A Prediction Method of PM2.5 Concentration Based on Game Neural Network

Abstract

A method for predicting PM2.5 concentration values ​​based on a game neural network, comprising the steps of: Step 1, collecting raw data, the raw data including PM2.5 concentration value historical data, PM2.5 concentration value index historical data and meteorological historical data; Step 2, use the generative network to generate simulated data; Step 3, use the discriminant network to judge the authenticity of the simulated data; Step 4, use the game neural network to predict the PM2.5 concentration value. In addition to performing nonlinear correlation analysis on PM2.5 concentration value historical data, PM2.5 concentration value related index historical data and meteorological historical data, the present invention also introduces a generating network to mix original data and noise to output simulated data, and The simulated data is sent to the discriminant network for discrimination, and iterated and adjusted according to the discriminant results. For small data sets, there is no need to preset the target model, and the temporal change law of PM2.5 concentration can be accurately described.

Description

technical field [0001] The invention relates to the technical field of prediction of air particulate matter PM2.5 concentration value, in particular to a method for predicting PM2.5 concentration value based on a game neural network. Background technique [0002] PM2.5 refers to particulate matter with a diameter less than or equal to 2.5 microns in the atmosphere. It is rich in a large amount of toxic and harmful substances and has a long residence time in the atmosphere and a long transportation distance. Therefore, it has a greater impact on human health and the quality of the atmospheric environment. Excessive PM2.5 also brought another impact - haze weather. Air pollution has become the focus of people's attention, and among the air pollution indicators, PM2.5 concentration has become a symbolic detection index to measure air quality. Nowadays, the prediction of PM2.5 concentration value in the future time period based on historical data has become a research problem w...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the shortcomings that the existing PM2.5 concentration value prediction method cannot train small data sets and the target model needs to be defined in advance, the present invention is based on PM2.5 concentration value historical data, PM2.5 concentration value related index historical data and meteorological historical data In addition to nonlinear correlation analysis, a generation network is also introduced to mix the original data and noise to output simulated data, and send the simulated data to the discriminant network for discrimination, and perform iterations and adjustments according to the discriminant results, providing a method for small data A game neural network-based PM2.5 concentration prediction method that can accurately describe the temporal variation of PM2.5 concentration without the need for a preset target model

Images