Artificial neural network model for air pollution prediction: case study of Moravica District, Serbia

Blagojević, Marija; Papić, Miloš; Vujičić, Momčilo; Šućurović, Marko

doi:10.37190/epe180110

Cited by 6 publications

(8 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…ANN closely follows the structure and functionality of the human brain and its neurons to solve complex problems faster with minimal human interventions, hence reducing error rates [6]. As ANN is evolving with newer algorithms, few studies [9][10][11][15][16][17][18] have considered their applicability in population health studies. As far as we know, most of the studies in the fields of population health and medicine have used different deep learning techniques to optimize classification of health outcomes and medical data [19,20], and disease screening/diagnosis [21,22].…”

Section: Introductionmentioning

confidence: 99%

Time series prediction of under-five mortality rates for Nigeria: comparative analysis of artificial neural networks, Holt-Winters exponential smoothing and autoregressive integrated moving average models

Adeyinka

Muhajarine

2020

BMC Med Res Methodol

View full text Add to dashboard Cite

Background Accurate forecasting model for under-five mortality rate (U5MR) is essential for policy actions and planning. While studies have used traditional time series modeling techniques (e.g., autoregressive integrated moving average (ARIMA) and Holt-Winters smoothing exponential methods), their appropriateness to predict noisy and non-linear data (such as childhood mortality) has been debated. The objective of this study was to model long-term U5MR with group method of data handling (GMDH)-type artificial neural network (ANN), and compare the forecasts with the commonly used conventional statistical methods—ARIMA regression and Holt-Winters exponential smoothing models. Methods The historical dataset of annual U5MR in Nigeria from 1964 to 2017 was obtained from the official website of World Bank. The optimal models for each forecasting methods were used for forecasting mortality rates to 2030 (ending of Sustainable Development Goal era). The predictive performances of the three methods were evaluated, based on root mean squared errors (RMSE), root mean absolute error (RMAE) and modified Nash-Sutcliffe efficiency (NSE) coefficient. Statistically significant differences in loss function between forecasts of GMDH-type ANN model compared to each of the ARIMA and Holt-Winters models were assessed with Diebold-Mariano (DM) test and Deming regression. Results The modified NSE coefficient was slightly lower for Holt-Winters methods (96.7%), compared to GMDH-type ANN (99.8%) and ARIMA (99.6%). The RMSE of GMDH-type ANN (0.09) was lower than ARIMA (0.23) and Holt-Winters (2.87). Similarly, RMAE was lowest for GMDH-type ANN (0.25), compared with ARIMA (0.41) and Holt-Winters (1.20). From the DM test, the mean absolute error (MAE) was significantly lower for GMDH-type ANN, compared with ARIMA (difference = 0.11, p-value = 0.0003), and Holt-Winters model (difference = 0.62, p-value< 0.001). Based on the intercepts from Deming regression, the predictions from GMDH-type ANN were more accurate (β0 = 0.004 ± standard error: 0.06; 95% confidence interval: − 0.113 to 0.122). Conclusions GMDH-type neural network performed better in predicting and forecasting of under-five mortality rates for Nigeria, compared to the ARIMA and Holt-Winters models. Therefore, GMDH-type ANN might be more suitable for data with non-linear or unknown distribution, such as childhood mortality. GMDH-type ANN increases forecasting accuracy of childhood mortalities in order to inform policy actions in Nigeria.

show abstract

Section: Introductionmentioning

confidence: 99%

Time series prediction of under-five mortality rates for Nigeria: comparative analysis of artificial neural networks, Holt-Winters exponential smoothing and autoregressive integrated moving average models

Adeyinka

Muhajarine

2020

BMC Med Res Methodol

View full text Add to dashboard Cite

show abstract

“…El método que permite alcanzar una mayor confiabilidad en los resultados fue el DL, por medio de una red neuronal no lineal auto-regresiva (auto-regressive nonlinear neural network, ARNNN). M. Blagojevic et al [108] proponen una MLP con un algoritmo de aprendizaje basado en BP para la clasificación del nivel de contaminación del aire en el distrito Moravica en Serbia, a través de datos como: municipio, sitio, año, niveles de hollín, dióxido de sulfuro (SO2), dióxido de nitrógeno (NO2) y material particulado. X. Feng et al [109] proponen un modelo híbrido que combina el análisis de la trayectoria de la masa de aire y la transformada wavelet para mejorar la precisión de una red MLP para el pronóstico de las concentraciones de material particulado (PM2.5).…”

Section: Gestión Y Administración De La Salud Públicaunclassified

Aplicaciones de las redes neuronales y el deep learning a la ingeniería biomédica

Ramos¹

2020

revuin

View full text Add to dashboard Cite

Hoy en día, las redes neuronales artificiales y el deep learning, son dos de las herramientas más poderosas del aprendizaje de máquina, que tienen por objetivo desarrollar sistemas que aprenden automáticamente, reconocen patrones, predicen comportamientos y generalizan información a partir de conjuntos de datos. Estasdos herramientas se han convertido en un potencial campo de investigación con aplicaciones a la ingeniería, no siendo la ingeniería biomédica la excepción. En este artículo se presenta una revisión actualizada de las principales aplicaciones de las redes neuronales y el deep learning a la ingeniería biomédica en las ramas de la ómica, la imagenología, las interfaces cerebro-máquina y hombre-máquina, y la gestión y administración de la salud pública; ramas que se extienden desde el estudio de procesos a nivel molecular, hasta procesos que involucran grandes poblaciones.Palabras clave:aprendizaje de máquina; inteligencia artificial; reconocimiento de patrones; ómica; bioinformática; biomedicina; imagenología; interfaces cerebro-máquina; interfaces hombre-máquina;salud pública.

show abstract

“…Measuring points were based on location, population density, and landscape and weather conditions. Soot levels were measured using nitrogen oxide and sulphur dioxide levels, while the PM concentration was measured by analysing soluble and insoluble particles in the air [13].…”

Section: Literature Reviewmentioning

confidence: 99%

“…Input layers contained input data, including the year, municipality name and measuring site as additional attributes. The hidden layer contained neurons for receiving the input and processing the output, while the additional attribute for this layer was the air pollution level [13].…”

Section: Literature Reviewmentioning

confidence: 99%

See 1 more Smart Citation

Air Pollution Prediction using Machine Learning Algorithms

Aljuaid,

Alwabel

2019

IJEAT

View full text Add to dashboard Cite

Air pollution has a serious impact on human health. It occurs because of natural and man-made factors. The major contribution of this research is that it provides a comparison between different methodologies and techniques of mathematical and machine learning models. The process began with integrating data from different sources at different time interval. The preprocessing phase resulted in two different datasets: one-hour and five-minute datasets. Next, we established a forecasting model for particulate matter PM2.5, which is one of the most prevalent air pollutants and its concentration affects air quality. Additionally, we completed a multivariate analysis to predict the PM2.5 value and check the effects of other air pollutants, traffic, and weather. The algorithms used are support vector regression, k-nearest neighbors and decision tree models. The results showed that for the one-hour data set, of the three algorithms, support vector regression has the least root-mean-square error (RMSE) and also lowest value in mean absolute error (MAE). Alternatively, for the five-minute dataset, we found that the auto-regression model showed the least RMSE and MAE; however, this model only predicts short-term PM2.5.

show abstract

Artificial neural network model for air pollution prediction: case study of Moravica District, Serbia

Cited by 6 publications

References 9 publications

Time series prediction of under-five mortality rates for Nigeria: comparative analysis of artificial neural networks, Holt-Winters exponential smoothing and autoregressive integrated moving average models

Time series prediction of under-five mortality rates for Nigeria: comparative analysis of artificial neural networks, Holt-Winters exponential smoothing and autoregressive integrated moving average models

Aplicaciones de las redes neuronales y el deep learning a la ingeniería biomédica

Air Pollution Prediction using Machine Learning Algorithms

Contact Info

Product

Resources

About