A New PM2.5 Concentration Predication Study Based on CNN-LSTM Parallel Integration

Prevention and control of haze is the hot topic in the study of air quality, and PM2.5 concentration prediction is one of the keys in the haze prevention and control. This paper proposes a new method of integrating Convolutional Neural Networks (CNN) and Long Short-Term Memory Networks (LSTM) in parallel to predict PM2.5 concentration. This method can learn the spatial and temporal characteristics of data at the same time, and has powerful integrated learning capabilities. Taking the prediction of PM2.5 concentration in Xi'an area as an example, the method in this article is compared with the method in relevant authoritative literature. The experimental results show that the method in this article has better prediction effect and is a more competitive deep learning prediction model.