Starting from medical big data, this article uses data mining technology to analyze and study the pathogenic factors of lung cancer based on the lung cancer electronic medical record data from the oncology department of the authoritative third grade A hospital for many years. With respect to the processing of huge data from electronic medical records for lung cancer, traditional serial Apriori algorithm has the disadvantages of scanning database frequently, running slowly and consuming large amount of memory resources. Therefore, an improved Apriori algorithm based on MapReduce distributed computing model of Hadoop platform is proposed. The experimental cluster and lung cancer data mining experiments show that the improved Apriori algorithm has higher execution efficiency and good system scalability in dealing with lung cancer big data, and can well mine the relationship between lung cancer and pathogenic factors, which has important guiding significance for assisting the clinical diagnosis and risk prediction of lung cancer.

本文从医学大数据入手，利用数据挖掘技术，以多年权威三级甲等医院肿瘤科的肺癌电子病历数据为基础，对肺癌的致病因素进行分析研究。对于处理来自肺癌的电子病历的海量数据，传统的串行Apriori算法具有扫描数据库频繁，运行缓慢和消耗大量内存资源的缺点。因此，提出了一种基于Hadoop平台MapReduce分布式计算模型的改进Apriori算法。实验集群和肺癌数据挖掘实验表明，改进的Apriori算法在处理肺癌大数据方面具有更高的执行效率和良好的系统可扩展性，