Statistical model selection between elastic and Newtonian viscous matrix models for the microboudin palaeopiezometer

• Main Authors: Tarojiro Matsumura, Tatsu Kuwatani, Toshiaki Masuda
• Format: Article
• Language: English
• Published: SpringerOpen 2017-06-01
• Series: Earth, Planets and Space
• Subjects: Microboudinage structure; AIC; Cross-validation; Statistical model selection; Elastic matrix model; Newtonian viscous matrix model
• Online Access: http://link.springer.com/article/10.1186/s40623-017-0669-4

Abstract We carried out statistical evaluations of two probability density functions for the microboudin palaeopiezometer using the Akaike information criterion (AIC) and the cross-validation (CV) technique. In terms of the relevant stress-transfer model, these functions are defined as the elastic matrix model and the Newtonian viscous matrix model, respectively. The AIC and CV techniques enable us to evaluate the relative quality of both models when applied to nine data sets collected from metachert samples containing tourmaline grains in a quartz matrix, collected from the East Pilbara Terrane, Western Australia. The results show that the elastic matrix model is the more appropriate probability density function for analysis of fracturing of tourmaline grains in a quartz matrix. This statistical evaluation shows the validity of the elastic matrix model for the microboudin palaeopiezometer when analysing such data sets. Graphical abstract .