Artificial Neural Networks to Estimate the Influence of Vehicular Emission Variables on Morbidity and Mortality in the Largest Metropolis in South America

Abstract

The emission of pollutants from vehicles is presented as a prime factor deteriorating air quality. Thus, seeking public policies encouraging the use and the development of more sustainable vehicles is paramount to preserve populations’ health. To better understand the health risks caused by air pollution and exclusively by mobile sources urges the question of which input variables should be considered. Therefore, this research aims to estimate the impacts on populations’ health related to road transport variables for São Paulo, Brazil, the largest metropolis in South America. We used three Artificial Neural Networks (ANN) (Multilayer Perceptron—MLP, Extreme Learning Machines—ELM, and Echo State Neural Networks—ESN) to estimate the impacts of carbon monoxide, nitrogen oxides, ozone, sulfur dioxide, and particulate matter on outcomes for respiratory diseases (morbidity—hospital admissions and mortality). We also used unusual inputs, such as road vehicles fleet, distributed and sold fuels amount, and vehicle average mileage. We also used deseasonalization and the Variable Selection Methods (VSM) (Mutual Information Filter and Wrapper). The results showed that the VSM excluded some variables, but the best performances were reached considering all of them. The ELM achieved the best overall results to morbidity, and the ESN to mortality, both using deseasonalization. Our study makes an important contribution to the following United Nations Sustainable Development Goals: 3—good health and well-being, 7—affordable and clean energy, and 11—sustainable cities and communities. These research findings will guide government about future legislations, public policies aiming to warranty and improve the health system.

Dynamic model to predict the association between air quality, COVID-19 cases, and level of lockdown

“…It seems clear that adjusting the hidden weights is an important step in nonlinear mapping applications, as is presented in this investigation. In this case, there are a set of inputs of variable nature (for example, temperature, humidity, and partial lockdown), and mapping these values to another variable is not a trivial task ( Kachba et al, 2020 ; Polezer et al, 2018 ).…”

“…It is a mathematical treatment that transforms the series of data into approximately stationary. Some studies have presented the importance of using such an approach ( Kachba et al, 2020 ; Siqueira et al, 2018 ).…”

“…We followed the premises from the literature of adopting the MSE as the most important error metric because this is reduced during the training (adjustment) of the neural models ( Araujo et al, 2020 ; Kachba et al, 2020 ; Siqueira et al, 2014 , 2018 , 2020 bib_Siqueira_et_al_2020 bib_Siqueira_et_al_2018 bib_Siqueira_et_al_2014 ).…”

“…Artificial Neural Networks (ANN), on the other hand, is a nonlinear methodology capable of mapping a set of inputs into an output, which is important to support decisions regarding preventive measures. This approach has been used in air pollution epidemiological studies ( Araujo et al, 2020 ; Kachba et al, 2020 ; Kassomenos et al, 2011 ; Polezer et al, 2018 ). In Araujo et al (2020) and Kassomenos et al (2011) , the ANN showed a better performance than linear approaches as Generalized Linear Models.…”

Comparative study of forecasting approaches in monthly streamflow series from Brazilian hydroelectric plants using Extreme Learning Machines and Box & Jenkins models

“…Artificial neural networks are structures of distributed processing of information, consisting of simple processing units, called artificial neurons, with a high degree of interconnectivity (Haykin, 2008;Lizot et al, 2020;Siqueira et al, 2012a;Tang et al, 1991;Thesari et al, 2019;Wang, 2011). One striking feature of this methodology is the ability of learning (Kachba et al, 2020;Zhang et al, 2018). Usually, the learning process involves the adaptation of the efficiencies of the connections between neurons, represented as weights that balance the transmitted signals, aiming to bring the desired response in the output of the network (Atiya et al, 1999;Francelin et al, 1996;Yaseen et al, 2016).…”