| Air is the basis for people’s survival.However,with the continuous advancement of science and technology,air pollution has become more and more serious.Air quality problems have become a key factor for China’s green and high-quality development.Because of its irregular and non-linear characteristics,air quality cannot be predicted by a simple mathematical model.An effective prediction of air quality can have an anchor to windward.Therefore,building a high-accuracy and fast air quality prediction model has positive and far-reaching effects on people’s daily life,social stability and well-being,and national development and construction.This paper proposes a CNN-LSTM-Attention model that is mainly used to predict the air quality index in the next day.The model is divided into three parts.First,convolutional neural network(CNN)is used to extract the advanced features that affect the air quality index,and then the data processed by CNN is extracted through the LSTM layer to extract time features,finally,the attention layer is used to extract the impact of data features at different times in the temporal data on the predicted values,and intervention is carried out through weighting,after the steps,can predict the air quality index.The degree of influence of data characteristics at different times in the time series data on the predicted value is intervened by weighting,and finally the air quality index is predicted.In order to verify the effect of the model,this paper selects the daily air pollution data set of Lanzhou City from 2013 to 2022 to train and test the model,and compares the prediction results based on the CNN-LSTM-Attention model with the CNN-LSTM model,the LSTM model and ARIMA model,through three evaluation indicators.The results show that CNN-LSTM-Attention has higher prediction accuracy and better model performance.Therefore,the model can provide an effective reference for people’s daily travel and the government to improve air quality. |
