Forecasting PM10 levels using ANN and MLR: A case study for Sakarya City

Abstract

In this study, potential of neural network to estimate daily mean PM10 concentration levels in Sakarya city, Turkey as a case study was examined to achieve improved prediction ability. The level and distribution of air pollutants in a particular region is associated with changes in meteorological conditions affecting air movements and topographic features. Thus, meteorological variables data for a two-year period for Sakarya city which is located in most industrialized and crowded part of Turkey were selected as input. Neural network models and multiple linear regression models have been statistically evaluated. The results of the study showed that ANN models were accurate enough for prediction of PM10 levels.

Forecasting Particulate Matter Concentration Using Linear and Non-Linear Approaches for Air Quality Decision Support

“…Another study proved that the ANN model is better in the prediction of particulate matter in Hong Kong, which is in agreement with our findings [72]. Prediction of PM 10 in Sakarya City in Turkey has proved that the nonlinear model (MLP) outperforms the linear model (MLR) with vast differences in the R 2 of 0.84 and 0.32, respectively [73]. Similarly, a previous study in the development of prediction model in industrial area confirmed that the MLP model is able to increase the accuracy by 29.9%, and can reduce the error by 69.3% as compared to the MLR model [35].…”

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

“…However, only two methods were found to be frequently used. The first method was a min-max normalization that used the minimum and the maximum values of each feature to change the original data to a new range (Sabri & Tarek, 2012;Özdemir & Taner, 2014;Fu et al, 2015;Zu et al, 2017;Ceylan & Bulkan, 2018;V. Yadav & Nath, 2018).…”

“…• The linear perception of ANN had lower performance than the MLP-ANN because of their simple structure and inability to predict the complex problem (Perez & Reyes, 2002). • The MLR was more frequently implemented when being compared with the MLP-ANN but had lower performance (Alam & McNabola, 2015;Biancofiore et al, 2017;Cai et al, 2009;Ceylan & Bulkan, 2018;Chaloulakou et al, 2003;Gualtieri et al, 2018;Li et al, 2017;Liu et al, 2015;McKendry, 2002;Özdemir & Taner, 2014;Papanastasiou et al, 2007;Slini et al, 2006;Vlachogianni et al, 2011;Voukantsis et al, 2011;Zu et al, 2017). Interestingly, only one research article argued that the MLR performed slightly better than the ANN (Russo et al, 2015).…”

Neuro-Fuzzy Transformation with Minimize Entropy Principle to Create New Features for Particulate Matter Prediction

“…MLR is a popular statistical model for comparing the model performance with the ANN, but the results showed that MLR is less effective than ANN [17][18][19][20]. ARIMA is a common model for time-series data.…”