At present,it is a challenge to present and analyz air quality prediction results from spatio-temporal perspective.Therefore,combining deep learning and big data analysis techniques,an interac-tive visual analysis method is proposed for multidimensional representation of air quality change dy-namics based on neural network prediction.Firstly,a multilayer two dimensional convolutional neural networks and long short-term memory neural network model(2D-CNN+LSTM)is proposed for extract-ing air quality spatio-temporal features for Air Quality Index(AQI)prediction.Secondly,from the prac-tical requirements,a visualization view and interaction method is designed for multidimensional repre-sentation of air quality situation information.Finally,a visual analysis system is constructed to utilize the air pollution dataset for case studies and analysis.The experimental results show that it can realize the spatial and temporal multidimensional expression and analysis of air quality situation through the synergistic interaction of multiple views such as time-series visualization,spatial visualization and at-tribute association visualization.It provides new ideas and methods for air quality prevention and man-agement problems.
air quality predictionspatio-temporal situation analysisair qualityvisual analysisneural network
国家科技期刊平台
NSTL
万方数据
