Font Size: a A A

A Study Of The Share Market-timing Model Based On Riemannian Kernel Function Support Vector Machine Algorithm

Posted on:2017-05-30Degree:MasterType:Thesis
Country:ChinaCandidate:S HuangFull Text:PDF
GTID:2359330566456100Subject:Applied Mathematics
Abstract/Summary:PDF Full Text Request
Share market timing refers to predicting the future general trend in some ways,thus deciding the rise,fall or consolidation.The decisions of buying or selling financial assets are made when there is a rising or falling trend respectively.If there is a market fluctuation,then the financial assets are sold in a high price and bought in a low price.The essence of market timing is a classification problem of predicting the rise or fall of the big board in the future by using the historical data.Support Vector Machine(SVM)is a machine learning method based on statistical learning theory.SVM resolves the small sample,nonlinearity,high dimension and local minimum problemsperfectly and has good generalization ability.The classification ability of SVM is very helpful to the share market timing question.How to choose the kernel function is the key to SVM algorithm.In this paper,we discuss the theory,algorithm and nature of SVM.By introducing the information geometry theory,we constructthe Riemann kernel function,improve the geometric structure of SVM and realize the improving of SVM algorithm.At the same time we establish the Riemann kernel function model and analyzethe historical data of stock market.Finally we propose the improvement scheme.The main innovation work in this paper is the transformation of the SVM kernel function from information geometry.By introducing a Riemann kernel function,we improve the SVM model and establish the simulation experiment on the stock market of CSI 300 index.The simulation results show that Riemann proposed kernel function SVM model has more profits than traditional SVM algorithm.
Keywords/Search Tags:Data mining, Share time-selecting, Riemannian kernel function, support vector machine
PDF Full Text Request
Related items