跨组织医疗健康大数据聚合与案例知识推理方法研究

Cross-Organizational Medical and Health Big Data Fusion and Knowledge Recommendation Method

【目的/意义】医疗健康大数据为智慧医疗提供了前所未有的机遇。然而"数据烟囱""信息孤岛"和低效的知识服务方法严重阻碍医疗健康服务模式创新。如何通过医疗健康大数据深度聚合和动态知识服务,实现面向全方位全周期智慧医疗服务的知识管理创新成为当前医疗信息资源管理领域的重要问题。【方法/过程】介绍了一种面向大规模多源异构医疗健康数据安全共享的联邦学习机制和深度聚合方法,提出了人机协同的医疗案例库构建方法和基于杰卡德距离算法的医疗案例知识推理方法。【结果/结论】该方法为智慧诊疗、临床教学和辅助科研提供了一体化知识管理服务框架。【创新/局限】该方法不仅为智慧医疗与精准健康管理提供了一种数据管理方法体系,还为5P智慧医疗服务新模式构建提供了新的思路。

【Purpose/significance】Healthcare big data provides unprecedented opportunities for smart healthcare. However, "data chimneys", "information silos" and inefficient knowledge service methods seriously hinder the innovation of medical and health service models. How to realize knowledge management innovation for all-round and full-cycle smart medical services through deep integration of medical and health big data and dynamic knowledge services has become an important issue in the field of medical information resource management.【Method/process】We introduce a federated learning mechanism and deep fusion method for the secure sharing of large-scale multi-source heterogeneous medical and health data, and propose a human-machine collaborative medical case database construction method and a medical case knowledge recommendation method based on Jaccard distance algorithm.【Result/conclusion】This method provides an integrated knowledge management service framework for smart diagnosis and treatment, clinical teaching and auxiliary scientific research.【Innovation/limitation】This method not only provides a data management method system for smart medical and precise health management, but also provides a new idea for the construction of a new model of 5 P smart medical service.