In view of the poor fitting of the traditional GM (1,1) model in the long-term prediction and its failure in data series with high volatility, a grey Markov prediction model is established by combining grey theory with Markov theory. The location of moving objects is predicted by this model, and more accurate prediction results are obtained. Firstly, the GM (1,1) model is constructed by abstracting the problem into a single variable grey system. Then, at one-minute interval, some GPS trajectory data in GeoLife are extracted and substituted into the prediction model, and the model is tested according to the results. Finally, the Markov model is introduced into GM (1,1) model for further experiments. Experiments show that compared with the traditional GM (1,1) model, the grey Markov prediction model has less error in the location prediction of moving objects and improves the accuracy of the prediction.
[1]
Mikolaj Morzy,et al.
Mining Frequent Trajectories of Moving Objects for Location Prediction
,
2007,
MLDM.
[2]
Hong Ji,et al.
Optimal Access Points Selection Based on Mobility Prediction in Heterogeneous Small Cell Networks
,
2018,
2018 IEEE/CIC International Conference on Communications in China (ICCC).
[3]
Zhen Yang,et al.
Next Location Prediction Based on an Adaboost-Markov Model of Mobile Users †
,
2019,
Sensors.
[4]
Jie Zhao,et al.
A review of moving object trajectory clustering algorithms
,
2016,
Artificial Intelligence Review.
[5]
Jian Yao,et al.
LP-HMM: Location Preference-Based Hidden Markov Model
,
2018,
ICSINC 2018 Fall.