基于贝叶斯公式的诊断模型

从一个颇有争议的诊疗案例出发,基于贝叶斯公式,根据我国目前的医疗体制,建立了一个定量化的诊断模型,给出了诊疗方案合理性的指标,用数据对模型进行了检验,并对模型进行了评价,进一步证实了在我国建立医疗责任保险制度的必要性与紧迫性。     

Bayes分析指导孤立性肺结节的临床决策

目的：探讨Bayes分析指导孤立性肺结节（solitary pulmonary nodules, SPN）临床决策的可行性及其临床意义。方法：利用Bayes分析法先从352例SPN训练集（恶性135例,良性217例）中求出恶性SPN的验前比及各临床和CT表现的似然比,再以此计算每个SPN的恶性概率。比较Bayes分析与医生常规判断132例SPN测试集（恶性61例,良性71例）样本的诊断效能,并分析在不同计算概率下的实际诊断结果。结果：Bayes分析诊断测试集SPN的敏感度、特异度、符合率、阳性预测值及阴性预测值分别为88.5％,85.9％,87.1％,84.4％,89.7％,其诊断符合率与高年资甲、乙医生比较无统计学差异（均P＞0.05）,但高于低年资丙、丁医生（均P＜0.05）;Bayes分析,高年资甲、乙医生及低年资丙、丁医生的Brier值分别为0.099,0.140,0.137,0.154,0.179;除外被错判的11例孤立性肺转移瘤,Bayes分析法估算概率＜20％的假阴性率为1.0％（5/484）。结论：运用Bayes分析诊断SPN性质的符合率高,预测恶性概率的精度高,且假阴性率低,用于指导SPN的临床决策具有一定的可行性。     

Bayesian model search for mixture models based on optimizing variational bounds

When learning a mixture model, we suffer from the local optima and model structure determination problems. In this paper, we present a method for simultaneously solving these problems based on the variational Bayesian (VB) framework. First, in the VB framework, we derive an objective function that can simultaneously optimize both model parameter distributions and model structure. Next, focusing on mixture models, we present a deterministic algorithm to approximately optimize the objective function by using the idea of the split and merge operations which we previously proposed within the maximum likelihood framework. Then, we apply the method to mixture of expers (MoE) models to experimentally show that the proposed method can find the optimal number of experts of a MoE while avoiding local maxima.     

Differentiation of integrals in higher dimensions

We prove a localization principle for directional maximal operators in L^p(ℝⁿ), with p > 1. The resulting bounds, which we conjecture hold for the largest possible class of directions, imply Lebesgue-type differentiation of integrals over tubes that point in the given directions.     

连续肌肉电信号的辨识

以前对肌电信号的处理都是针对手工分割后的信号进行辨识，但实际应用中需要处理的是连续肌电信号。本文中采用了时域参数与频域参数相结合的方法对连续的肌电信号进行自动分割，然后再辨识分割后的肌电信号，取得了很好的结果。     

Testing simultaneous hypotheses in pharmaceutical trials: a bayesian approach

Consider the cohort design and suppose that the outcome of primary interest is a continuous random variable observed repeatedly over time. Suppose that the value of a “clinical marker,” which is thought to be predictive of the primary outcome, is also recorded. We would like to determine whether there is an association between the two variables as they evolve over time. We might also want to predict the pattern for the primary outcome conditionally on a specific profile of clinical interest for the serial marker. A model is developed to address these issues. One regression model is created for the primary outcome while a second regression model is developed for the clinical marker. The vector autoregressive model, i.e., VAR(1), is used to characterize the covariance structure between the two sets of repeated observations. Hypotheses of interest are described and procedures for testing them are elaborated. The Diabetes Control and Complications Trial is used to illustrate the procedures.     

基于Bayes公式的医生辅助诊断系统的研究与开发

近几十年来，随着计算机性能的不断提高，它在医学、生物学等各个领域的应用已越来越受到研究者的重视。其中，计算机辅助诊断已成为众多医学工作者的研究热点。重点介绍了基于Bayes公式医生辅助诊断系统的研究与开发，该系统将成为医生诊断及治疗的重要辅助工具。     

Orientation uncertainty reduces perceived obliquity

The influence of prejudice on perception should be greatest when certainty about stimulus identity is least. We exploited this relationship to reveal visual biases for the cardinal orientations: vertical and horizontal. Specifically, when we increased the variance of orientations in an array of grating patches, estimates of the mean became less oblique. This result is consistent with a stable prior, or prejudice, for those orientations most prevalent in natural scenes.     

Bayes判别分析在海南省疟区分层中的应用

本文试对与疟疾流行有关的社会经济和地理诸因素进行Bayes逐步判别分析,进而建立高、中、低三类疟区的判别函数式,以探索疟区分层的新方法。
1990年在海南省12个县选择了不同流行程度的自然村55个,以自然村为调查单位,进行社会经济因素调查,以其33个自然村作为建模样本,22个自然村为考核样本。用包括12项社会经济和地理诸因素的统一调查表,逐户进行访问调查。将收集的资料输入IBM/PC-XT微机进行Bayes逐步判别分析。结果选出地形地貌、反映经济水平的劳动力比例、人均收入、居住条件,以及疟防知识和人的特殊行为(上山作业过夜者的比例)等6项对疟区分层有判别意义的变量,并以此6项指标建立了代表高、中、低3个不同流行水平的判别函数式。经建模样本的回代检验,并对照原疟区分层,正确判别率为91.0%。又经考核样本的外部检验,正确判别率为77.3%。表明此种判别分析方法在疟区分层中具有实用价值,在特定条件下有可能代替发病率、原虫率、媒介等指标进行疟区分层。     

Passive smoking in the workplace: classical and Bayesian meta-analyses

There are currently several classical and Bayesian methods of meta-analysis available for combining epidemiological results. We describe and compare these in a consistent framework, and apply them to published studies of the relative risk of lung cancer associated with exposure to environmental tobacco smoke in the workplace. We find that although all methods give reasonably similar combined estimates of relative risk of lung cancer associated with this exposure (none of which is significantly raised above unity, in either a frequentist or a Bayesian sense), the approximations arising from classical methods appear to be nonconservative and should be used with caution. The Bayesian methods, which account more explicitly for possible inhomogeneity in studies, give slightly lower estimates again of relative risk and wider posterior credible intervals, indicating that inference from the non-Bayesian approaches might be optimistic.