Research And Application Of Air Quality Index Prediction Algorithm Based On Deep Learning

Air quality is related to people’s health.Severe air pollution can cause respiratory problems.Therefore,air quality prediction is necessary.Air quality index is a general index to evaluate air quality.It is obtained by quantifying the concentration of 6 air pollutants and selecting the maximum value.In this paper,air quality data and weather data of 20 city stations in Jiangsu province are processed,so as to predict air quality index.Aiming at the problem of too many weather factors,feature selection is used to reduce the useless factors for the air quality index prediction,so as to ensure the prediction accuracy and reduce the time complexity.In view of the complex and diverse trend of air quality data in many cities,this paper first clusters several consecutive days of different city air quality to obtain the categories with similar air quality trends.On that basis,the air quality index prediction model based on clustering is obtained by training different sub-models of the air quality data of different classes formed after clustering.In order to solve the problem that the variation trend of air quality data in a single city is not obvious,this paper can obtain the variation trend of air quality in different time scales by using empirical mode decomposition.According to the changing trend under different time scale,the long short term memory network is used to train,and the corresponding predicted value is obtained.The empirical modal decomposition fuzzy prediction model of air quality index based on long short term memory network is obtained.The experiments show that the two models have good predictive performance of air quality index.In addition,this paper designs an air quality index prediction system,which can display the historical air quality and obtain the predicted results.It can help the environmental management department to improve the ecological security early warning ability.