Daily PM2.5 concentration prediction based on principal component analysis and LSSVM optimized by cuckoo search algorithm

Increased attention has been paid to PM2.5 pollution in China. Due to its detrimental effects on environment and health, it is important to establish a PM2.5 concentration forecasting model with high precision for its monitoring and controlling. This paper presents a novel hybrid model based on principal component analysis (PCA) and least squares support vector machine (LSSVM) optimized by cuckoo search (CS). First PCA is adopted to extract original features and reduce dimension for input selection. Then LSSVM is applied to predict the daily PM2.5 concentration. The parameters in LSSVM are fine-tuned by CS to improve its generalization. An experiment study reveals that the proposed approach outperforms a single LSSVM model with default parameters and a general regression neural network (GRNN) model in PM2.5 concentration prediction. Therefore the established model presents the potential to be applied to air quality forecasting systems. •PCA is adopted to extract original features and reduce dimension for input selection.•LSSVM model is firstly proposed for PM2.5 concentration prediction.•The novel hybrid model PCA-CS-LSSVM outperforms single LSSVM and GRNN models in terms of prediction precision.•The model presents strong potential to be applied to air quality forecasting systems.