摘要
在医疗系统中,人们通常使用决策树对患者的发病类型以及概率进行有效地分类预测。随着信息技术的普及,医疗系统中产生了大量的历史医疗记录,处理和分析这些海量的医疗数据给医疗系统带来了极大地挑战。本文针对海量医疗数据问题,提出了分布式构建决策树算法。该算法分布式逐层构建决策树,可以高效地构建决策树,快速有效地完成医疗系统中的预测工作。该算法是基于现有流行的云计算平台,使用MapReduce分布式框架设计的分布式算法。实验结果表明,该算法具有很好的扩展性和高效性。
In the medical system,people usually use decision tree to predict the type of disease or the possibility of getting one disease for patients.With the development of information technology,it produces a huge amount of medical records in the system.And,it takes great challenge to the medical system to deal with the huge scale data.In this paper,focusing on mass medical data,we propose a distributed decision tree algorithm.This algorithm builds the decision tree layer by layer.It could build decision tree effectively and complete the predication in medical system.The algorithm is based on popular cloud computing platform with MapReduce distributed programming framework.The experiment results show that the algorithm has good scalability and efficiency.
出处
《科技通报》
北大核心
2013年第2期45-47,共3页
Bulletin of Science and Technology