Clustering and Classification Using Mixture Models: Towards Big Data

使用混合模型进行聚类和分类：迈向大数据

• 批准号: RGPIN-2017-05255
• 负责人: McNicholas, Paul
• 金额: $ 3.13万
• 依托单位: McMaster University
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2017
• 资助国家: 加拿大
• 起止时间: 2017-01-01 至 2018-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=636037
• 关键词: Clustering; Classification; Using; Mixture; Models

Modern scientific research produces ever-growing masses of data. Considering that scientists rely on statistics to interpret results, the availability of suitable statistical methods is essential. Unfortunately, some research questions in the big data era cannot be adequately answered due to the lack of appropriate statistical techniques. The proposed research will develop novel approaches to the labelling of observations as belonging to one of a number of groups; such approaches are called clustering or classification, depending on whether or not some of the points are labelled. The techniques that will be developed concern web streaming data and "long" longitudinal data. Furthermore, some work will be devoted to discovering a gold standard for classification using mixtures. These techniques, once developed, will be released for use by the broadest possible research community through packages for the R software. This work will extend research on model-based clustering and classification towards big data. The impact of this work will be felt in many disciplines in fields as diverse as biology, food science, security, and sociology.

现代科学研究产生了不断增长的大量数据。考虑到科学家依靠统计数据来解释结果，适当的统计方法的可用性是必不可少的。不幸的是，由于缺乏适当的统计技术，一些大数据时代的研究问题无法得到充分的回答。拟议的研究将开发新的方法来标记属于若干组之一的观察结果；这种方法被称为聚类或分类，取决于是否有一些点被标记。将开发的技术涉及网络流数据和“长”纵向数据。此外，一些工作将致力于发现使用混合物进行分类的黄金标准。这些技术一旦开发出来，将通过R软件的软件包发布给尽可能广泛的研究社区使用。这项工作将把基于模型的聚类和分类研究扩展到大数据。这项工作的影响将在生物学、食品科学、安全和社会学等领域的许多学科中感受到。