一种对时间连续数据进行分析查询的 SQL 扩展语言

摘要

There is the physiological time series data in home e-health platform.To realise its analysis and query is the basis of data utilisation.In light of the characteristics of such data, we propose that taking the relational database structure query language ( SQL) as the basis to define the similarity query algorithm and to find out fast and effectively the similarity subsequence in accordance with a certain similarity level from given signal sequence.The similarity matching uses two algorithms, the segmented aggregation approximation and the wavelet transform.They are easy to implement the similarity grading of subsequence.Experimental results show that this test system can efficiently and accurately complete the designed function.%在家庭电子医疗平台中,存在生理学时间连续数据,实现对其分析和查询是数据利用的基础. 针对这类数据的特点,提出以关系型数据库结构化查询语言SQL为基础,定义相似查询算法,快速有效地从给定信号序列中找出符合某一相似度等级的相似子序列. 相似匹配使用了分段聚合近似法和小波变换两种算法,易于进行子序列的相似度分级. 实验结果表明,该试验系统能够高效、准确地完成设计功能.