Abstract:
raditional approach to predictive modeling of large-scale sequential curves is to build model separately according to every curve, which causes heavy and complicated modeling work inevitably. Therefore the existing approach is la ck of manipulation in the application. This paper proposes a new method to solve this problem. By reducing model types of curves, clustering curves and modeling by clusters, the new method simplifies modeling work to a large extent and rese rves original information as much as possible in the meantime. This paper specif ies the theory and algorithm, and applies it to predict GDP curves of multi-reg ions, which confirms practicability and validity of the presented approach.