Interpretable vs. noninterpretable machine learning models for data-driven hydro-climatological process modeling

Chakraborty, Debaditya; Başağaoğlu, Hakan; Winterle, James

doi:10.1016/j.eswa.2020.114498

Cited by 73 publications

(29 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Similar results have been obtained. Similar results were concluded in different settings [13][14][15][16][17][18][19][20][21][22][23] when handling missing data using ML/AI techniques, depends on the scenario and type of the data structure to be imputed.…”

Section: Discussionsupporting

confidence: 72%

“…In the last decade, the use of artificial intelligence (AI) and machine-learning (ML) techniques has increased substantially in a diverse range of disciplines, such as climatology, industry, and biomedicine [13][14][15][16][17][18][19][20]. There is a vast and diverse number of ML algorithms which may be mostly categorized based on the approach taken toward the data set as well as the type of data processed.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Imputation for Repeated Bounded Outcome Data: Statistical and Machine-Learning Approaches

Larracoechea

Borges

2021

Mathematics

View full text Add to dashboard Cite

Real-life data are bounded and heavy-tailed variables. Zero-one-inflated beta (ZOIB) regression is used for modelling them. There are no appropriate methods to address the problem of missing data in repeated bounded outcomes. We developed an imputation method using ZOIB (i-ZOIB) and compared its performance with those of the naïve and machine-learning methods, using different distribution shapes and settings designed in the simulation study. The performance was measured employing the absolute error (MAE), root-mean-square-error (RMSE) and the unscaled mean bounded relative absolute error (UMBRAE) methods. The results varied depending on the missingness rate and mechanism. The i-ZOIB and the machine-learning ANN, SVR and RF methods showed the best performance.

show abstract

Section: Discussionsupporting

confidence: 72%

Section: Introductionmentioning

confidence: 99%

Imputation for Repeated Bounded Outcome Data: Statistical and Machine-Learning Approaches

Larracoechea

Borges

2021

Mathematics

View full text Add to dashboard Cite

show abstract

“…In Ref. [59], we had compared the predictive performance of the non-probabilistic ML models-including XGBoost, support vector machines (SVM), long short-term memory networks (LSTM), deep learning (DL), random forest (RF), and linear regression (LR)-in predicting ET o from structured tabular datasets acquired from multiple weather stations. In that study, the top performing interpretable XGBoost model exhibited comparable predictive accuracy to the top performing noninterpretable DL model.…”

Section: Et Predictions Using Probabilistic ML Modelsmentioning

confidence: 99%

“…Based on the analysis in Ref. [59], we chose RF as the baseline model in this study to establish a point of reference for analyzing the performance of the hybrid NGBoost-XGBoost ML model. Table 1 shows that the hybrid model exhibited better predictive performance than RF in terms of point predictions of both ET o and ET a on the testing data.…”

Section: Et Predictions Using Probabilistic ML Modelsmentioning

confidence: 99%

Reliable Evapotranspiration Predictions with a Probabilistic Machine Learning Framework

Başağaoğlu¹,

Chakraborty

Winterle³

2021

Water

Self Cite

View full text Add to dashboard Cite

Evapotranspiration is often expressed in terms of reference crop evapotranspiration (ETo), actual evapotranspiration (ETa), or surface water evaporation (Esw), and their reliable predictions are critical for groundwater, irrigation, and aquatic ecosystem management in semi-arid regions. We demonstrated that a newly developed probabilistic machine learning (ML) model, using a hybridized “boosting” framework, can simultaneously predict the daily ETo, Esw, & ETa from local hydroclimate data with high accuracy. The probabilistic approach exhibited great potential to overcome data uncertainties, in which 100% of the ETo, 89.9% of the Esw, and 93% of the ETa test data at three watersheds were within the models’ 95% prediction intervals. The modeling results revealed that the hybrid boosting framework can be used as a reliable computational tool to predict ETo while bypassing net solar radiation calculations, estimate Esw while overcoming uncertainties associated with pan evaporation & pan coefficients, and predict ETa while offsetting high capital & operational costs of EC towers. In addition, using the Shapley analysis built on a coalition game theory, we identified the order of importance and interactions between the hydroclimatic variables to enhance the models’ transparency and trustworthiness.

show abstract

“…Many research projects have aimed to address this limitation and understand the pros and cons of various types of ML models. In [76], there was a focus on the tension between the interpretable and non-interpretable ML models by examining the predictive capabilities of three models of the former type (linear regression, random forest (RF; ensembles of decision trees), and extreme gradient boosting) and three models of the latter type (SVM and two NN architectures: a LSTM and a dense, fully-connected neural network, also known as perceptron). The hydro-climatological problem they tackled concerned modeling of crop evapotranspiration.…”

Section: Related Workmentioning

confidence: 99%

Machine Learning Modeling of Climate Variability Impact on River Runoff

Norel

Krawiec

Kundzewicz

2021

Water

View full text Add to dashboard Cite

The hypothesis of this study was one of existence of spatially organized links between the time series of river runoff and climate variability indices, describing the oscillations in the atmosphere–ocean system: ENSO (El Niño–Southern Oscillation), PDO (Pacific Decadal Oscillation), AMO (Atlantic Multidecadal Oscillation), and NAO (North Atlantic Oscillation). The global river flow reconstructions (ERA-20-CM-R) for 18 study areas on six continents and climate variability indices for the period 1901–2010 were used. The split-sample approach was applied, with the period 1901–2000 used for training and 2001–2010 used for testing. The quality measures used in this paper were mean absolute error, dynamic time warping, and top extreme events error. We demonstrated that a machine learning approach (convolution neural network, CNN) trained on climate variability indices can model the river runoff better than the long-term monthly mean baseline, both in univariate (per-cell) and multivariate (multi-cell, regionalized) settings. We compared the models to the baseline in the form of heatmaps and presented results of ablation experiments (test time ablation, i.e., jackknifing, and training time ablation), which suggested that ENSO is the primary determinant among the considered indices.

show abstract

Interpretable vs. noninterpretable machine learning models for data-driven hydro-climatological process modeling

Cited by 73 publications

References 44 publications

Imputation for Repeated Bounded Outcome Data: Statistical and Machine-Learning Approaches

Imputation for Repeated Bounded Outcome Data: Statistical and Machine-Learning Approaches

Reliable Evapotranspiration Predictions with a Probabilistic Machine Learning Framework

Machine Learning Modeling of Climate Variability Impact on River Runoff

Contact Info

Product

Resources

About