Machine-learning methods for hydrological imputation data: analysis of the goodness of fit of the model in hydrographic systems of the Pacific - Ecuador

Abstract

Computational methods based on machine learning have had extensive development and application in hydrology, especially for modelling systems that do not have enough data. Within this problem, there are data series that are missing, and that should not necessarily be discarded; this is achieved by means of the imputation of the same ones, obtaining complete sets. For this reason, this research proposes a comparison of computer-learning techniques to identify those best suited for hydrographic systems of the Pacific of Ecuador. For the elaboration of this investigation, the hydro-meteorological records of the monitoring stations located in the watersheds of the Esmeraldas, Cañar and Jubones Rivers were used for 22 years, between 1990 and 2012. The variables that were imputed were precipitation and flow. Automatic learning machines of the Python Scikit_Learn module were used; these modules integrate a wide range of automated learning algorithms, such as Linear Regression and Random Forest. Finally, results were obtained that led to a minimum useful mean square error for Random Forest as an automatic machine-learning imputation method that best fits the systems and data analyzed.

Regionalization of Rainfall Regimes Using Hybrid RF-Bs Couple with Multivariate Approaches

“…Random forests (RF) is a supervised machine learning algorithm that has recently begun to gain popularity in applications for water management [23,24]. In statistical analyses, missing data are common, and imputation methods based on RF have become popular for handling them, especially in climate research.…”

Methodological approaches for imputing missing data into monthly flows series

“…Because missing data imputation is a useful tool in water-resource management studies (Barnetche and Kobiyama, 2006), several authors have worked on the application of techniques for imputing missing data in hydrological studies resulting in a variety of methods ranging from simple imputation by mean or median to widely used statistical methods such as Regional Weighting (Ely et al, 2021); interpolations (linear, quadratic and cubic) (Gyau-Boakye andSchultz, 1994, Hamzah et al, 2020); methods based on linear regressions (single and multiple) (Kamwaga et al, 2018;Khalifeloo et al, 2015); Self Organizing Map (SOM) and Soil and Water Assessment Tool (SWAT) (Kim et al, 2015); to more advanced and robust methods, such as different Artificial Neural Network approaches (Canchala-Nastar et al, 2019;Elshorbagy et al, 2000;Nkiaka et al, 2016;Starrett et al, 2010;Vega-Garcia et al, 2019); machine learning methods (Heras and Matovelle, 2021;Rado et al, 2019); satellite radar altimetry and multiple imputation (Ekeu-Wei et al, 2018); combination of regression and autoregressive integrated moving average (ARIMA) models called dynamic regression (Tencaliec et al, 2015); Singular Spectrum Analysis (SSA) and Multichannel Singular Spectrum Analysis (MSSA) (Semiromi and Koch, 2019); among many others. The many methods that can be used for hydrological missing data imputation resulted in literature reviews as can be seen in Ben Aissia et al (2017) and Hamzah et al (2020).…”