AP-I:一种快速预测路网中移动对象未来位置的索引

发布时间：2018-06-26 12:17

本文选题：Predictive + Query　；参考：《计算机科学》2017年S1期

【摘要】：随着智能交通、基于位置的广告投放、移动对象监测等应用的广泛发展,如何快速预测未来某一时间点的对象的位置成为目前的一个研究热点。提出了一种新颖的AP-I(Adaptive Predication-Index)索引,其在历史轨迹数据缺乏的情况下,能够追踪移动对象的当前位置,大幅提高预测查询的运行效率。与现有的Predictive Tree~([4])索引相比,AP-Index能有效地挖掘移动对象之间的路径关联性,避免大量的索引更新和重建操作,提高索引效率。同时,通过引入AP(Adaptive Probability)以及Pruning操作,进一步减小AP-I,提高索引的命中率和查询效率。实验表明,与Predictive Tree相比,在保证同等查询效率的基础上,AP-I实现了更优的准确度、更新效率和空间效率。
[Abstract]:With the wide development of intelligent transportation, location-based advertising, mobile object monitoring and other applications, how to quickly predict the location of objects at a certain point in the future has become a research hotspot at present. In this paper, a novel AP-I (Adaptive Predication-Index) index is proposed, which can track the current position of moving objects in the absence of historical track data and greatly improve the efficiency of prediction query. Compared with the existing predictive tree ([4]) index, AP-Index can effectively mine the path correlation between moving objects, avoid a large number of index updating and reconstruction operations, and improve the index efficiency. At the same time, by introducing AP (Adaptive probability) and pruning operations, the AP-I is further reduced, and the hit rate and query efficiency of the index are improved. Experimental results show that AP-I achieves better accuracy update efficiency and spatial efficiency than predictive Tree on the basis of ensuring the same query efficiency.
【作者单位】：深圳职业技术学院计算机工程学院;
【分类号】：O157.5

【相似文献】