Font Size: a A A

Monte Carlo Tests For Multivariate Mixture Normality

Posted on:2007-12-02Degree:MasterType:Thesis
Country:ChinaCandidate:W H YanFull Text:PDF
GTID:2120360182999202Subject:Probability theory and mathematical statistics
Abstract/Summary:PDF Full Text Request
In actually ,including biology statistics and econometrics data is not necessarily subject to the normal distribution .This data is often assumed that from the mixture distribution .For example, Morgan Bank adopted assumed the rate of return belongs to the mixture normal distribution to measuring the market risk.Since mixture normal distribution is the non-mixed ,the estimation and test have many difficulties .Most testing methods for normal distribution can not apply to the testing for mixture distribution. This paper is based on the empirical characteristic function and Monte Carlo test solution this problem,and a lot of simulations proved the practical of this measure.
Keywords/Search Tags:Multivariate Mixture Normality, Monte Carlo tests, Empirical characteristic function, The test statistic, E-M algorithm Estimation
PDF Full Text Request
Related items