基于粒计算的时间粒建模及其基本运算
ime Granular Modeling on Temporal Big Data and its Basic Computing
时序数据的形式化研究是解决复杂时间序列大数据问题的理论基础。基于等价关系、全序关系等二元关系,本文研究了时间序列大数据分析中普遍存在的时间粒度化现象,提出了时间粒系统(TGS)。在证明TGS中时间粒集合是完全格后,构造了一种新型时间粒哈斯图。通过时间粒哈斯图方法研究了TGS中的基本元素即时间粒的语义、时间粒间的二元关系及多个时间粒间关系的转换。最后定义了时间粒的一元计算和二元计及其逻辑等价公式。这些计算定义和逻辑基础可以实现时间粒的不同粒度互转化、不同时间粒之间的二元关系计算及时间粒间复杂关系的逻辑推理。可以证明TGS中的时间粒元素集合及其基础运算是一套完备的格代数系统。
he formalism of temporal data is the theoretical fundament for the settlement of time-related big data problems with complicated property. This paper studied the generality towards the analysis of temporal big data from the perspective of binary relation including equivalence relation and linear ordering. A novel information system Temporal Granular System (TGS) is thus proposed . The Hasse diagram is presented after demonstrating that TGS is a complete lattice. This diagram can provide a more intuitive perception to elaborate the syntax of time granular, the basic component of TGS, as well as the intrinsic relation and converted practicability among different time granular. Furthermore, the unary operators, as well as binary operators, are defined and equivalent forms of whose are further deduced. The aforementioned logic and computing operators enable the transformation of time granular in different hierarchies and the relation reasoning of different time granular. Finally, it is proved that TGS, together with its calculus, is a complete lattice algebraic system.
徐健锋、苗夺谦、张远健
计算技术、计算机技术数学
时间序列二元关系完全格时间粒代数系统
time seriesbinary relationcomplete latticetemporal granularalgebraic system
徐健锋,苗夺谦,张远健.基于粒计算的时间粒建模及其基本运算[EB/OL].(2017-04-19)[2025-08-10].http://www.paper.edu.cn/releasepaper/content/201704-216.点此复制
评论