Abstract: Data mining based on rough set theory in process industry is applied to prediction of product qualities. The original data mining model based on the rough set is improved to be suitable for process industry data that would be of large dimension, complexity and continuous in time. Combined with FCM, a method of data mining technique based on the rough set theory in process industry is designed. The method obtains good performance in the experiment on the real industrial data of a practical acetone refining process.

Key words: Rough set, Data mining, Acetone refining, Product prediction

摘要： 将基于粗糙集理论的数据挖掘方法应用于丙酮精制过程产品质量的预报。针对流程工业数据高维、构成复杂、连续性强等特点，改进了基本的粗糙集数据挖掘算法，并与模糊聚类等技术相结合，提出了一种适用于流程工业数据的粗糙集数据挖掘方法。在采用实际丙酮精制生产数据作为样本的实验中应用效果良好，表明该方法具有一定的实用价值。

关键词: 粗糙集, 数据挖掘, 丙酮精制, 产品质量预报