How to statistically analyze nano exposure measurement results: using an ARIMA time series approach |
| |
Authors: | Rinke H. Klein Entink Wouter Fransman Derk H. Brouwer |
| |
Affiliation: | (1) TNO, Research Group Quality and Safety, P.O. Box 360, 3700 AJ Zeist, The Netherlands |
| |
Abstract: | Measurement strategies for exposure to nano-sized particles differ from traditional integrated sampling methods for exposure assessment by the use of real-time instruments. The resulting measurement series is a time series, where typically the sequential measurements are not independent from each other but show a pattern of autocorrelation. This article addresses the statistical difficulties when analyzing real-time measurements for exposure assessment to manufactured nano objects. To account for autocorrelation patterns, Autoregressive Integrated Moving Average (ARIMA) models are proposed. A simulation study shows the pitfalls of using a standard t-test and the application of ARIMA models is illustrated with three real-data examples. Some practical suggestions for the data analysis of real-time exposure measurements conclude this article. |
| |
Keywords: | |
本文献已被 SpringerLink 等数据库收录! |
|