Abstract
As an emerging class of spatial trajectory data, mobile user trajectory data can be used to analyze individual or group behavioral characteristics, hobbies and interests. Besides, the information extracted from original trajectory data is widely used in smart cities, transportation planning, and anti-terrorism maintenance. In order to identify the important locations of the target user from his trajectory data, a novel division method for preprocessing trajectory data is proposed, the feature points of original trajectory are extracted according to the change of trajectory structural, and then important locations are extracted by clustering the feature points, using an improved density peak clustering algorithm. Finally, in order to predict next location of mobile users, a multi-order fusion Markov model based on the Adaboost algorithm is proposed, the model order k is adaptively determined, and the weight coefficients of the 1~k-order models are given by the Adaboost algorithm according to the importance of various order models, a multi-order fusion Markov model is generated to predict next important location of the user. The experimental results on the real user trajectory dataset Geo-life show that the prediction performance of Adaboost-Markov model is better than the multi-order fusion Markov model with equal coefficient, and the universality and prediction performance of Adaboost-Markov model is better than the first to third order Markov models.
Highlights
In recent years, with the rapid development of mobile communication technology and the increasingly powerful functions of smart mobile terminal networks, smart terminal devices such as mobile phones and tablet computers have gradually surpassed personal computers and become the most widely used information devices for people
In order to make the service more forward-looking, in recent years, a large number of domestic and foreign scholars have turned their attention to the location prediction of mobile users [2,3,4,5]
Location prediction concentrates on predicting the location of Sensors 2019, 19, 1475; doi:10.3390/s19061475
Summary
With the rapid development of mobile communication technology and the increasingly powerful functions of smart mobile terminal networks, smart terminal devices such as mobile phones and tablet computers have gradually surpassed personal computers and become the most widely used information devices for people. DJ-Cluster to extract important locations in the original trajectory data, in view of the shortcomings of the standard Markov model, they proposed an extended mobility Markov chain called n-MMC to improve the prediction accuracy [32]. Gidófalvi et al used a grid-based approach to preprocess the original trajectory data, periodically extracted and managed important locations that the user frequently visited, they proposed an inhomogeneous continuous-time Markov model to predict when the user will leave his current region and where he will move [33]. The conclusions of our work are presented in the “Conclusions” section
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
