基于遗传算法的扩展NelsonSiegel模型及实证研究.PDFVIP

26 1 20111 Vol. 26 No. 1 Statistics Information Forum Jan. , 2011 基于遗传算法的扩展NelsonSiegel 模型及实证研究 1 1 2 苏云鹏, 杨宝臣, 李冬连 ( 1. , 300072; 2. , 0 0061) : Nelson- Siegel , , Nelson- Siegel Nelson- Siegel , Nelson- Sie gel , , , , : ; ; Nelson- Siegel ; : F830. 91 : A : 1007- 3116(2011) 01- 001 - 0 : 2010- 08- 07; : 2010- 10- 22 : !H eath- Jarrow - Morton ∀(7077107 ) ; !∀(200800 60032) ; ! ∀ : , , , , , : , ; , , , , , : , ; , , , , : , Maximum a Posteriori Estimate of Hierarchical Linear Model ZHANG Xuan ( School of Statistics, Renmin University of China, Beijing 100084, China) Abstract: Maximum a posteriori ( MA P) estimation is an important method in statistical estimation. This research develops MA P estimation for Hierarchical Linear Model ( HLM) , and gives a detailed process of estimation and computational program. The expectation maximum ( EM) algorithm w as adapted to the concrete calculation, and the second derivatives of log likelihood function help us to obtain the esti mators of variance and covariance components in this model according to asymptotic distribution theory of maximum likelihood estimator. At last, w e compare the results of MAP estimate to the general maximum likelihood ( M L) estimate, noting that EM algorithm is exploited into the tw o different estimates. Simula tion show s that the estimators of fixed effects using the tw o methods are almost consistent, but the MAP estimators of variance and covariance components are closer to real values than the ML estimate counter part w hen the group size is small, at the same time, the iterations needed in M