Uncertainty evaluation in time-dependent measurements. (15th June 2022)

• Record Type: Journal Article
• Authors: Silva, Maria Almeida; Amado, Conceição; Ribeiro, Álvaro; Loureiro, Dália
• Is Part Of: Measurement. Volume 196(2022)
• Publication Date: 2022-06-15
• Subjects: Model-based approach -- Resampling -- Time-dependent measurements -- Uncertainty evaluation; Weights and measures -- Periodicals; Measurement -- Periodicals; Measurement; Weights and measures; Periodicals; 530.8
• DOI: 10.1016/j.measurement.2022.111196
• Languages: English
• ISSNs: 0263-2241

Abstract

Abstract: Data quality plays a strong role in modelling and is fundamental to ensure that obtained analyses and results are reliable. Data uncertainty is related to the measurement's imperfections that will cause an error in the obtained numerical value. In this way, data quality could be measured through uncertainty. Although there are well-established methods for uncertainty evaluation in independent measurements, there is a gap of methods in the case of time-dependent measurements. This study aims to propose a general methodology to evaluate the uncertainty of time-dependent measurements. Fitting an appropriate model to the measurements and then resampling and refitting a model, the uncertainty estimation of each data point is possible. The approach was validated using measurements obtained from typical time series models and compared with theoretical values. Further, real data from a water system is also used to illustrate the method' capacity for uncertainty estimation of single time measurements. Highlights: An approach to evaluate the time-dependent measurements' uncertainty is proposed. The approach is based on fitting a model and then resampling and refitting the model. The approach was validated using typical time series models and theoretical values. Real data from a water system is used to illustrate the capacity of the methodology.