Selection of relevant input variables in storm water quality modeling by multiobjective evolutionary polynomial regression paradigm

Abstract

The growing availability of field data, from information and communication technologies (ICTs) in ''smart'' urban infrastructures, allows data modeling to understand complex phenomena and to support management decisions. Among the analyzed phenomena, those related to storm water quality modeling have recently been gaining interest in the scientific literature. Nonetheless, the large amount of available data poses the problem of selecting relevant variables to describe a phenomenon and enable robust data modeling. This paper presents a procedure for the selection of relevant input variables using the multiobjective evolutionary polynomial regression (EPR-MOGA) paradigm. The procedure is based on scrutinizing the explanatory variables that appear inside the set of EPR-MOGA symbolic model expressions of increasing complexity and goodness of fit to target output. The strategy also enables the selection to be validated by engineering judgement. In such context, the multiple case study extension of EPR-MOGA, called MCS-EPR-MOGA, is adopted. The application of the proposed procedure to modeling storm water quality parameters in two French catchments shows that it was able to significantly reduce the number of explanatory variables for successive analyses. Finally, the EPR-MOGA models obtained after the input selection are compared with those obtained by using the same technique without benefitting from input selection and with those obtained in previous works where other data-modeling techniques were used on the same data. The comparison highlights the effectiveness of both EPR-MOGA and the input selection procedure.

A hybrid multi-step sensitivity-driven evolutionary polynomial regression enables robust model structure selection

“…In this work, 100 Monte Carlo simulations were considered for each case study. We follow previous work (Berardi et al, 2008;Creaco et al, 2016) and utilize a vector of exponents with a step of 0.1 (𝐄𝐒 = [−2, −1.9, … , 1.9, 2]). This step size provides a good compromise between the CPU costs of our EPR method and the corresponding accuracy of the optimal model structure.…”

“…The correlation between LL and I P was found to be noteworthy (𝑟 = 0.91), indicating that LL could be excluded from the modeling approach. Since the objective of this paper is on model structural selection rather than on the influence of explanatory variables on models' performance (e.g., Creaco et al, 2016), LL was kept in the modeling for purposes of comparison with previous work (Jin et al, 2019b). The model's response to I P variation, available in Fig.…”

“…Most notably, MO-based EPR enables us to handle multiple objectives within the search strategy. For instance, several adopted MO-based EPR to maximize the model fitness to data and to minimize the number of polynomial terms (Alani and Faramarzi, 2014;Balf et al, 2018;Berardi et al, 2008;Creaco et al, 2016;Giustolisi and Savic, 2009;Rezania et al, 2008). The MO procedure thus returns a Pareto-efficient subset of feasible non-dominated solutions, that is, optimal model structures based on different criteria (Giustolisi and Savic, 2009).…”

Estimation of natural streams longitudinal dispersion coefficient using hybrid evolutionary machine learning model

“…Also, FS is known as parameter selection or attribute selection, is an independent process generally combined with the learning procedure of ML models, which aims to reduce the number of attributes to improve the performance of a predictor (Creaco et al, 2016). The central premise of FS is to remove the irrelevant or redundant features from the dataset; thus, the quality of the final estimations does not deteriorate (Meng & Li, 2017).…”

Advances in Modeling and Management of Urban Water Networks

“…Nevertheless, some problems of numerical instability have not yet been fully solved, calling for more research efforts. Finally, numerous experimental and numerical works have focused on water quality, e.g., [29][30][31], and sediment transport, e.g., [32][33][34], in sewers. However, some aspects, such as the sediment transport in pressurized pipes and the impact of water use reduction on the operation of UDSs, need to be analyzed in more depth.…”