知识服务中基于WKNN-MF与NRS的非完备案例知识匹配OA
Incomplete Case Knowledge Matching Based on WKNN-MF and NRS in Knowledge Service Process
为改善知识服务效益、解决非完备案例知识属性值缺失的问题,设计了基于 WKNN-MF 与 NRS 的知识匹配方法.该方法利用 WKNN 算法对非完备案例知识集实施预填充,作为缺失森林算法相应计算的基础,并由 WKNN-MF 获得完备案例知识集;在此基础上,对案例视图进行 NRS 约简,剔除冗余属性;进而基于知识视图相似度搜寻类似案例并确定最终结果.基于 3 个不同 UCI 数据集的实验结果表明,相较于既有数据填补算法,本方法对非完备案例知识集具有更优的填补实效.
To improve the benefits of knowledge services and solve the problem of missing attribute values of incomplete case knowledge,a knowledge matching method based on WKNN-MF and NRS was de-signed.The WKNN algorithm was used to pre-populate the non-complete case knowledge set as the basis of MF calculation,and WKNN-MF was used to obtain the complete case knowledge set.On this basis,NRS simplification was performed to eliminate redundant attributes,and then similar cases were searched by the similarity of knowledge views to determine the knowledge matching results.Experimental results based on three different UCI datasets showed that the proposed method had a better filling effect on in-complete case knowledge sets compared with existing data-filling algorithms.
张建华;曹子傲;周晓倩;温丹丹;贺龙飞
郑州大学 管理学院 河南 郑州 450001郑州大学 管理学院 河南 郑州 450001郑州大学 管理学院 河南 郑州 450001郑州大学 管理学院 河南 郑州 450001郑州大学 管理学院 河南 郑州 450001
信息技术与安全科学
知识服务知识匹配非完备案例知识加权最近邻缺失森林邻域粗糙集
knowledge serviceknowledge matchingnon-complete case knowledgeweighted KNNmissing forestNRS
《郑州大学学报（理学版）》 2026 (3)
77-85,9
国家社会科学基金项目(19BTQ035)
评论