An improved particle swarm optimization (PSO): method to enhance modeling of airborne particulate matter (PM10)

Modeling air pollution by integrating ANFIS and metaheuristic algorithms

Air pollution is increasing for many reasons, such as the crowding of cities, the failure of planning to consider the benefit of society and nature, and the non-implementation of environmental legislation. In the recent era, the impacts of air pollution on human health and the ecosystem have become a primary global concern. Thus, the prediction of air pollution is a crucial issue. ANFIS is an artificial intelligence technique consisting of artificial neural networks and fuzzy inference systems, and it is widely used in estimating studies. To obtain effective results with ANFIS, the training process, which includes optimizing its premise and consequent parameters, is very important. In this study, ANFIS training has been performed using three popular metaheuristic methods: Genetic Algorithm (GA), Particle Swarm Optimization (PSO), and Differential Evolution (DE) for modeling air pollution. Various air pollution parameters which are particular matters: PM2.5 and PM10, sulfur dioxide (SO2), ozone (O3), nitrogen dioxide (NO2), carbon monoxide (CO), and several meteorological parameters such as wind speed, wind gust, temperature, pressure, and humidity were utilized. Daily air pollution predictions in Istanbul were obtained using these particular matters and parameters via trained ANFIS approaches with metaheuristics. The prediction results from GA, PSO, and DE-trained ANFIS were compared with classical ANFIS results. In conclusion, it can be said that the trained ANFIS approaches are more successful than classical ANFIS for modeling and predicting air pollution.

Performance evaluation of a recurrent deep neural network optimized by swarm intelligent techniques to model particulate matter

Atmospheric pollution refers to the presence of substances in the air such as particulate matter (PM) which has a negative impact in population ́s health exposed to it. This makes it a topic of current interest. Since the Metropolitan Zone of the Valley of Mexico's geographic characteristics do not allow proper ventilation and due to its population's density a significant quantity of poor air quality events are registered. This paper proposes a methodology to improve the forecasting of PM10 and PM2.5, in largely populated areas, using a recurrent long-term/short-term memory (LSTM) network optimized by the Ant Colony Optimization (ACO) algorithm. The experimental results show an improved performance in reducing the error by around 13.00% in RMSE and 14.82% in MAE using as reference the averaged results obtained by the LSTM deep neural network. Overall, the current study proposes a methodology to be studied in the future to improve different forecasting techniques in real-life applications where there is no need to respond in real time.Implications: This contribution presents a methodology to deal with the highly non-linear modeling of airborne particulate matter (both PM10 and PM2.5). Most linear approaches to this modeling problem are often not accurate enough when dealing with this type of data. In addition, most machine learning methods require extensive training or have problems when dealing with noise embedded in the time-series data. The proposed methodology deals with this data in three stages: preprocessing, modeling, and optimization. In the preprocessing stage, data is acquired and imputed any missing data. This ensures that the modeling process is robust even when there are errors in the acquired data and is invalid, or the data is missing. In the modeling stage, a recurrent deep neural network called LSTM (Long-Short Term Memory) is used, which shows that regardless of the monitoring station and the geographical characteristics of the site, the resulting model shows accurate and robust results. Furthermore, the optimization stage deals with enhancing the capability of the data modeling by using swarm intelligence algorithms (Ant Colony Optimization, in this case). The results presented in this study were compared with other works that presented traditional algorithms, such as multi-layer perceptron, traditional deep neural networks, and common spatiotemporal models, which show the feasibility of the methodology presented in this contribution. Lastly, the advantages of using this methodology are highlighted.

Short-term prediction of carbon emissions based on the EEMD-PSOBP model

The recovery of carbon emissions in the past 2 years has alerted us that carbon emissions are a long-term process, and setting short-term emission reduction targets can more effectively curb the rising trend of carbon emissions. Therefore, the research on short-term prediction of carbon emissions is particularly important. In this paper, the idea of "decomposition-prediction" is put forward in the short-term prediction of carbon emissions, and the combined model of "decomposition-prediction" is constructed. The model is composed of ensemble empirical mode decomposition (EEMD) and the backpropagation neural network based on particle swarm optimization (PSOBP). It is also the first time that EEMD has been applied to the field of carbon emission prediction. Firstly, EEMD is used to decompose the daily carbon emission monitoring data into 6 modal functions and one residual sequence, and the partial autocorrelation function (PACF) is used to determine the input of each modal function. Then, PSOBP was used to predict. Finally, adding the prediction results of each sequence to get the final prediction results. To verify the effectiveness and superiority of the EEMD-PSOBP model, 14 comparative models were constructed, and the prediction effect of the models was evaluated by R², RMSE, and MAPE. All the prediction results show that the proposed model has the best prediction performance (R²=0.9507, RMSE=0.3431, MAPE=0.093). Compared with PSOBP, the R² of EEMD-PSOBP was increased by 63.58%, and RMSE and MAPE were decreased by 65.18% and 64.23%, respectively. The accuracy of prediction can be improved significantly by decomposing before predicting. It was also found that EEMD had the highest predictive performance improvement. Therefore, this model will have broad development prospects in the field of short-term carbon emission prediction in the future.