基于多维深层数据关联的医学知识挖掘研究进展

数据科学和情报学方法的核心在于如何从数据中挖掘出知识和见解。在与生命健康密切相关的医学和医疗领域,大数据分析应在相关性挖掘基础上揭示因果关系,增强重复性和解释性。基于因果关系的数据关联对于智库研究和情报感知具有重要意义。文章提出基于多维数据关联和深层数据关联的医学知识挖掘思路,介绍了相关数据平台和研究进展。一是实验室—临床知识转化测度与临界分析;二是科学的技术影响力测度;三是交叉性、变革性创新前沿识别;四是基于全文本、融合文献计量学与计算语言学的不确定性医学知识挖掘。前三个方面拓展了医学知识的空间,包括从实验室到临床,从科学空间到技术空间。对于确定性/不确定性医学证据和论断挖掘深化了对医学知识的因果关系的揭示和解释。

Biomedical Knowledge Discovery Based on Big Data Linkage Analysis

Extracting knowledge and insights from the data is the core of data science and informatics approach. In the medical field, big data analysis is applied to reveal causal relationships and enhance its repeatability and inter-pretability based on correlation mining. Analysis of data association with causality is of great significance for think tank research and intelligence perception. To reveal the causal relationship between knowledge, the paper introduces relevant data platforms and research progress, and proposes a medical knowledge mining ideas based on multi-space and deep data. One is the measurement and critical analysis of laboratory-clinical knowledge trans-formation;the other is scientific technological influence measurement;the third is the identification of cross-cutting and innovative frontiers;the last one is the mining of medical knowledge based on the combination of full-text, bib-liometrics and computational linguistics. The first three approaches expand the space of medical knowledge, includ-ing from basic research space to applied research space, and from scientific space to technological space. The fourth way deepens the disclosure and explanation of the causality of medical knowledge based on certainty or uncertainty of the medical knowledge.