A Study of Pattern Prediction in the Monitoring Data of Earthen Ruins with the Internet of Things.
A Study of Pattern Prediction in the Monitoring Data of Earthen Ruins with the Internet of Things.
复制标题
物联网土遗址监测数据模式预测研究
DOI：
10.3390/s17051076
复制
发表时间：
2017-05-11
期刊：
影响因子：
--
通讯作者：
Fang D
中科院分区：
文献类型：
--
作者：
Xiao Y;Wang X;Eshragh F;Wang X;Chen X;Fang D
了解土遗址夯土温度的变化，对土遗址的保护具有重要意义。为了利用土遗址监测数据的气温模式预测夯土温度模式，提出了一种基于兴趣模式挖掘和关联的模式预测方法——PPER。PPER首先在空气温度序列和夯土温度序列中发现了有趣的模式。为了减少处理时间，提出了两种剪枝规则和一种新的基于r树的数据结构。然后挖掘了空气温度模式与夯土温度模式之间的相关规律。将相关规则合并为夯土温度型的预测规则。实验结果表明了所提方法的准确性和剪枝规则的有效性。利用明代长城数据对算法进行验证，得到了从气温到夯土温度基于兴趣模式的6条预测规则，平均准确率达到89.8%。该模型和预测规则将为土遗址保护中夯土温度的预测提供依据。
An understanding of the changes of the rammed earth temperature of earthen ruins is important for protection of such ruins. To predict the rammed earth temperature pattern using the air temperature pattern of the monitoring data of earthen ruins, a pattern prediction method based on interesting pattern mining and correlation, called PPER, is proposed in this paper. PPER first finds the interesting patterns in the air temperature sequence and the rammed earth temperature sequence. To reduce the processing time, two pruning rules and a new data structure based on an R-tree are also proposed. Correlation rules between the air temperature patterns and the rammed earth temperature patterns are then mined. The correlation rules are merged into predictive rules for the rammed earth temperature pattern. Experiments were conducted to show the accuracy of the presented method and the power of the pruning rules. Moreover, the Ming Dynasty Great Wall dataset was used to examine the algorithm, and six predictive rules from the air temperature to rammed earth temperature based on the interesting patterns were obtained, with the average hit rate reaching 89.8%. The PPER and predictive rules will be useful for rammed earth temperature prediction in protection of earthen ruins.