Least Squares Support Vector Prediction for Daily Atmospheric Pollutant Level

Abstract

Multi-layer perceptrons (MLP) have been employed to solve a variety of problems. The practical applications of MLP however suffer from different drawbacks such as local minima and over-fitting, such that good generalization may not be obtained. Least squares support vector machines (LS-SVM), a novel type of machine learning technique based on statistical learning theory, can be used for regression and time series prediction. In this study, meteorological and pollutions data are collected daily at monitoring stations of a city. This pollutant-related information can be used to build an early warning system, which provides forecast and also alarms health advice to local inhabitants by medical practicians and local government. Through experiment, we found that LS-SVM could overcome most of the drawbacks of MLP and had been reported to show promising results.