The Application of The Kernel Density Estimates in Predicting VaR
Linjun Tang · Shuxue de shijian yu renshi · 2005
After studying the theory of kernel density,propose a kind of Laplace Kernel Function,which adapts estimating the distribution of finance time series.Based on single variable kernel density estimator,constructs forecast model of VaR,and by weight variation coefficient get two weighted forecast model of VaR.Finally,Using Shanghai′s securities index tests 3 models,shows weighted methods higher efficiency.