Location Prediction on Trajectory Data:   A Review

Abstract

Location prediction is the key technique in many location based services including route navigation, dining  location recommendations, and traffic planning and control, to mention a few. This survey provides a comprehensive  overview of location prediction, including basic definitions and concepts, algorithms, and applications. First, we  introduce the types of trajectory data and related basic concepts. Then, we review existing location-prediction  methods, ranging from temporal-pattern-based prediction to spatiotemporal-pattern-based prediction. We also  discuss and analyze the advantages and disadvantages of these algorithms and briefly summarize current  applications of location prediction in diverse fields. Finally, we identify the potential challenges and future research  directions in location prediction.

Existing System 

Urban planning, relieving traffic congestion, and  effective location recommendation systems are  important objectives worldwide and have received  increasing attention in recent years. Spatiotemporal  data mining is the key technique involved in these  practical applications Trajectory data brings  new opportunities and challenges in the mining of  knowledge about moving objects. To present, many  researchers have used trajectory data to mine latent  patterns that are hidden in data. These patterns can  also be extracted for the analysis of the behavior  of moving objects. Location prediction, as the  primary task of spatiotemporal data mining, predicts  the next location of an object at a given time. In  recent years, researchers in location prediction have made much progress. For instance, early studies  traced student ID cards to identify frequent temporal  patterns and used these patterns to predict their next  location Since then, location prediction has had  a wide range of applications in daily life, e.g., travel  recommendation, location-aware advertisements, and  early warning of potential public emergencies, to  mention a few.

Proposed System 

Introduce  these two types of trajectory data below.  _ Active recording trajectory data: People actively  record their locations when they login to social  networks or travel to places of interest and share  their life experiences. Typical data types include  check-in data (e.g., Twitter, Weibo, QQ, etc.),  and location-based data such as travel photos.  Figure 1 shows the correlation of users and  locations in social networks. In Flickr, a sufficient  number of geotagged photos can be formulated as a trajectory, whereby each photo is associated  with a location tag and a time stamp. Due to  the random behaviors of users, active recording  data is typically characterized by its sparsity.  When mining this trajectory data, additional social  information is usually added.  _ Passive recording trajectory data: With  the development of positioning techniques,  many moving objects are equipped with  positioning position devices that record  location information. These include global  positioning systems GPS in vehicles and radiofrequency  identification devices for tracing  objects.

Conclusion 

In this article, we provided an overview of location  prediction ranging from trajectory data preprocessing  to forecasting location and the evaluation of location  prediction systems. First, we introduced the basic  concepts of location prediction, the different types of  data sources, the challenges associated with location  predictions and the location prediction framework. We  introduced trajectory data preprocessing methods and  then identified the classification of location prediction  model types and discussed these models in detail. Next,  we categorized location-prediction models as either  single-object or group models or shared insights about  these approaches. We also listed the available public  datasets and evaluation methods to help readers conduct  their own research. Lastly, we discussed locationprediction  applications and future work.

References 

[1] Y. L. Wang, Y. Zheng, and Y. X. Xue, Travel time  estimation of a path using sparse trajectories, in Proc. 20th  ACM SIGKDD Int. Conf. Knowledge Discovery and Data  Mining, New York, NY, USA, 2014, pp. 25–34.

[2] H. K. Pao, J. Fadlil, H. Y. Lin, and K. T. Chen, Trajectory  analysis for user verification and recognition, Knowl.-  Based Syst., vol. 34, pp. 81–90, 2012.

[3] J. Yuan, Y. Zheng, X. Xie, and G. Z. Sun, Driving  with knowledge from the physical world, in Proc. 17th  ACM SIGKDD Int. Conf. Knowledge Discovery and Data  Mining, San Diego, CA, USA, 2011, pp. 316–324.

[4] A. L. Barab´asi, The origin of bursts and heavy  tails in human dynamics, Nature, vol. 435, no. 7039, pp.  207–211, 2005.

[5] D. Brockmann, L. Hufnagel, and T. Geisel, The scaling  laws of human travel, Nature, vol. 439, no. 7075, pp. 462–  465, 2006.

[6] A. Alahi, K. Goel, V. Ramanathan, A. Robicquet, F.  F. Li, and S. Savarese, Social LSTM: Human trajectory  prediction in crowded spaces, in IEEE Conf. Computer  Vision and Pattern Recognition, Las Vegas, NV, USA,  2016, pp. 961–971.

[7] F. Wu and Z. H. Li, Where did you go: Personalized  annotation of mobility records, in Proc. 25th ACM  Int. Conf. Information and Knowledge Management,  Indianapolis, IN, USA, 2016, pp. 589–598.

[8] C. L. Wang, D. De, and W. Z. Song, Trajectory  mining from anonymous binary motion sensors in Smart  Environment, Knowl.-Based Syst., vol. 37, pp. 346–356,  2013.

[9] W. Y. Zhu, W. C. Peng, L. J. Chen, K. Zheng, and X.  F. Zhou, Modeling user mobility for location promotion  in location-based social networks, in Proc. 21th ACM  SIGKDD Int. Conf. Knowledge Discovery and Data  Mining, Sydney, Australia, 2015, pp. 1573–1582.

[10] M. Q. Lv, L. Chen, Y. B. Shen, and G. C. Chen,  Measuring cell-id trajectory similarity for mobile phone  route classification, Knowl.-Based Syst., vol. 89, pp. 181–  191, 2015.

[11] Y. Zheng, Trajectory data mining: An overview, ACM  Trans. Intell. Syst. Technol., vol. 6, no. 3, p. 29, 2015.

[12] J. J. Liu, K. Zhao, P. Sommer, S. Shang, B. Kusy, J.  G. Lee, and R. Jurdak, A novel framework for online  amnesic trajectory compression in resource-constrained  environments, IEEE Trans. Knowl. Data Eng., vol. 28, no.  11, pp. 2827–2841, 2016.

[13] K. F. Richter, F. Schmid, and P. Laube, Semantic trajectory  compression: Representing urban movement in a nutshell,  J . Spat. Inf. Sci., vol. 2012, no. 4, pp. 3–30, 2012.

[14] Y. T. Jia, Y. Z. Wang, X. L. Jin, and X. Q. Cheng,  Location prediction: A temporal-spatial Bayesian model,  ACM Trans. Intell. Syst. Technol., vol. 7, no. 3, p. 31, 2016.

[15] W. C. Lee and J. Krumm, Trajectory Preprocessing. New  York, NY, USA: Springer, 2011.