Fine particle prediction and traceability method based on the combination of Gaussian process regression and firefly algorithm

Abstract

The invention belongs to the fine particle detection field, in particular to a fine particle prediction and traceability method based on the combination of Gaussian process regression and firefly algorithm. The method comprises the following steps of optimizing the firefly algorithm to obtain a firefly optimization algorithm; creating a sample training set of the model and selecting a covariance function for Gaussian process regression model; carrying out the parameter optimization on the equation of the Gaussian process regression model by using an improved particle swarm optimization algorithm; according to the optimized super-parameters, obtaining the model and outputting the prediction. The invention provides the fine particle prediction and traceability method based on the combinationof Gaussian process regression and firefly algorithm which combines the Gaussian process regression with the firefly algorithm, can not only predict the spatial distribution and short-term variationtrend of fine particles, but also trace back the multi-pollution sources dynamically, thereby accurately locating the multi-pollution sources of fine particles.

Description

technical field [0001] The invention belongs to the field of fine particle detection, and in particular relates to a fine particle prediction and traceability method based on the combination of Gaussian process regression and firefly algorithm. Background technique [0002] In recent years, environmental problems brought about by economic development have become increasingly prominent, and air quality has become the focus of public attention. Among them, fine particles can penetrate into the alveoli and blood circulation system of the human body, causing great harm to human health. Facing the serious problem of fine particle pollution, government departments at all levels have actively adopted comprehensive control measures and achieved certain results. However, smog control is a long process. The relevant departments not only need a long-term control plan, but also need a scientific and quantitative prevention and real-time monitoring mechanism to achieve both symptoms and...

AI Technical Summary

Problems solved by technology

[0006] In summary, the existing technology has problems such as uncertain application conditions, cumbersome parameter setting, long time consumption, large error, and inability to deal with small samples.

Images