27 September 2018 in Water Resources Research
Regional flood frequency analysis is commonly applied in situations where there exists insufficient data at a location for a reliable estimation of flood quantiles. We develop a Bayesian hierarchical modeling framework for a regional analysis of data from 203 large catchments in Norway with the generalized extreme value distribution as the underlying model. Generalized linear models on the parameters of the generalized extreme value distribution are able to incorporate location‐specific geographic and meteorological information and thereby accommodate these effects on the flood quantiles. A Bayesian model averaging component additionally assesses model uncertainty in the effect of the proposed covariates. The resulting regional model is seen to give substantially better predictive performance than the regional model currently used in Norway.
3 January 2018 in Statistics in Medicine
15 September 2021 in Journal of Marine Engineering & Technology
Thermal protection in marine electrical propulsion motors is commonly implemented by installing temperature sensors on the windings of the motor. An alarm is issued once the temperature reaches the alarm limit, while the motor shuts down once the trip limit is reached. Field experience shows that this protection scheme in some cases is insufficient, as the motor may already be damaged before reaching the trip limit. In this paper, we develop a machine learning algorithm to predict overheating, based on past data collected from a class of identical vessels. All methods were implemented to comply with real-time requirements of the on-board protective systems with minimal need for memory and computational power. Our two-stage overheating detection algorithm first predicts the temperature in a normal state using linear regression fitted to regular operation motor performance measurements, with exponentially smoothed predictors accounting for time dynamics. Then it identifies and monitors temperature deviations between the observed and predicted temperatures using an adaptive cumulative sum (CUSUM) procedure. Using data from a real fault case, the monitor alerts between 60 to 90 min before failure occurs, and it is able to detect the emerging fault at temperatures below the current alarm limits.
24 April 2020 in Multivariate Behavioral Research
Psychiatric syndromes in dementia are often derived from the Neuropsychiatric Inventory (NPI) using principal component analysis (PCA). The validity of this statistical approach can be questioned, since the excessive proportion of zeros and skewness of NPI items may distort the estimated relations between the items. We propose a novel version of PCA, ZIBP-PCA, where a zero-inflated bivariate Poisson (ZIBP) distribution models the pairwise covariance between the NPI items. We compared the performance of the method to classical PCA under zero-inflation using simulations, and in two dementia-cohorts (N = 830, N = 1349). Simulations showed that component loadings from PCA were biased due to zero-inflation, while the loadings of ZIBP-PCA remained unaffected. ZIBP-PCA obtained a simpler component structure of “psychosis,” “mood” and “agitation” in both dementia-cohorts, compared to PCA. The principal components from ZIBP-PCA had component loadings as follows: First, the component interpreted as “psychosis” was loaded by the items delusions and hallucinations. Second, the “mood” component was loaded by depression and anxiety. Finally, the “agitation” component was loaded by irritability and aggression. In conclusion, PCA is not equipped to handle zero-inflation. Using the NPI, PCA fails to identify components with a valid interpretation, while ZIBP-PCA estimates simple and interpretable components to characterize the psychopathology of dementia.
1 February 2017 in Scandinavian Journal of Statistics
10 January 2022 in Scandinavian Journal of Statistics