Improvement on PM-10 Forecast by Using Hybrid ARIMAX and Neural Networks Model for the Summer Season in Chiang Mai

Brazilian truckers’ strike and particulate matter (PM10) concentration: Temporal trend and time series models

“…In this study, Autoregressive Integrated Moving Average (ARI-MA) models were combined with regression models to assess the PM 10 concentration levels in the cities of Limeira and Campinas. These models are called ARIMAX and have been used in different studies to predict the concentration of air pollutants (Catalano et al, 2016;Wongsathan and Chankham, 2016;Ahani et al, 2019). Formally, the ARIMA process is based on the orders p and q, which are the orders of the autoregressive process (p) and moving averages (q), considered by taking the d th differences of the time series variables.…”

“…The autocorrelation (ACF) and partial autocorrelation (PACF) functions of the residuals were analyzed, as well as the plot of residuals over time and the residuals plot versus the percentiles of the normal distribution (qq-plot), all used to evaluate the statistical assumptions of the ARIMAX model (Catalano et al, 2016;Wongsathan and Chankham, 2016). To assess the normality of the residuals, Anderson-Darling and Shapiro-Wilks tests were performed (Moeeni and Bonakdari, 2018), as well as the Ljung-Box and Box-Pierce tests, to assess the autocorrelation of the residuals (Pinto et al, 2018).…”

Investigation of PM10 prediction utilizing data mining techniques: Analyze by topic

“…Yildirim & Bayramoglu, 2006;S. Lira et al, 2007;Sfetsos & Vlachogiannis, 2009;Su arez S anchez et al, 2011;Polat & Durduran, 2012;Yetilmezsoy & Abdul-Wahab, 2012;Muñoz et al, 2013;Asha B. Chelani, 2015;Cortina-Januchs et al, 2015;Hamid etal., 2016;Magaña-Villegas et al, 2016;Wongsathan & Chankham, 2016;W. Li et al, 2017;García Nieto et al, 2018a, 2018b.…”

“…Thirdly, the sigmoid activation function was mostly selected because it was simple and speed up the convergence during the learning process (Fu et al, 2015). Finally, the MLP-ANN model could be combined with other models to improve the model performance such as ARIMA (Díaz-Robles et al, 2008), ARIMAX (Wongsathan & Chankham, 2016), and k-NN (Perez, 2012).…”

“…Yildirim & Bayramoglu, 2006;Taşpınar, 2015). The last method applied a combination of two models in the model creation process as shown in Section 7 (Díaz-Robles et al, 2008;Goyal et al, 2006;Perez, 2012;Sabri & Tarek, 2012;Wongsathan & Chankham, 2016).…”