Author(s)

Dan Chen

Affiliation(s)

Department of Statistics, Guangxi Normal University, Guilin 541006, Guangxi, China.

Corresponding Author

Dan Chen

ABSTRACT

With the acceleration of urbanization and the improvement of industrialization, air quality has become an important factor affecting the quality of life and health of urban residents. Under this background, the air quality problem has become increasingly prominent and has become a key factor affecting the quality of life and health of urban residents. Therefore, how to accurately predict the air quality index (AQI) in order to take timely and effective measures to protect the health of residents has become a major challenge we face. This study takes Guilin city as the object, and constructs an air quality prediction model by integrating big data and artificial intelligence technology. Based on the historical monitoring data such as PM2.5 and NO2, a training set/test set division method is adopted, and the ARIMA time series model and the LSTM depth learning network are comprehensively applied: the former captures the trend and seasonal characteristics of the data, and the latter handles the complex sequence relationship. By introducing the rolling prediction mechanism to update the model parameters in real time, the dynamic adaptability of the prediction system is effectively improved. Experiments show that the hybrid model significantly improves the prediction accuracy and stability of AQI, and provides a scalable technical scheme for urban air quality management.

KEYWORDS

LSTM model; ARIMA model; Rolling forecast; Air quality index forecast

CITE THIS PAPER

Dan Chen. Prediction of air quality index based on arima and lstm models—take guilin as an example. AI and Data Science Journal. 2025, 2(1): 106-115. DOI: https://doi.org/10.61784/adsj3017.

REFERENCES

[1]Fang Xiaoping, Chen Xiuluan, Chu Qi, et al. prediction of air quality index in Yangtze river economic zone based on PSOGSA-LSTM model. Mathematical statistics and management, 2023, 42(01).

[2]Zheng Yangyang, Bai Yanping, Hou Yuchao. Application of LSTM model based on Keras in prediction of air quality index. Mathematical practice and understanding, 2019, 49(07): 138-143.

[3]Ling Jin. Short-term wind speed prediction method based on rolling prediction and its error analysis. Hefei University of Technology, 2017.

[4]Sun Jing, Zhu Jianlin, Li Wanlan, et al. Research on rolling forecast of consumer confidence index based on internet data. Journal of Xi 'an Jiaotong University (Social Science Edition), 2021, 41(06).

[5]Ma Siyuan, Jiao Jiahui, Ren Shengqi, et al. Missing Value Filling of Urban Multiple Air Quality Data Based on Attention Mechanism. Computer Engineering and Science, 2023, 45(08): 1354-1364.

[6]Guo Kuo. Applicability Analysis of ARIMA Model in Precipitation Forecast in Fuxin Area. Heilongjiang Water Conservancy Science and Technology, 2024, 52(04): 37-39+79.

[7]Gu Keming. Research on air quality prediction based on spatio-temporal depth neural network. Zhejiang University of Technology, 2024.

[8]YU Hui Qing, HUANG Peng Cheng, ZHAO Zhen Yu, et al. Standard cell delay prediction method based on RNN. Journal of Zhengzhou University (Science Edition), 2024, 1-7.

[9]Tian Xuwei, Yang Kai, Yin Tong, et al. Prediction of port ambient air quality index based on SSA-Bi-LSTM. Transportation Energy Conservation and Environmental Protection, 2023, 19(05): 32-36+41.

[10]Zhang Tingzhong, Zhang Qinghui, Xing Qiang, et al. Short-term wind speed rolling prediction method based on combination of LMD and GA-BP neural network. Shandong Power Technology, 2019, 46(11): 13-20.

[11]Atakan J, Ayse B O. Forecasting air pollutant indicator levels with geographic model 3 days in advance using neural networks. Expert Systems with Applications, 2010, 37(12): 7986-7992.

[12]A Monterio, M Vieira, C Gama, et al. Miranda. 2017. Towards an improve air quality index. Air Quality,  Atmosphere And Health, 10(4): 447-445.

[13]Navares R, Aznarte J L. Predicting air quality with deep learning LSTM: Towards comprehensive models. Ecological Informatics, 2019, 55: 101-119.