Research on Time Series Data Mining Based on Intelligent Integrated Particle Swarm Optimization Algorithm
ZHANG Jian;College of Computer Science and Engineering,Sanjiang University;
An intelligent integrated architecture is proposed to address the problem that a single algorithm has the defect that can't dig all information in dealing with complex time series data. Four kinds of integration architecture have been given and their applications have been analyzed. Aiming at the time series data of a class of random noise interference,a series nested modeling structure is adopted,and the autoregressive moving average model of multiple double-population particle swarm optimization algorithm is proposed to dig the randomness trend in data. Meanwhile,The least squares support vector machine( LSSVM) based on probability density functions control( PDF) is proposed to dig the certainty trend in data,the parallel compensation of two models realizes the full excavation of data information. Through a set of experiments,the effect of proposed method is verified.