplete for various reasons, preventing the use of standard multidimensional scaling (ordination) and cluster analysis packages. Although there exist a few resemblance functions that allow missing scores, is no theoretical background software support most distance similarity coefficients potentially applied in multivariate analysis. We provide general framework precise mathematical redefinition large set originally developed complete sets with presence-absence (binary) or ratio-scale variables. Included are which consider double absences abundance data. Potential problems these discussed, conclusion incompleteness would rarely if ever influence greatly interpretability ordinations classifications. An R function described Appendix represents link to R. also stand-alone WINDOWS application users other computer programs. The new will packages perform using wide variety even whatever reason.