基于spark技术的心脏病预测平台研究与设计

Research and design of a heart disease prediction platform based on spark technology

针对目前心脏病预测存在的准确率低、技术平台分散、耦合差等问题,本文采用spark大数据处理技术,结合HDFS分布式数据存储技术,设计一种心脏病预测平台。该平台选用spark机器学习库中的决策树算法实现心脏病预测建模,利用SpringBoot技术搭建Web服务器,结合Mysql数据库实现预测模型与Web服务器的数据实时交互。本文以美国某区域ICU开源的体测数据为数据源,对平台进行验证,通过测试,该平台可以根据用户提供的体态数据实时预测是否患有心脏病,准确率达到89.2%。系统具有稳定可靠、操作简单、实时性强等特点。

Aiming at the problems of low accuracy,decentralized technology platform,and poor coupling in current heart disease prediction,this paper uses spark big data processing technology,combined with HDFS distributed data storage technology,to design a heart disease prediction platform.The platform uses the decision tree algorithm in the spark machine learning library to implement heart disease prediction modeling,uses SpringBoot technology to build a web server,and combines with Mysql database to achieve real-time data interaction between the prediction model and the web server.This article uses the open source physical test data of a certain area of the United States as a data source to verify the platform.Through testing,the platform can predict whether there is a heart disease in real time based on the posture data provided by the user,with an accuracy rate of 89.2%.The system has the characteristics of stability,reliability,simple operation,and strong real-time performance.