基于形式概念分析的新冠肺炎疫情大数据挖掘

COVID-19 Big Data Mining Based on Formal Concept Analysis

形式概念分析是一种分析数据和提取规则的有力工具,其核心结构-概念格体现了对象与属性间的统一,通过构造概念格可以挖掘数据中隐含的概念,以及概念之间的层次关系。因此,在大数据的背景下,运用形式概念分析理论,从大量数据中进行规则提取和发现是切实可行并且大有可为的。该文将之引入到新冠肺炎患者医疗数据的分析中来,通过构建形式背景和概念格来探究新冠肺炎病理学潜在模式,并通过统计学方法分析患者性别、年龄、临床症状分布规律;然后,用Pearson相关系数探究新冠肺炎与不同影响因素间的隐含联系,并进行了相关实验分析。得出结论:新冠肺炎各年龄段人群普遍易感,其中,中老年人和慢性病患者感染的可能性更大。新冠肺炎临床症状主要表现为发热和咳嗽,部分个体会伴随有肌肉疼痛、咽喉疼、疲劳等症状。最后,以上述分析结果为基础,提出相应的防控策略。

Formal concept analysis(FCA) is a powerful tool for data analysis and rule extraction.Its core structure,concept lattice,reflects the unity between objects and attributes.By constructing concept lattice,implicit concepts in data and hierarchical relationships between concepts can be mined.Therefore,in the context of big data,it is feasible and promising to use FCA to extract and discover rules from a large amount of data.We introduce FCA into the analysis of medical data of COVID-19 patients to explore potential pathological patterns of COVID-19 by constructing formal background and concept lattice,and to analyze the distribution rules of gender,age and clinical symptoms of patients by statistical methods.Then,we explore the implicit connection between the COVID-19 and different influencing factors with Pearson correlation coefficient,and carry out relevant experimental analysis.It is concluded that COVID-19 is widespread among people of all ages,with middle-aged and elderly people and those with chronic diseases more likely to be infected.The clinical symptoms of COVID-19 are mainly fever and cough,and some people will experience symptoms such as muscle pain,sore throat,and fatigue.Finally,the corresponding prevention and control strategies are suggested based on the analysis results.