点击次数：

发表时间：2017-01-01

发表刊物：自动化学报

所属单位：电子信息与电气工程学部

期号：6

页面范围：944-954

ISSN号：0254-4156

摘要：It is of great significance to predict production process indicators in process industry for production scheduling,safety production and energy saving.Currently,various data-driven approaches for predicting these indicators are proposed,including the following three aspects:feature selection,prediction model construction and model parameter optimization.This paper surveys the above three aspects and summaries the merits and demerits of these approaches.Finally,future research directions of production process prediction of key indicators in process industry are suggested with respect to industrial big data and knowledge automation.

备注：新增回溯数据