Enhanced prediction of ozone concentrations using an artificial neural network model
DOI:
https://doi.org/10.52292/j.laar.2025.3489Keywords:
Artificial neural networks, Ozone prediction, Air quality monitoring, Environmental modeling, Particulate matter, Predictive analyticsAbstract
The primary goal of this research was to develop an Artificial Neural Network (ANN) model to predict ozone (O3) concentrations using hourly data obtained from a monitoring station in Samsun City, located in the Middle Black Sea Region of Turkey. The dataset utilized encompassed the years from 2016 to 2020. The ANN architecture incorporated eleven input nodes representing various parameters: month, hour, concentrations of particulate matter (PM2.5 and PM10), nitrogen oxides (NOx), wind direction, relative humidity, air temperature, wind speed, cabin temperature of the measuring station, and air pressure. The focus of the model’s output was on predicting the O3 concentration. During the training and testing phases, the ANN model displayed outstanding performance, as evidenced by correlation coefficients nearing one. The model also registered minimal values for Mean Absolute Percentage Error (MAPE, %), Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). In the training phase, the model achieved a Training R-value of 0.9993, an RMSE of 0.7424, a MAPE of 4.3221%, and a MAE of 0.5301. The testing phase showed equally strong results, with a Test R-value of 0.9990, an RMSE of 0.8595, a MAPE of 4.5642%, and an MAE of 0.5823. These outcomes emphasize the model's ability to accurately predict ozone concentrations, markedly enhancing the precision compared to previous models based on traditional statistical methods. The findings of this study highlight the potential of this ANN model in providing precise ozone concentration readings in the atmosphere. The proposed ANN model distinguishes itself from previous studies by incorporating more representative variables as inputs, significantly boosting prediction accuracy. Additionally, the removal of outliers during preprocessing enhances data quality, thereby increasing the reliability of the predictions. Despite its simple structure, the model demonstrates high performance, making it both innovative and effective in comparison to earlier models. Moreover, the model's superior performance may reduce the need for additional measurement devices at newly established monitoring stations.
Published
Issue
Section
License
Copyright (c) 2025 Latin American Applied Research - An international journal

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Once a paper is accepted for publication, the author is assumed to have transferred its copyright to the Publisher. The Publisher will not, however, put any limitation on the personal freedom of the author to use material from the paper in other publications. From September 2019 it is required that authors explicitly sign a copyright release form before their paper gets published. The Author Copyright Release form can be found here