PM2.5 concentration prediction method based on data space-time characteristics

Abstract

The invention discloses a PM2.5 concentration prediction method based on data space-time characteristics. The method comprises the steps of collecting PM2.5 and weather data of all monitoring stations; cleaning all the acquired data; carrying out correlation analysis, setting a threshold value, and fusing data of the to-be-predicted station and data of adjacent stations with correlation coefficients larger than or equal to the threshold value to construct model input data; normalizing the input data, and sliding and segmenting the data sequence according to a time step length T to obtain a plurality of data components; building a PM2.5 concentration prediction model based on a CNN capable of extracting data space features and a bidirectional LSTM capable of extracting time features; training the model by using the training set, and optimizing by using a grid search method to obtain an optimal parameter; and carrying out PM2.5 prediction by using the trained model. Compared with the prior art, the model provided by the invention has the advantages that the prediction stability is improved while the prediction precision is improved, and the robustness to abnormal values is higher.

Description

technical field [0001] The patent of the present invention relates to a PM2.5 concentration prediction method, in particular to a PM2.5 concentration prediction method considering the temporal and spatial characteristics of data. Background technique [0002] In recent years, the problem of environmental pollution has received extensive attention. Especially the air pollution caused by PM2.5 has been and will continue to be a major health hazard to be solved for a long time in the future. Living in a heavily polluted environment for a long time, the human respiratory system, cardiovascular system, and reproductive system will gradually develop pathological changes. At the same time, these air pollutants can scatter and absorb visible light, reducing the visibility of the atmosphere and causing more traffic accidents, which also affect people's normal life. Therefore, accurate and stable air quality prediction is very important for regional early warning and reducing safety...

AI Technical Summary

Problems solved by technology

However, these models are only applied to the data research of a single monitoring station, and due to the high degree of circulation in the atmosphere, there is generally a certain relationship between the pollutant data of multiple adjacent stations except for individual mutations at certain moments , these models do not make good use of this connection when predicting

Images