基于分层聚类的心血管疾病协同过滤推荐模型研究

Research on the Collaborative Filtering Recommendation Model of Cardiovascular Disease Based on Hierarchical Clustering

目的:医院收集了大量医疗数据,这些数据中提取的信息可用于临床决策和治疗推荐。传统的协同过滤推荐算法面临着数据稀疏性和可扩展性差等问题,这会降低系统的准确性和效率。方法:提出了一种基于聚类和子聚类的实时协同过滤推荐模型。该方法被用于四种不同类型的心血管疾病数据集,包括心绞痛、非心源性胸痛、隐形缺血和心肌梗死。将患者数据按照对应的疾病类别进行分层聚类,查询患者被引导到正确的疾病分区并从较小的子集中获得相似度评分。结果:实验结果表明,模型减少了查询患者的搜索范围,从而压缩提供准确建议所需时间,提升了推荐的准确性和效率。结论:模型对心血管疾病相关治疗方案的推荐具有重要的意义。

Objective:Hospitals collected large amounts of medical data,and the information extracted from this data could be used for clinical decision-making and treatment recommendations.The traditional collaborative filtering recommendation algorithm was faced with the problems of data sparsity and poor scalability,which will reduce the accuracy and efficiency of the system.Methods:A real-time collaborative filtering recommendation model based on clustering and sub-clustering was proposed.The method was used in four different types of cardiovascular disease data sets,including angina,non-cardiogenic chest pain,silent ischemia and myocardial infarction.The patient data were clustered and sub-clustered according to the corresponding disease category and the patients were asked to be directed to the correct disease partition and the similarity score was obtained from the smaller subset.Results:The experimental results show that the recommendation model reduces the searching range of patients to reduce the time needed to provide accurate suggestions,and improves the accuracy and efficiency of recommendations.Conclusion:The model is of great significance for the recommendation of medical programs related to cardiovascular diseases.