需求驱动的医疗健康大数据挖掘模型构建

将跨行业数据挖掘标准流程模型(简称“CRISP-DM模型”)和需求、数据、数据挖掘方法三者的对应关系相结合构建需求驱动的数据挖掘模型。该模型主要包括需求、数据和数据挖掘方法3个核心概念与需求理解、数据理解、数据准备、建模、评估和部署6个过程步骤,分为需求层、数据层、算法层和部署层4个层次。需求驱动的医疗健康大数据挖掘模型可打破“数据先入为主”的习惯性思维,降低盲目进行数据挖掘研究引发的失败风险。整理需求、数据和数据挖掘方法之间的对应关系,可优化医疗健康大数据挖掘路径,降低数据挖掘新手的学习成本,对临床医疗和健康管理实践都具有现实意义,可用于未来的医疗健康大数据开发策略研究。

Demand-driven establishment of medical and health big data mining model

The demand-driven data mining model was established by combining the cross-industry standard process for data mining with the corresponding relationships of demand, data and data mining methods, which includes need concept, data concept, data mining methods concept, and demand understanding step, data understanding step, data preparation step, modeling step, evaluation step, deployment step, and are divided into demand layer, data layer, algorithm layer, deployment layer. The demand-driven medical and health big data mining model can break the habitual thinking of "taking the first impression as the strongest", and reduce the risk of failure induced by studies on blind data mining. It is of practical significance for medical and health management to rearrange the corresponding relation-ships of demand, data and data mining, to optimize the medical and health big data mining methods, and reduce the learning cost of new data miners, and can thus be used in studies on the strategies for future medical and health big data development.