基于位置相似性与Markov模型的移动轨迹预测算法
Moving Trajectory Prediction Algorithm Based on Position Similarity and Markov Model
李佳泽 1高全力 1郭帅 1胡发丽 1李庆敏1
作者信息
- 1. 西安工程大学计算机科学学院 西安 710048
- 折叠
摘要
实时、准确的移动轨迹预测是移动端新兴业务发展的重要支撑.目前的移动轨迹预测大多采用Markov模型,但是Markov模型存在对历史轨迹信息利用不充分的问题.为此,论文提出了一种基于位置相似性与Markov模型的移动轨迹预测算法.首先,通过迭代网格划分方法实现原始轨迹数据的序列化;然后,根据用户当前轨迹位置找出历史轨迹集合中与之相似的历史轨迹;最后,由相似的历史轨迹建立转移概率矩阵,完成对用户未来区域的预测.在大规模数据集上的试验结果表明,相比于传统Markov模型,该方法的平均预测准确率提高了13.8%.
Abstract
Real time and accurate mobile trajectory prediction is an important support for the development of emerging mobile services.At present,Markov model is mostly used to predict the moving trajectory,but the Markov model does not make full use of the historical trajectory information.Therefore,this paper proposes a moving trajectory prediction algorithm based on position simi-larity and Markov model.Firstly,the original trajectory data is serialized by iterative meshing method.Then,the similar historical trajectories in the historical trajectory set are found according to the user's current trajectory position.Finally,the transition probabil-ity matrix is established from the similar historical trajectories to complete the prediction of the user's future region.Experimental re-sults on large-scale datasets show that the average prediction accuracy of this method is improved by 13.8%compared with the tradi-tional Markov model.
关键词
推荐服务/轨迹预测/位置相似性/Markov模型Key words
recommendation service/trajectory prediction/location similarity/Markov model引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据