A hybrid Prophet-LSTM Model for Prediction of Air Quality Index
Landi Zhoul, Ming Chenl, Qingjian Ni
Abstract
In recent years, air pollution has caused widespread concern in society. This paper takes Nanjing as an example to analyze the changes in its Air Quality Index (AQI) and make predictions, which is of great significance for improving local air quality. Based on the air quality monitoring data of Nanjing from June 2014 to May 2019, a Prophet model is constructed to perform statistical analysis on the data and predict the AQI index from June 2019 to April 2020. At the same time, this paper designs Prophet-SVR hybrid model and Prophet-LSTM hybrid model to optimize the prediction accuracy of Prophet model. The results show that the Prophet-LSTM hybrid model has the best performance, its prediction accuracy is higher than that of the single model, and it has obvious advantages in the prediction of the air quality index.