Bayesian machine learning ensemble approach to quantify model uncertainty in predicting groundwater storage change

2021 ◽  
Vol 769 ◽  
pp. 144715
Author(s):  
Jina Yin ◽  
Josué Medellín-Azuara ◽  
Alvar Escriva-Bou ◽  
Zhu Liu
Keyword(s):  
Machine Learning ◽  
Model Uncertainty ◽  
Groundwater Storage ◽  
Ensemble Approach ◽  
Bayesian Machine Learning ◽  
Storage Change

Precipitation reconstruction from climate-sensitive lithologies using Bayesian machine learning

2021 ◽  
Vol 139 ◽  
pp. 105002
Author(s):  
Rohitash Chandra ◽  
Sally Cripps ◽  
Nathaniel Butterworth ◽  
R. Dietmar Muller
Keyword(s):  
Machine Learning ◽  
Precipitation Reconstruction ◽  
Bayesian Machine Learning

Bayesian Machine Learning: EEG\/MEG signal processing measurements

2016 ◽  
Vol 33 (1) ◽  
pp. 14-36 ◽  
Author(s):  
Wei Wu ◽  
Srikantan Nagarajan ◽  
Zhe Chen
Keyword(s):  
Machine Learning ◽  
Signal Processing ◽  
Bayesian Machine Learning

Estimating the water balance and uncertainty bounds in a highly groundwater-dependent and data-scarce areas: An example for the upper Citarum basin

2021 ◽  
Author(s):  
Steven Reinaldo Rusli ◽  
Albrecht Weerts ◽  
Victor Bense
Keyword(s):  
Water Balance ◽  
Water Storage ◽  
Total Water ◽  
Groundwater Abstraction ◽  
Water Balance Components ◽  
Groundwater Storage ◽  
Actual Evaporation ◽  
Basin Scale ◽  
Basin Water ◽  
Storage Change

<p>In this study, we estimate the water balance components of a highly groundwater-dependent and hydrological data-scarce basin of the upper reaches of the Citarum river in West Java, Indonesia. Firstly, we estimate the groundwater abstraction volumes based on population size and a review of literature (0.57mm/day). Estimates of other components like rainfall, actual evaporation, discharge, and total water storage changes are derived from global datasets and are simulated using a distributed hydrological wflow_sbm model which yields additional estimates of discharge, actual evaporation, and total water storage change. We compare each basin water balance estimate as well as quantify the uncertainty of some of the components using the Extended Triple Collocation (ETC) method.</p><p>The ETC application on four different rainfall estimates suggests a preference of using the CHIRPS product as the input to the water balance components estimates as it delivers the highest r<sup>2</sup>  and the lowest RMSE compared to three other sources. From the different data sources and results of the distributed hydrological modeling using CHIRPS as rainfall forcing, we estimate a positive groundwater storage change between 0.12 mm/day - 0.60 mm/day. These results are in agreement with groundwater storage change estimates based upon GRACE gravimetric satellite data, averaged at 0.25 mm/day. The positive groundwater storage change suggests sufficient groundwater recharge occurs compensating for groundwater abstraction. This conclusion seems in agreement with the observation since 2005, although measured in different magnitudes. To validate and narrow the estimated ranges of the basin water storage changes, a devoted groundwater model is necessary to be developed. The result shall also aid in assessing the current and future basin-scale groundwater level changes to support operational water management and policy in the Upper Citarum basin.</p>

Optimization of Joint Equalization of High-Speed Signals using Bayesian Machine Learning

2021 ◽  
Author(s):  
N. Dikhaminjia ◽  
G. Tsintsadze ◽  
Z. Kiguradze ◽  
J. He ◽  
M. Tsiklauri ◽  
...  
Keyword(s):  
Machine Learning ◽  
High Speed ◽  
Bayesian Machine Learning

scholarly journals Evaluation of Groundwater Storage Variations in Northern China Using GRACE Data

Geofluids ◽  
2017 ◽  
Vol 2017 ◽  
pp. 1-13 ◽  
Author(s):  
Wenjie Yin ◽  
Litang Hu ◽  
Jiu Jimmy Jiao
Keyword(s):  
Land Surface ◽  
Snow Water Equivalent ◽  
Groundwater Resources ◽  
Dynamic Change ◽  
Northern China ◽  
Groundwater Storage ◽  
Grace Data ◽  
Depletion Rate ◽  
Terrestrial Water ◽  
Storage Change

Dynamic change of groundwater storage is one of the most important topics in the sustainable management of groundwater resources. Groundwater storage variations are firstly isolated from the terrestrial water storage change using the Global Land Data Assimilation System (GLDAS). Two datasets are used: (1) annual groundwater resources and (2) groundwater storage changes estimated from point-based groundwater level data in observation wells. Results show that the match between the GRACE-derived groundwater storage variations and annual water resources variation is not good in six river basins of Northern China. However, it is relatively good between yearly GRACE-derived groundwater storage data and groundwater storage change dataset in Huang-Huai-Hai Plain and the Song-Liao Plain. The mean annual depletion rate of groundwater storage in the Northern China was approximately 1.70 billion m3 yr−1 from 2003 to 2012. In terms of provinces, the yearly depletion rate is higher in Jing-Jin-Ji (Beijing, Tianjin, and Hebei province) and lowest in Henan province from 2003 to 2012, with the rate of 0.70 and 0.21 cm yr−1 Equivalent Water Height (EWH), respectively. Different land surface models suggest that the patterns from different models almost remain the same, and soil moisture variations are generally bigger than snow water equivalent variations.

scholarly journals Forecasting the Suitability of Online Mental Health Information for Effective Self-Care Developing Machine Learning Classifiers Using Natural Language Features

2021 ◽  
Vol 18 (19) ◽  
pp. 10048
Author(s):  
Meng Ji ◽  
Wenxiu Xie ◽  
Riliu Huang ◽  
Xiaobo Qian
Keyword(s):  
Mental Health ◽  
Machine Learning ◽  
Health Information ◽  
Self Care ◽  
Health Resources ◽  
Health Issues ◽  
Mental Health Information ◽  
Learning Classifier ◽  
Mental Health Issues ◽  
Bayesian Machine Learning

Background: Online mental health information represents important resources for people living with mental health issues. Suitability of mental health information for effective self-care remains understudied, despite the increasing needs for more actionable mental health resources, especially among young people. Objective: We aimed to develop Bayesian machine learning classifiers as data-based decision aids for the assessment of the actionability of credible mental health information for people with mental health issues and diseases. Methods: We collected and classified creditable online health information on mental health issues into generic mental health (GEN) information and patient-specific (PAS) mental health information. GEN and PAS were both patient-oriented health resources developed by health authorities of mental health and public health promotion. GENs were non-classified online health information without indication of targeted readerships; PASs were developed purposefully for specific populations (young, elderly people, pregnant women, and men) as indicated by their website labels. To ensure the generalisability of our model, we chose to develop a sparse Bayesian machine learning classifier using Relevance Vector Machine (RVM). Results: Using optimisation and normalisation techniques, we developed a best-performing classifier through joint optimisation of natural language features and min-max normalisation of feature frequencies. The AUC (0.957), sensitivity (0.900), and specificity (0.953) of the best model were statistically higher (p < 0.05) than other models using parallel optimisation of structural and semantic features with or without feature normalisation. We subsequently evaluated the diagnostic utility of our model in the clinic by comparing its positive (LR+) and negative likelihood ratios (LR−) and 95% confidence intervals (95% C.I.) as we adjusted the probability thresholds with the range of 0.1 and 0.9. We found that the best pair of LR+ (18.031, 95% C.I.: 10.992, 29.577) and LR− (0.100, 95% C.I.: 0.068, 0.148) was found when the probability threshold was set to 0.45 associated with a sensitivity of 0.905 (95%: 0.867, 0.942) and specificity of 0.950 (95% C.I.: 0.925, 0.975). These statistical properties of our model suggested its applicability in the clinic. Conclusion: Our study found that PAS had significant advantage over GEN mental health information regarding information actionability, engagement, and suitability for specific populations with distinct mental health issues. GEN is more suitable for general mental health information acquisition, whereas PAS can effectively engage patients and provide more effective and needed self-care support. The Bayesian machine learning classifier developed provided automatic tools to support decision making in the clinic to identify more actionable resources, effective to support self-care among different populations.

Sign in / Sign up

Export Citation Format

Share Document