PM2.5 hour concentration prediction method and system fusing SSAE deep feature learning and LSTM network

Abstract

The invention relates to a PM2.5 hour concentration prediction method and system fusing SSAE deep feature learning and an LSTM network. The method comprises steps that an SSAE-LSTM (Spatial Spatial Absorption Emphasis-Long Short Term Memory) model is constructed; an air pollutant time sequence with a certain time step length is input into the model; an SSAE network is adopted to extract abstract features of input data through an unsupervised method, the extracted features serve as input features of an LSTM network, feature distribution of air pollutant information within a certain time step length is obtained, and finally the PM2.5 hour concentration is predicted in combination with a full-connection network. According to the invention, the prediction accuracy can be effectively improved.

Description

technical field [0001] The present invention relates to PM 2.5 In the field of prediction, especially a PM that integrates SSAE deep feature learning and LSTM network 2.5 Hourly concentration prediction method and system. Background technique [0002] In recent years, PM 2.5 Haze pollution, which is an important component, is highly frequent and has brought serious harm to human life and the ecological environment. Taking into account the timeliness and dynamics of the atmospheric environment, to achieve accurate PM 2.5 The hourly concentration prediction can effectively improve the forecast and early warning ability of air pollution, and it is also an important research direction of air quality forecast and prevention. [0003] At present, air pollutant concentration prediction models can be roughly divided into two types: mechanism model and non-mechanism model. Among them, the mechanism model predicts the concentration of pollutants by simulating the diffusion proces...

AI Technical Summary

Problems solved by technology

In the non-mechanistic model, the long-short term memory (LSTM) neural network with temporal memory is used in PM 2.5 Some results have been achieved in prediction, but there are still two problems: 1) Some existing methods only consider the influence of meteorological factors on PM 2.5 or only consider the impact of air pollutants on PM 2.5 effect, ignoring the PM 2.5 spatial correlation without PM 2.5 2) Existing methods select PM from a large number of influencing characteristics according to certain evaluation criteria or empirical knowledge. 2.5 Features with important influence, it is difficult to select the optimal combination

Images