概率粗糙集三支决策在线快速计算算法研究

随着大数据和物联网技术的不断发展，动态在线计算已经成为了一种常见的计算模式，在动态在线计算中进行不确定问题的推理和求解是一项具有挑战性的新议题。概率粗糙集三支决策理论是一种处理不确定性知识挖掘的有效工具，根据在线计算模式中数据同步增减的动态特点，提出了一种概率粗糙集三支决策的在线计算方法。首先，以内存滑动窗口模式对在线动态计算的数据变化特点进行理论建模；然后，根据上述模型中在线计算的数据变化模式，推导出不同类型数据变化模式下的三支决策条件概率及三支区域的变化规律；最后，提出了一种新型在线快速计算算法，其获取的三支决策规则与经典概率三支决策算法是等效的。通过与经典三支决策计算算法的多组对比实验，验证了提出的在线快速计算算法的高效性与稳定性。
With the continuous development of big data and IoT (Internet of Things), dynamic online computation has become a common computing pattern; however the field of dynamic online computation faces challenges in deducing and solving uncertainty problems. A three-way decision theory with probabilistic rough set method is an efficient tool for mining uncertain knowledge; thus a dynamic online computing approach of three-way decision theory with probabilistic rough set is proposed in this paper, in accordance with the features of data dynamic synchronization. First, a data model is established to describe the inherent features of dynamic online computation via memory sliding window mode. In terms of the variational features of dynamic online computation of the above model, a three-way decision conditional probability and the change rule of three-way area are deduced as diverse variational patterns of data. Finally, a novel algorithm of online rapid computation is proposed. The obtained three-way decision rule is identical with the three-way decision algorithm of classic probability. By comparison with the classic three-way decision algorithm through multiple experiments, the proposed online rapid computation algorithm is confirmed to have high efficiency and stability