CAJ | 학술논문

在现代金融数据分析中，金融收益率数据的一个重要性质---在较长时间服从同一类分布---与收益率数据的其他典型特征：非对称性，尖峰厚尾，波动群聚性等有着同样的重要性。本文研究了方差伽玛分布的可加性及其在金融数据分析上的应用，并运用其验证了金融数据收益率数据满足上面的性质。作者采用贝叶斯方法对分布的参数进行了估计，并与R软件中的ghyp程序包的结果进行了比较，最后也对贝叶斯方法产生的Markov链的收敛性运用CODA软件进行了诊断。
재현대금융수거분석중，금융수익솔수거적일개중요성질---재교장시간복종동일류분포---여수익솔수거적기타전형특정：비대칭성，첨봉후미，파동군취성등유착동양적중요성。본문연구료방차가마분포적가가성급기재금융수거분석상적응용，병운용기험증료금융수거수익솔수거만족상면적성질。작자채용패협사방법대분포적삼수진행료고계，병여R연건중적ghyp정서포적결과진행료비교，최후야대패협사방법산생적Markov련적수렴성운용CODA연건진행료진단。
There is an important property in today’s financial analysis ,which is that data follow a distri-bution in a longer term .Along with other topical characteristics :asymmetry ,excess kurtosis and vola-tility clustering ,the property plays an important role .T his paper investigates the additivity of variance Gamma distribution and its application in analyzing data ,and proves the property above .Bayesian meth-od is employed for the estimation of the parameters of the distribution .T he result from Bayesian analy-sis is compared with ghyp package in software R .Extensive convergence diagnostics for the chain ob-tained from BUGS are performed by using the CODA software .