Research Of Air Quality Prediction Model Based On Deep Learning

In recent years,with the rapid development of China's economy and the rapid advancement of urbanization,the air pollution problem is getting worse and worse.Air pollution not only seriously affects people's daily life and physical and mental health,but also poses a huge obstacle to the sustainable development of society,which has aroused great attention from all walks of life.Therefore,the accurate evaluation and prediction of air quality is of important practical significance and social value.With the rise of big data and artificial intelligence technology,traditional air quality prediction methods can no longer meet the needs of intelligent processing of big data.Many scholars began to use intelligent methods to evaluate and predict air quality based on big data.Deep learning is an important branch of artificial intelligence technology.It has powerful feature extraction and data fitting abilities,and is widely used in image recognition and classification prediction.Therefore,in this paper,deep learning was introduced into air quality prediction,and the training sample database was built with air quality monitoring data of xi 'an from November 2013 to February 2019.Based on Deep Belief Network(DBN)and Automatic Encoder(AE),three deep learning-based air quality prediction models named DBN,DBN-ELM and DAE-BP were constructed.The main work of the research is as follows:(1)By reading a large number of domestic and foreign relevant literatures,six pollutants such as carbon monoxide,nitrogen dioxide,sulfur dioxide,PM2.5,PM10 and ozone were selected as evaluation factors of air quality prediction model,and the AQI value was used as the target variable.(2)The original test data of this paper were pre-processed such as cleaning and filling,and a sample library of air quality prediction model was constructed.(3)Aiming at the shortcomings of traditional regression model and shallow machine learning algorithm in air quality prediction,this paper constructs a deep quality belief network(DBN)based air quality evaluation and prediction model based on the powerful feature extraction ability of deep belief network.Results show that the method constructed in this paper has better classification prediction effect than the traditional prediction algorithm,which verifies the effectiveness of deep belief network in air quality prediction,and provides a new idea for the research of air quality prediction.(4)Aiming at the deficiency of traditional deep belief network(DBN)in feature extraction and parameter training,the cross entropy sparse penalty factor mechanism is introduced into the feature extraction process of DBN,and DBN and ELM algorithm are combined to propose an air quality prediction model based on DBN-ELM algorithm.The experimental results show that the proposed method has better classification prediction effect than the traditional deep belief network and the traditional shallow machine learning prediction algorithm.(5)A deep quality automatic encoder(DAE)and BP neural network are combined to construct an air quality prediction model based on DAE-BP algorithm.Firstly,multiple self-editors(AE)are stacked together to construct a deep feature extractor(DAE)layer-by-layer feature extraction of the air quality data set;then the extracted features are input into the BP neural network,using the self-encoder network.Weights are used to initialize the BP neural network.Finally,the back propagation algorithm of the BP neural network is used to adjust and optimize the parameters of the model to achieve accurate prediction of air quality.The experimental results show that the DAE-BP model has better prediction effect than the traditional shallow machine learning prediction algorithm,and proves the effectiveness of the deep self-encoder model in air quality prediction.The DBN,DAE-BP and DBN-ELM air quality evaluation and prediction models based on deep learning established in this paper provide new ideas for air quality prediction,and also provide new theoretical basis and prediction method for the guidance of air pollution treatment and people's daily life.