Application of Chaotic Time Series Prediction in Forecasting of Library Borrowing Flow

Abstract

Prediction of library borrowing flow plays an important role in controlling the quality of book collection and cataloging work in library. Traditional time series prediction methods are hard to model the library borrowing flow because it is a nonlinear dynamical process and has nonstationary and stochastic character. Based on support vector machine and the theory of chaotic time series prediction, a new method is proposed to model and predict the library borrowing flow. The experiments show this method is reasonable to solve the nonlinear problem in library borrowing flow and is of certain value in both theory and practice.