With the development of technology in data collection and storage, new types of higher order tensorial information streams are available in chemical and biological manufacturing processes, which contain valuable information about the process condition and product quality. However, tensorial data have not been fully utilized yet and the application of tensorial data analytics to manufacturing processes has not been thoroughly investigated. In this article, different types of higher order data in manufacturing processes are described, and their potential usage is addressed. Then some perspectives are provided on the application of tensorial data analytics to manufacturing processes, with an emphasis on multilinear subspace learning problems. In particular, the most representative multilinear subspace learning methods are reviewed. Looking into the future, the potential and research needs for tensorial data analytics are briefly discussed.

随着数据收集和存储技术的发展，化学和生物制造过程中可以使用新型的高阶张量信息流，其中包含有关过程条件和产品质量的宝贵信息。然而，张量数据尚未得到充分利用，并且张量数据分析在制造过程中的应用还没有得到充分研究。在本文中，描述了制造过程中不同类型的高阶数据，并探讨了它们的潜在用途。然后提供一些关于张量数据分析在制造过程中的应用的观点，重点是多线性子空间学习问题。特别是，回顾了最具代表性的多线性子空间学习方法。