摘要
Abstract
Considering the strong correlation and weak independence of measurement indicators in public healthcare perform-ance,the performance evaluation indicators have ambiguity.Traditional evaluation methods based on keyword similarity may be misled under the interference of ambiguity,resulting in ineffective removal of redundant evaluation indicators and poor indicator mining results.Design a hospital performance evaluation indicator mining method based on knowledge graph and collaborative filtering.By using the historical performance evaluation indicators of the hospital,a data logic analysis model is constructed,and the semantic correlation of medical performance evaluation indicators is excavated through knowledge graph to fill the information gap between projects caused by semantic gaps during search.By using collaborative filtering to remove subjective evaluation indicators,and relying on complementary semantic filtering to filter out redundant data and residual meaningful in-formation,performance evaluation indicators are classified using a fused knowledge graph to achieve optimal mining.Experi-mental results show that when the number of iterations is 5,the mining accuracy of the proposed method is 0.93.The support rate of the indicator system after mining has increased by about 10％,which can effectively eliminate the redundant evaluation indicators of hospital performance.关键词
知识图谱/协同过滤/医院绩效/考核指标体系/数据挖掘Key words
knowledge graph/collaborative filtering/hospital performance/evaluation indicator system/data mining分类
信息技术与安全科学
