Comprehensive learning particle swarm optimization enabled modeling framework for multi-step-ahead influenza prediction

Epidemics of influenza are major public health concerns. Since prediction always relies on the weekly clinical or laboratory surveillance data, typically Influenza-like illness (ILI) rate series, accurate multi-step-ahead predictions using ILI series is great importance, especially, to potential coming outbreaks. This study proposes Comprehensive Learning Particle Swarm Optimization based Machine (CLPSO-ML) framework incorporating support vector regression (SVR) and multilayer perceptron (MLP) for prediction. A comprehensive examination comparison performance three commonly used modeling strategies, including iterated strategy, direct strategy multiple-input multiple-output (MIMO) was conducted from both Southern Northern China. The results show that: (1) MIMO achieves best prediction, potentially more adaptive longer horizon; (2) demonstrates special potentials deriving least time difference between occurrence predicted peak value true an outbreak; (3) For in China, SVR model implemented with performs best, also shows remarkable especially during outbreak periods; while MLP models have competitive