Semantic web framework for software system design, testing and maintenance for syndromic surveillance applications

综合症监测是利用诸如急诊就诊、远程保健电话、非处方药销售、实验室报告、人口信息等数据流监测人口健康状况。这种监测的一项主要活动是使用畸变检测算法提醒公共卫生官员注意可能的疾病暴发情况。虽然这些系统已经成功地应用于现场，但它们还没有被证明是容易扩展的，因此许多新的和有趣的像差检测算法没有得到测试和部署。该研究计划正在开发一个框架，以帮助使用语义网的概念和工具开发畸变检测软件系统。该研究将以本体的形式概念化信息系统开发的整个过程，用于算法的选择、组成、部署、测试和评估。由此产生的本体论和相关推理系统将有助于将信息系统开发过程抽象为公共卫生领域专家（或者实际上，任何不具备软件专业知识的领域专家）可以访问的描述形式。将围绕随机算法（如人工神经网络（ann））的情况进行概念验证。人工神经网络通常被认为是复杂、神秘的，对最终用户来说太难处理。人工神经网络的设计、部署、测试和评估有许多不同于其他类型算法的阶段，信息系统开发系统可以使用这些阶段的本体论描述来正确地将人工神经网络方法集成到畸变检测系统中。还可以开发一种算法实施阶段的进一步本体论描述，以方便监视系统以对用户透明的方法使用分布式计算资源。再一次，这是对最终用户来说过于复杂的另一个领域（网格或云计算），但在扩展许多应用程序（特别是生命科学领域）的计算能力方面有很大的潜力，这些应用程序通常没有利用可用的计算资源。

Syndromic surveillance is the monitoring of the health of a population using datastreams such as emergency department visits, calls to telehealth, over-the-counter drug sales, laboratory reports, demographic information, etc. One major activity in this surveillance is the use of aberration detection algorithms to alert public health officials to a possible disease outbreak situation. While these systems have been successfully used in the field, they have not proven to be easy to extend and thus many new and interesting aberration detection algorithms are not being tested and deployed. This research program is developing a framework to aid in the development of aberration detection software systems using the concepts and tools of the semantic web. The research will conceptualize the entire process of information system development in the form of an ontology for algorithm selection, composition, deployment, testing and evaluation. The resulting ontology and related reasoning systems will facilitate the abstraction of the information system development process into a form of description that is accessible to the public health domain expert (or in fact, any domain expert who does not have software expertise). A proof of concept will be developed around the case of stochastic algorithms such as artificial neural networks (ANNs). ANNs are often perceived as complex, mysterious and too difficult to handle for the end user. There are many stages in the design, deployment, testing and evaluation of ANNs that are distinct from other types of algorithms and an ontological description of this could be used by an information system development system to correctly integrate ANN methods into an aberration detection system. A further ontological description of the implementation phase of an algorithm could also be developed that would facilitate the use of distributed computing resources by a surveillance system in a method that is transparent to the user. Once again, this is another area (Grid or Cloud Computing) that is overly complex for the end user but has great potential to extend the computational capability of many applications (particularly in the life sciences) that often do not take advantage of available computational resources.