Applications of SVR-PSO Model and Multivariate Linear Regression Model in PM2.5 Concentration Forecasting

Applications of SVR-PSO Model and Multivariate Linear Regression Model in PM2.5 Concentration Forecasting

Guo-Feng Fan (School of Mathematics &Statistics Science, Ping Ding Shan University, Ping Ding, China), Meng-Qi Liang (School of Mathematics &Statistics Science, Ping Ding Shan University, Ping Ding, China), Jing-Ru Li (School of Economics and Management, Ping Ding Shan University, Ping Ding, China) and Wen-Lu Ma (School of Economics and Management, Ping Ding Shan University, Ping Ding, China)
Copyright: © 2017 |Pages: 17
DOI: 10.4018/IJAEC.2017100105

Abstract

At present, the fog and haze problem is intensified, which has a great impact on the production of enterprises and living of the residents. PM2.5 is an important indicator of air pollution and it also receives much concern. This article collects the reliable data of PM2.5 in the five industrial cities in Henan Province from Weather Report Network, and PM2.5 Data Network since 2015. The effective approaches to forecast PM2.5 concentration is proposed, i.e., the improved multivariate linear regression (namely IMLR) model and support vector regression with particle swarm optimization algorithm (namely SVR-PSO) model. The empirical results demonstrate that the proposed IMLR and SVR-PSO forecasting models are effective, and also, could be an instructive reference for weather quality forecasting, safe travel, and safe production.
Article Preview

2. The Data Sources And Processing

In this paper, the data are collected from the Weather Report and the PM2.5 Data Network, in which five cities of Henan Province, China, are selected. These five selected cities include Zheng Zhou City (the provincial capital of Henan Province), An Yang City, Luo Yang City, Ping Ding Shan City, and Jiao Zuo City; in which, the previous two cities are the industrial cities and the other three cities are belonged to the resource-based cities (Zhang et al., 2015b). The data contents mainly contain PM2.5 concentration and the dependent variable AQI index. The period of the collected data is from 23 April 2015 to 31 July 2015; 24 May 2016 to 31 August 31 2016; and 23 June 2017 to 30 September 30 2017; totally 300 days in the collected data set. In which, 200 days are used as the training set to establish the training model, the other 100 days are used to establish the testing model (Yang et al., 2013). In the training stage, the parameters of an SVR model would be determined with minimal forecasting errors, and the un-used raw data will be employed to calculate the forecasting errors. The testing results has demonstrated that the proposed approach receives more accurate forecasting performances, which could provide valuable and meaningful governance guidance than the simple linear regression model (Lv et al., 2016).

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 10: 4 Issues (2019): Forthcoming, Available for Pre-Order
Volume 9: 4 Issues (2018): 2 Released, 2 Forthcoming
Volume 8: 4 Issues (2017)
Volume 7: 4 Issues (2016)
Volume 6: 4 Issues (2015)
Volume 5: 4 Issues (2014)
Volume 4: 4 Issues (2013)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing