基于Hadoop的医疗云存储与医生推荐系统的研究与实现

Research and Implementation of Hadoop-based Medical Cloud Storage and Doctor Recommendation System

随着医疗行业的快速发展,医疗领域数据量呈现几何级增长,但是这些数据的存储和二次利用往往得不到解决,造成资源的浪费。本文设计实现了一套基于Hadoop的医疗大数据存储与医生推荐系统,利用Hadoop集群的特性为医疗信息提供高容错性和高可靠性的存储;同时利用Map-Reduce编程框架实现了基于项目的协同过滤算法为用户推荐医生,Map-Reduce具有良好的并行性,是天然的大数据分析框架,实现了医疗大数据的二次利用。通过实验测试表明,本系统可以很好的存储结构化、半结构化和非结构化医疗大数据,并且推荐结果具有实时性和较好的准确性。

With the rapid development of the medical industry, medical data volume showed a geometric growth, but the storage and secondary use of these data is often not resolved, resulting in waste of resources. This paper designs a set of medical data storage and medical recommendation system based on Hadoop, which utilizes the characteristics of Hadoop cluster to provide high fault tolerance and high reliability storage for medica/information. At the same time, we use the Map-Reduce programming framework to implement item-based Collaborative Filtering algorithm recommending doctors that users need. Map-Reduce has a good parallelism, and is a natural large data analysis framework, to achieve the secondary use of medical data. Experiments demonstrate that the system can store the structured, semi-structured and unstructured medical large data well, and the results are real-time and accurate.