1. Bayesian P-splines and advanced computing in R for a changepoint analysis on spatio-temporal point processes
- Author
-
Fedele Pasquale Greco, E.M. Scott, Linda Altieri, Janine B. Illian, Daniela Cocchi, Altieri, L., Cocchi, D., Greco, F., Illian, J.B., Scott, E.M., University of St Andrews. School of Mathematics and Statistics, University of St Andrews. Scottish Oceans Institute, and University of St Andrews. Centre for Research into Ecological & Environmental Modelling
- Subjects
Parallel computing ,QA75 ,Statistics and Probability ,010504 meteorology & atmospheric sciences ,QA75 Electronic computers. Computer science ,Bayesian probability ,NDAS ,spatial effect ,P splines ,computer.software_genre ,01 natural sciences ,Point process ,010104 statistics & probability ,Bayesian P-spline ,Modelling and Simulation ,Bayesian hierarchical modeling ,QA Mathematics ,0101 mathematics ,QA ,0105 earth and related environmental sciences ,Mathematics ,Bayesian P-splines ,parallel computing ,log-Gaussian Cox processe ,Applied Mathematics ,Perspective (graphical) ,Probability and statistics ,Spatial effect ,Data set ,Spatio-temporal point processes ,Earthquake data ,Modeling and Simulation ,Changepoint analysis ,62M30 ,spatio-temporal point processe ,Data mining ,Statistics, Probability and Uncertainty ,Focus (optics) ,Log-Gaussian Cox processes ,62H11 ,changepoint analysi ,computer - Abstract
As regards authors Linda Altieri and Fedele Greco, the research work underlying this paper was partially funded by an FIRB 2012 [grant number RBFR12URQJ]; title: Statistical modelling of environmental phenomena: pollution, meteorology, health and their interactions) for research projects by the Italian Ministry of Education, Universities and Research. This work presents advanced computational aspects of a new method for changepoint detection on spatio-temporal point process data. We summarize the methodology, based on building a Bayesian hierarchical model for the data and declaring prior conjectures on the number and positions of the changepoints, and show how to take decisions regarding the acceptance of potential changepoints. The focus of this work is about choosing an approach that detects the correct changepoint and delivers smooth reliable estimates in a feasible computational time; we propose Bayesian P-splines as a suitable tool for managing spatial variation, both under a computational and a model fitting performance perspective. The main computational challenges are outlined and a solution involving parallel computing in R is proposed and tested on a simulation study. An application is also presented on a data set of seismic events in Italy over the last 20 years. Postprint
- Published
- 2016
- Full Text
- View/download PDF