Air quality prediction SVM model construction method

Abstract

The invention discloses an air quality prediction SVM (Support Vector Machine) model construction method. The method comprises the following steps: collecting air quality data, meteorological data and pollution source continuous emission data; the collected data is processed through sumif function calculation; model variables are constructed, air quality data are processed into conventional air quality variables through an In function, meteorological data are processed into conventional meteorological parameter variables through calculation, pollution source emission data are weighted, and pollution source pollutant emission variables are introduced into pollution sources through calculation processing; and a prediction model is established, the model is established by adopting an SVM method, modeling is carried out, and the model is subjected to trial operation after modeling is completed. According to the invention, big data and information of the existing Internet of Things and Internet are fully utilized, big data statistical analysis thinking and methods and tools are innovated, the urban air quality management decision principle is served, high-level innovation is made in the aspect of scientific research, high-level researchers, doctors and basic-level professionals are cultivated, and the scientific research efficiency is improved. And a statistical prediction diagnosis technical support is provided for air quality management of a heavily polluted area.

Description

technical field [0001] The invention relates to the technical field of air quality prediction, in particular to a method for constructing an air quality prediction SVM model. Background technique [0002] Most of the existing literature on air quality prediction adopts the neural network method to model. From the perspective of explanatory variable selection, most studies only considered the impact of meteorological factors on the concentration of monitoring points, and no research has considered the impact of pollution sources on the concentration of monitoring points. Zhou Shuhua (2017) used the stepwise regression analysis method to establish a statistical forecast model of PM2.5 concentration in different seasons and daily scales in Yibin City within two years, and comprehensively analyzed the relationship between PM2.5 concentration and the concentration of six pollutants in the previous day. At the same time, the relationship between PM2.5 and meteorological elements ...

AI Technical Summary

Problems solved by technology

However, the above studies did not try to incorporate pollution sources into the model, and the significance of the model for the control and planning of pollution sources is out of the question.

Images