Article Details

Abstract

Aiming at the problem that the traditional BP Neural Network has slow convergence speed and is easy to fall into local optimal solution, this paper proposes a PM2.5 concentration prediction model based on Particle Swarm Optimization (PSO) algorithm optimized BP Neural Network, which can quickly converge and get the global optimal solution. Firstly, the pollutant indexes with high correlation with PM2.5 concentration are selected as input variables by Pearson correlation analysis. Secondly, the PSO algorithm is used to optimize the initial weights and thresholds of BP Neural Network, which overcomes the shortcomings of BP Neural Network, such as easy to fall into local optimum and slow convergence speed. Finally, the model is trained and tested using air pollutant data from July 2021 to June 2024 in Chengdu. The results show that the R2 of the test set is 0.944, the MAE of the test set is 4.231, and the RMSE of the test set is 6.364. Compared with the unoptimized BP Neural Network model, the PSO-BP model has higher prediction accuracy and faster convergence speed, and can effectively predict the PM2.5 concentration of the next day in Chengdu.

Keywords

PM2.5 concentration, prediction model, PSO algorithm, BP Neural Network

References

1. [1] ZHOU B H，WANG J，LIU S X，et al. Extrapolation of an Thropogenic Disturbances on Hazard Elements in PM2.5 Inatypical Heavy Industrial City in Northwest China [J].Environmental Scienceand Pollution Research International，2022，29（43）：64582-64596.
2. [2] Wu Tingting The impact of atmospheric PM2.5 on the number of cardiovascular disease hospitalizations and the mechanism of myocardial cell damage in urban areas of Lanzhou City [D]. Lanzhou: Lanzhou University, 2022
3. [3] YI YANG HU，HAI BIN LIAO，LI YUAN，et al. MTLPM: A Long-Term Fine-Grained PM2.5 Prediction Method Based on Spatio-Temporal Graph Neural Network [J].Environmental Monitoring and Assessment，2024，196（12）：1240-1240.
4. [4] Yang Xinxing, Feng Lihua, Wei Peng Atmospheric particulate matter PM2.5 and its hazards [J]. Frontier Science, 2012, 6 (1): 22-31
5. [5] Liu Xiaojian, Wu Yongsheng, Fu Yingbin, etc Generalized additive model analysis of PM2.5 and cardiovascular disease mortality in Shenzhen [J]. Chinese Journal of Disease Control and Prevention, 2016, 20 (2): 207-209
6. [6] Wang Yuanyuan, Shen Yu, Chen Xiuchuan, etc PM2.5 concentration prediction based on improved spatiotemporal causal convolutional network [J]. Automation Technology and Application, 2025, 44 (2): 21-25
7. [7] Ding Chengliang, Zheng Hongbo Research on PM2.5 concentration prediction model based on improved machine learning [J]. Journal of Dalian University of Technology, 2024, 64 (4): 353-360
8. [8] Lin Maijin, Zhang Lulu, Tang Youbing, etc PM2.5 concentration prediction based on CNN-GRU-SA combined model [J]. Science, Technology and Engineering, 2024, 24 (31): 13269-13276
9. [9] Zhang Danning, Wu Qiaoli, Zhang Bo A PM2.5 concentration prediction model based on BP neural network [J]. Building Materials Technology and Applications, 2023 (2): 9-13
10. [10] ZHANG S Q，ZHOU H Y. Beijing Population Migration Analysis Based on WPD-BP Neural Network [C]//Proceedings Volume 12510, International Conference on Statistics, Data Science, and Computational Intelligence (CSDSCI 2022).Chongqing：SPIE，2023：1-5.
11. [11] Peng Haojie, Zhou Yang, Hu Xiaofei, etc A PM2.5 concentration prediction model based on deep learning and random forest [J]. Journal of Remote Sensing, 2023, 27 (2): 430-440
12. [12] Zheng Junbao, Hua Sijie A PM2.5 concentration prediction model based on GA-BPNN [J]. Software Guide, 2021, 20 (9): 28-32
13. [13] TIAN J W，LIU Y，ZHENG W F，et al. Smog Prediction Based on the Deep Belief-BP Neural Network Model (DBN-BP) [J/OL].Urban Climate，2022，41：101078[2024-10-20].https://doi.org/10.1016/j.uclim.2021.101078.
14. [14] KENNEDY J，EBERHART R. Particle Swarm Optimization [C]//Proceedings of ICNN'95-International Conferenceon Neural Networks. Perth，1995，4：1942-1948.

Funding

Key Project of Sichuan Health Rehabilitation Vocational College (CWKY-2019Z-02); Sichuan Health Rehabilitation Vocational College School level Research Team (CWKY-TD24-10)

About the authors

Li Jialin (1987-), male, Han nationality, from Neijiang, Sichuan, lecturer, doctoral candidate, research direction: intelligent computing and deep learning; Hou Liming (1987-), male, Han ethnicity, from Xinxiang, Henan Province, associate professor, PhD, research direction: deep learning; Zhang Cong (1992-), male, Han ethnicity, from Zigong, Sichuan, is an engineer currently pursuing a master's degree in machine learning.