Journal of Jishou University(Natural Sciences Edition)

• Mathematics • Previous Articles     Next Articles

Prediction of Stock Index Based on Fractal Interpolation and SVM

LI Hong,WANG Hongyong   

  1. (School of Applied Mathematics,Nanjing University of Finance and Economics,Nanjing 210023,China)
  • Online:2018-05-25 Published:2018-06-27


In order to better analyze and predict the short-term trend of stock index time series,we propose a new method to determine the free parameters of fractal interpolation,and establish an improved fractal interpolation model.This model is the combined with the support vector machine model to establish a mixed prediction model.The daily closing data of Shanghai composite index is selected as the research object which is shown to have long-range dependence thorugh R/S analysis.The time series of Shanghai composite index are analyzed and predicted by the mixed prediction model.The empirical results show that and the new mixed model proposed in this paper has good fitting performance and  higher  accuracy in short-term prediction.

Key words: fractal interpolation;SVM , model;stock index series;prediction

WeChat e-book chaoxing Mobile QQ