Traffic Behavior Recognition And Modeling Techniques With Personal Social Information

It is difficult for traditional traffic information service to adapt to its surroundings,and there are strong human needs to be able to personally utilize the services.Because of the rapid development of the Internet industry chain,spatial orientation technology matures,it is possible to provide travelers with the real-time,personalized location-based services.The behavior recognition for travelers by fusing their location information in social networks was conducted in this paper.Firstly,the framework used to identify the recognition under the traffic environment was provided,following analyses of the services' features.Then,three kinds of stay using the historical trajectories of traveler were put forwards,and the sequence-based behavior model with the time stamp under specific constraints was built.Using the check-in data,the behavior model was extent for gathering and updating in time.Meanwhile,considering the complexity of different location data,we introduced the multi-resource data fusion algorithm to deal with the problem of the sparse data in social networks and improve the accuracy of behavior recognition.The experiment results shown that the recognition proposed in this paper was able to identify the travel behaviors,and the rate of recognition was more than 90%,while the recall can reach 84.9%.It can be found that the algorithm presented can to some extent improve the recognition with a large number of social data.