Self-Organizing Maps for identification of zeolitic diagenesis patterns in closed hydrologic systems.

Zeolites are one of the most common authigenic silicate minerals found in paleo lacustrine environments on the Earth. The spatial and stratigraphic patterns of these minerals found in paleolakes are potentially diagnostic for their lacustrine origin. But, identifying predominant mineral paragenesis patterns/mineral assemblages in these environments are often difficult and complex. Therefore, in this exercise we use Self-Organizing Maps, a statistical learning (Machine Learning) method to look for interesting patterns in the lake bed mineralogy without prescribing any specific interpretation. We have complied published bulk mineralogy data from 13 saline-alkaline paleolakes from USA, Tanzania, Mexico and Greece. This work done in RStudio. Dataset named "ZeoliteDB.csv". Code used in this study named "SOM_Zeolite_code.Rmd". There is an additional code called "Radial_plot.Rmd" to plot data in a radial plot.