论文部分内容阅读
水雾化粉末质量预报系统具有原始数据不全的特点,是一个不完备的信息系统。本文依据此特点,结合变精度粗糙集方法,在相容关系的基础上引入集对分析方法,提出一种基于集对势容差关系的扩充粗糙集方法,并应用于该信息系统的建模中,通过对α的调节和控制,有效保证了制粉生产条件属性容差类划分的准确性和灵活性。最后,本文以该模型为基础,使用基于属性重要性的贪心算法讨论了制粉生产条件属性约简和知识获取的方法、步骤和决策规则。经实际生产验证了获取的预报知识的合理性和有效性,对制粉工艺和生产真正起到了优化作用。
Water atomized powder quality forecast system with incomplete original data is an incomplete information system. Based on this feature and combined with the variable precision rough set method, this paper introduces a set pair analysis method based on the compatibility relationship, and proposes an extended rough set method based on the tolerance of the set pairwise potential. The method is applied to the modeling of the information system , Through the regulation and control of α, effectively ensure the accuracy and flexibility of the tolerance classification of the properties of milling production conditions. Finally, based on the model, this paper discusses the methods, steps and decision rules of attributes reduction and knowledge acquisition of milling production conditions using the greedy algorithm based on attribute importance. The actual production has verified the reasonableness and effectiveness of the acquired forecasting knowledge and has played an optimal role in milling technology and production.