Premium
Testing information redundancy in environmental monitoring networks
Author(s) -
Sarno Emma
Publication year - 2005
Publication title -
environmetrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.68
H-Index - 58
eISSN - 1099-095X
pISSN - 1180-4009
DOI - 10.1002/env.689
Subject(s) - autoregressive integrated moving average , redundancy (engineering) , computer science , estimator , probabilistic logic , data mining , autoregressive model , dimension (graph theory) , time series , econometrics , machine learning , statistics , artificial intelligence , mathematics , pure mathematics , operating system
There is a vast literature on optimal monitoring network designs. Most proposals arise from spatial statistics and they often overlook the time dimension of data collected by environmental detectors. In this work, we introduce a new concept of optimal design based on information redundancy, employing a time series approach. We model data generating processes corresponding to monitoring stations by ARIMA models and, subsequently, we measure the structural discrepancy between such models with the autoregressive distance estimator. Therefore, within a probabilistic framework we seek significant differences between models and, applying graph theory, we show how to elicit the optimal network design. Finally, a case study regarding the air monitoring network of Rome is illustrated to show how the whole procedure works in practice. Copyright © 2004 John Wiley & Sons, Ltd.