Post
Published on September 6, 2022
Research by Dahdaleh Global Health Graduate Scholar Michael De Santi (lead author) and his coauthors, including DI Research Fellow Syed Imran Ali and DI Faculty Fellow Usman Khan, has recently been published in PLOS WATER – an open-access journal that brings together research relevant to the study of water, sanitation, and hygiene (WASH) and water resources for people and planet.
Modelling point-of-consumption residual chlorine in humanitarian response: Can cost-sensitive learning improve probabilistic forecasts?
Abstract
Ensuring sufficient free residual chlorine (FRC) up to the time and place water is consumed in refugee settlements is essential for preventing the spread of waterborne illnesses. Water system operators need accurate forecasts of FRC during the household storage period. However, factors that drive FRC decay after water leaves the piped distribution system vary substantially, introducing significant uncertainty when modelling point-of-consumption FRC. Artificial neural network (ANN) ensemble forecasting systems (EFS) can account for this uncertainty by generating probabilistic forecasts of point-of-consumption FRC. ANNs are typically trained using symmetrical error metrics like mean squared error (MSE), but this leads to forecast underdispersion forecasts (the spread of the forecast is smaller than the spread of the observations). This study proposes to solve forecast underdispersion by training an ANN-EFS using cost functions that combine alternative metrics (Nash-Sutcliffe efficiency, Kling Gupta Efficiency, Index of Agreement) with cost-sensitive learning (inverse FRC weighting, class-based FRC weighting, inverse frequency weighting). The ANN-EFS trained with each cost function was evaluated using water quality data from refugee settlements in Bangladesh and Tanzania by comparing the percent capture, confidence interval reliability diagrams, rank histograms, and the continuous ranked probability. Training the ANN-EFS using the cost functions developed in this study produced up to a 70% improvement in forecast reliability and dispersion compared to the baseline cost function (MSE), with the best performance typically obtained by training the model using Kling-Gupta Efficiency and inverse frequency weighting. Our findings demonstrate that training the ANN-EFS using alternative metrics and cost-sensitive learning can improve the quality of forecasts of point-of-consumption FRC and better account for uncertainty in post-distribution chlorine decay. These techniques can enable humanitarian responders to ensure sufficient FRC more reliably at the point-of-consumption, thereby preventing the spread of waterborne illnesses.
De Santi M, Ali SI, Arnold M, Fesselet J-F, Hyvärinen AMJ, Taylor D, et al. (2022) Modelling point-of-consumption residual chlorine in humanitarian response: Can cost-sensitive learning improve probabilistic forecasts? PLOS Water 1(9): e0000040. https://doi.org/10.1371/journal.pwat.0000040
Join us on Wednesday, September 7 to hear from the authors directly. Register here.
Themes | Global Health & Humanitarianism |
Status | Active |
Related Work | |
Updates |
N/A
|
People |
Usman T. Khan, Faculty Fellow, Lassonde School of Engineering Active
Syed Imran Ali, Research Fellow, Global Health and Humanitarianism Active Matthew Arnold, Technical Advisor, Safe Water Optimization Tool Alum Michael De Santi, Dahdaleh Global Health Graduate Scholar, Lassonde School of Engineering Active |
You may also be interested in...
Recap — Urban Inequality During and After COVID-19 in Johannesburg, Chicago, and Toronto
In this seminar on May 24, Dahdaleh graduate student scholar Hillary Birch and Dahdaleh faculty fellow Roger Keil along with their colleagues shared their findings on how communities and governments in Johannesburg, Chicago, and Toronto ...Read more about this Post
Key Moments from the Trust and Political Sociology of Health Workshop
The "Trust and Political Sociology of Health" research workshop took place on April 4th at the Dahdaleh Institute for Global Health Research, York University. This forum brought together senior and emerging scholars, practitioners, and students ...Read more about this Post
Highlighting the SWOT’s Risk-Based Approach at the 2022 Emergency Environmental Health Forum
Dahdaleh Institute member Michael DeSanti is presenting on the innovative Safe Water Optimization Tool (SWOT) at the 2022 Emergency Environmental Health Forum. He will discuss SWOT’s risk-based approach to generating chlorination guidance to ensure safe ...Read more about this Post