Bayesian Data Mining for Health Surveillance
Bayesian Data Mining for Health Surveillance
复制标题
DOI：
10.1002/0470092505.ch12
复制
发表时间：
2005-01-01
期刊：
影响因子：
--
通讯作者：
Madigan, David
中科院分区：
文献类型：
--
作者：
Madigan, David
数据挖掘关注从数据中提取有用的知识。统计工具和思想显然是数据挖掘的核心，但由于数据挖掘通常（但并不总是）关注大规模的数据存储库，因此计算问题就显得尤为突出。与统计学教科书相比，数据挖掘教科书描述了算法，并仔细关注可行性和规模问题（例如，参见（Hand et al.，2001年））。“健康监测数据挖掘”是指数据挖掘在健康相关的、观察性的、有时间戳的数据中的应用。许多应用程序还将涉及空间参考数据。监视”对这些数据进行持续的监视，并区分正常情况和某种异常情况。贝叶斯统计分析和数据挖掘方法计算给定观测数据的感兴趣量（如未来可观测量）的条件概率分布。贝叶斯分析通常开始与全概率模型-一个联合概率分布的所有可观察和不可观察的数量研究-然后使用贝叶斯定理计算必要的条件概率分布。事实上，该定理规定了统计学习的基础，
Data mining concerns the extraction of useful knowledge from data. Statistical tools and ideas obviously lie at the core of data mining, but since data mining usually (but not always) focuses on larger-scale data repositories, computing issues come to the fore. Data mining textbooks, by contrast with Statistics textbooks, describe algorithms and pay careful attention to issues of feasibility and scale (see, for example, the outstanding text of (Hand et al., 2001)). By\data mining for health surveillance" I mean applications of data mining to health-related, observational, timestamped data. Many applications will additionally concern spatially-referenced data.\Surveillance" performs ongoing monitoring of such data and discriminates between normal conditions and anomolous conditions of one sort or another.The Bayesian approach to statistical analysis and data mining computes conditional probability distributions of quantities of interest (such as future observables) given the observed data. Bayesian analyses usually begin with a full probability model-a joint probability distribution for all the observable and unobservable quantities under study-and then use Bayes' theorem to compute the requisite conditional probability distributions. In fact, the theorem prescribes the basis for statistical learning in the