Classifying hot water chemistry: Application of multivariate statistics

Author
Prihadi Sumintadireja, Dasapta Erwin Irawan, Rina Herdianita, Yuano Rezky, Prana Ugiana Gio, Anggita Agustin, and Ali Lukman
View Count
2624
License
Creative Commons CC BY 4.0
Abstract

The following paper is a try out on the application of multivariate analysis (regression tree, principal component analysis, and cluster analysis) for classifying hot water chemistry. The number of sample analysed was 416 from all over Indonesia. Regression tree technique has failed to read the data structure due to multi-collinearity effect therefore PCA and cluster analysis were applied. We used open source R statistical packages to do the calculation. Such technique classifies hot water samples into three major clusters: cluster 1 (pure hot water), cluster 2 (mixing water), and cluster 3 (cold-meteoric water). Similar clustering were also detected in the PCA plot. The statistical is able to detect the close and open geothermal system based on data structure. This robust method should be applied to more geothermal system with larger dataset to see its performance.

Classifying hot water chemistry: Application of multivariate statistics