Nonparametric bayesian lifetime data analysis using dirichlet process lognormal mixture model

We propose a nonparametric Bayesian lifetime data analysis method using the Dirichlet process mixture model with a lognormal kernel. A simulation‐based algorithm that implements the Gibbs sampling is developed to fit the Dirichlet process lognormal mixture (DPLNM) model using rightly censored failure time data. Five examples are used to illustrate the proposed method, and the DPLNM model is compared to the Dirichlet process Weibull mixture (DPWM) model. Results indicate that the DPLNM model is capable of estimating different lifetime distributions. The DPLNM model outperforms the DPWM model in all the examples, and the DPLNM model shows promising potential to be applied to analyze failure time data when an appropriate parametric model for the data cannot be specified. © 2013 Wiley Periodicals, Inc. Naval Research Logistics, 2013     

静磁场Dirichlet问题Galerkin法和牛顿法

作者在［１］,［２］中给出了静磁场Ｄｉｒｉｃｈｌｅｔ问题和Ｎｅｕｍａｎｎ问题广义解的存在与唯一性。本文从理论上给出静磁场Ｄｉｒｉｃｈｌｅｔ问题的一种Ｇａｌｅｒｋｉｎ近似解法,并给出了在各向同性或无电流区情形下的误差估计和牛顿选代序列。     

基于非负张量表示的扩展LDA标注模型

由于“语义鸿沟”的存在,自动图像标注是一项极具挑战性的工作。考虑到图像低层视觉特征与高层语义概念的差异,分别从图像表示与语义建模两个方面来实现自动图像标注。在图像表示方面,提出了一种正则化约束下的非负张量表示方法,用以提取符合人眼视觉直观理解的图像高阶结构特征。在语义建模方面,提出了一种三层贝叶斯模型——扩展隐Dirichlet分配。该模型利用隐变量来实现图像与标注词的关联,并通过一种基于变分推理的期望最大值方法来估计其参数。实验结果表明,ELDA模型在大规模数据库NUS-WIDE上的标注结果相较于现有方法有了显著的提高。     

指数寿命产品可靠性增长试验的Bayes分析

针对指数寿命产品的定时、定数截尾试验方案,推广了Mazzuchi-Soyer模型的应用范围。首先引入模型假设,以狄氏分布作为先验分布,综合利用产品研制的历史信息和专家信息,结合产品研制各阶段试验数据,给出了各阶段可靠性的联合后验分布。然后利用Gibbs抽样算法解决后验推断计算问题,得到各阶段产品可靠性的Bayes点估计和区间估计。最后给出产品可靠性增长分析实例,表明了模型的优越性。     

多特征融合文本聚类的新闻话题发现模型

融合新闻命名实体、新闻标题、新闻重要段落、文本语义等多特征影响,提出基于多特征融合文本聚类的新闻话题发现模型。模型根据新闻的多特征影响,提出一种多特征融合文本聚类方法。该方法针对新闻标题、新闻重要段落等特征因素构建向量空间模型及相似度算法,基于潜在狄利克雷分配模型构建主题空间模型及相似度算法,针对命名实体构建命名实体模型及相似度算法,并将三种相似度算法形成最优融合。基于多特征融合文本聚类方法,模型改进了用于新闻话题发现的Single-Pass算法。实验是在真实新闻数据集上开展的,实验结果表明:该模型有效地提高了新闻话题发现的准确率、召回率和综合评价指标,并具有一定的自适应能力。     

Bayesian strategies for dynamic pricing in e‐commerce

E‐commerce platforms afford retailers unprecedented visibility into customer purchase behavior and provide an environment in which prices can be updated quickly and cheaply in response to changing market conditions. This study investigates dynamic pricing strategies for maximizing revenue in an Internet retail channel by actively learning customers' demand response to price. A general methodology is proposed for dynamically pricing information goods, as well as other nonperishable products for which inventory levels are not an essential consideration in pricing. A Bayesian model of demand uncertainty involving the Dirichlet distribution or a mixture of such distributions as a prior captures a wide range of beliefs about customer demand. We provide both analytic formulas and efficient approximation methods for updating these prior distributions after sales data have been observed. We then investigate several strategies for sequential pricing based on index functions that consider both the potential revenue and the information value of selecting prices. These strategies require a manageable amount of computation, are robust to many types of prior misspecification, and yield high revenues compared to static pricing and passive learning approaches. © 2006 Wiley Periodicals, Inc. Naval Research Logistics, 2007     

结合非负张量表示与扩展隐Dirichlet分配模型的图像标注

由于"语义鸿沟"的存在,自动图像标注是一项极具挑战性的工作。考虑到图像低层视觉特征与高层语义概念的差异,分别从图像表示与语义建模两个方面来实现自动图像标注。在图像表示方面,提出了一种正则化约束下的非负张量表示方法,用以提取符合人眼视觉直观理解的图像高阶结构特征。在语义建模方面,提出了一种三层贝叶斯模型——扩展隐Dirichlet分配。该模型利用隐变量来实现图像与标注词的关联,并通过一种基于变分推理的期望最大值方法来估计参数。实验结果表明,ELDA模型在大规模数据库NUS-WIDE上的标注结果相较于现有方法有了显著的提高。