基于贝叶斯准则的偏态分布拟合效果分析

Analysis of Fitting Effect of Skewed Distribution Based on Bayesian Information Criterion

摘要: 正态分布、t分布、双指数分布等概率统计分布已经广泛地应用于社会的各个领域，但是由于实际数据的复杂性，上述分布对数据的拟合并不能很好地表现数据的特征，因此偏态分布逐渐引起了人们的注意。为了进一步拓展偏态分布的应用范围，借助R软件，将偏态分布应用到了台风灾害造成的经济损失数据中，并与常用的分布进行对比分析，利用贝叶斯信息准则评价分布的优劣性。通过研究发现，偏t分布对台风灾害造成的经济损失数据表现出了较好的拟合效果，能够更好的体现数据特征，这无论是对台风灾害的研究还是偏t分布的研究都具有一定的意义。

Abstract: Probability statistical distributions such as normal distribution, t-distribution and double exponen-tial distribution have been widely used in various fields of society. However, due to the complexity of actual data, the data fitting of the above distribution can not well represent the characteristics of the data, so the skewed distribution has gradually attracted people’s attention. In order to further expand the application scope of the skewed distribution, R software is used to apply the skewed distribution to the economic loss data caused by typhoon disasters, and the comparison analysis is made with the commonly used distribution, the advantages and disadvantages of the distribution are evaluated by the Bayesian Information Criterion. Through the study, it is found that the skew-t distribution has a better fitting effect on the economic loss data caused by typhoon disasters, which can better reflect the characteristics of the data, which has certain significance for both the study of typhoon disasters and the study of skew-t distribution.