《Machine Learning_ A Bayesian and Optimization Perspective》 作者：Sergios Thedoridis

贝叶斯网络参数学习 课程项目-COL884（Spring'18）：人工智能的不确定性 创作者：Navreet Kaur [2015TT10917] 客观的： 警报贝叶斯网络给定数据的贝叶斯参数学习，每行最多有一个缺失值。 使用的算法： 期望最大化 目标： 这项任务的目的是获得学习贝叶斯网络的经验，并了解它们在现实世界中的价值。 设想： 医学诊断。 一些医学研究人员创建了贝叶斯网络，该网络对（某些）疾病和观察到的症状之间的相互关系进行建模。 作为计算机科学家，我们的工作是根据健康记录来学习网络的参数。 不幸的是，在现实世界中，某些记录缺少值。 我们需要尽力计算网络参数，以便以后可以将其用于诊断。 问题陈述： 我们得到了由研究人员创建的贝叶斯网络（如BayesNet.png所示），注意此处对八种诊断进行了建模：血容量不足，左心衰竭，过敏React，镇痛不足，肺栓塞，插管，弯管和断线。

Bayesian methods are increasingly becoming attractive to researchers in many fields. Econometrics, however, is a field in which Bayesian methods have had relatively less influence. A key reason for this absence is the lack of a suitable advanced undergraduate or graduate level textbook. Existing Bayesian books are either out-dated, and hence do not cover the computational advances that have revolutionized the field of Bayesian econometrics since the late 1980s, or do not provide the broad coverage necessary for the student interested in empirical work applying Bayesian methods. For instance, Arnold Zellner’s seminal Bayesian econometrics book (Zellner, 1971) was published in 1971. Dale Poirier’s influential book (Poirier, 1995) focuses on the methodology and statistical theory underlying Bayesian and frequentist methods, but does not discuss models used by applied economists beyond regression. Other important Bayesian books, such as Bauwens, Lubrano and Richard (1999), deal only with particular areas of econometrics (e.g. time series models). In writing this book, my aim has been to fill the gap in the existing set of Bayesian textbooks, and create a Bayesian counterpart to the many popular non-Bayesian econometric textbooks now available (e.g. Greene, 1995). That is, my aim has been to write a book that covers a wide range of models and prepares the student to undertake applied work using Bayesian methods. This book is intended to be accessible to students with no prior training in econometrics, and only a single course in mathematics (e.g. basic calculus). Students will find a previous undergraduate course in probability and statistics useful; however Appendix B offers a brief introduction to these topics for those without the prerequisite background. Throughout the book, I have tried to keep the level of mathematical sophistication reasonably low. In contrast to other Bayesian and comparable frequentist textbooks, I have included more computer-related material. Modern Bayesian econometrics relies heavily on the computer, and developing some basic programming skills is essential for the applied Bayesian. The required level of computer programming skills is not that high, but I expect that this aspect of Bayesian econometrics might be most unfamiliar to the student brought up in the world of spreadsheets and click-and-press computer packages. Accordingly, in addition to discussing computation in detail in the book itself, the website associated with the book contains MATLAB programs for performing Bayesian analysis in a wide variety of models. In general, the focus of the book is on application rather than theory. Hence, I expect that the applied economist interested in using Bayesian methods will find it more useful than the theoretical econometrician.

逻辑回归matlab 代码 2018-MLSP-sparse-bayesian-logistic-regression Matlab code to reproduce some of the results of the paper. Maxime Vono, Nicolas Dobigeon, Pierre Chainais, , Proc. of MLSP, 2018. Copyright Copyright (c) 2018 Maxime Vono.

Machine Learning A Bayesian and Optimization Perspective.pdf

Karl-Rudolf Koch Introduction to Bayesian Statistics Second, updated and enlarged Edition

The Bayesian method is the natural approach to inference, yet it is hidden from readers behind chapters of slow, mathematical analysis. The typical text on Bayesian inference involves two to three chapters on probability theory, then enters what Bayesian inference is. Unfortunately, due to mathematical intractability of most Bayesian models, the reader is only shown simple, artificial examples. This can leave the user with a so-what feeling about Bayesian inference. In fact, this was the author's own prior opinion.

结构模态识别+多测点频域贝叶斯+快速Bayesian

贝叶斯反演 贝叶斯统计，尤其是贝叶斯逆问题中使用的算法的集合

针对现有剩余寿命预测研究中需要多个同类设备历史数据离线估计模型参数的问题，本文提出了一种基于退化数据建模的服役设备剩余寿命自适应预测方法． 该方法，利用指数随机退化模型来建模设备的退化过程，基于退化监测数据运用Bayesian 方法更新模型的随机参数，进而得到剩余寿命的概率分布函数及点估计． 区别于现有方法，本文方法基于设备到当前时刻的监测数据，利用期望最大化算法对模型中的非随机未知参数进行在线估计，由此.无需多个同类设备历史数据． 最后，通过数值仿真与实例分析，验证了本文方法在剩余寿命预测时的有效性．