Research On The Construction Of User Behavior Characteristic Model Based On Semantic Stay Point

With the development and popularization of smart mobile devices and the continuous improvement and development of positioning technology,the GPS trajectory of people in daily life can be easily obtained and recorded.GPS trajectory data contains a large number of meaningful data,can truly reflect the user in the real world activities trajectory,and the ability to show the user the information such as the travel mode and social attribute,so the study of user GPS trajectory is a popular research direction.Most of the research by analyzing the user's GPS trajectory stay point to get to the user's information,but didn't add semantic information,therefore,this article takes the user trajectory starting point,put forward a kind of stay point extraction algorithm based on time and semantics,the semantic information fusion to stay point to get to the user of the semantics of the spots,and then stay in semantic point on the basis of further analysis of user behavior,get the user's behavior characteristic,finally set up the user behavior model based on semantic stops.Specifically,the work is as follows:Firstly,in the case that other clustering algorithms do not consider the factors of time and semantics in the process of obtaining stop points,an HST-OPTICS algorithm based on semantics and time is proposed.The algorithm improves the accuracy of the stay points by increasing the time factor,and then obtains the semantic information such as residence time,date type and map annotation by the data point set of the stop point.Secondly,the user's semantic stop point is obtained by using the hst-optics algorithm,and the user behavior characteristic model based on the semantic stop point is constructed.Stay in user semantic point on the basis of the analysis of user behavior,get the user's behavior characteristics,and then using the analytic hierarchy process(AHP)and TF-IDF method to calculate the weights of each user behavior feature,according to the behavior feature and weight of the each item behavior characteristics to build a user behavior model.Thirdly,a new algorithm for updating user model is presented.After the user model is not invariable,influenced by factors such as time and interest,user behavior will stay with the semantic characteristics of the change and development,aiming at the effects of these factors,the forgetting curve and LRU algorithm is proposed on the basis of a user behavior model updating algorithm.Through the verification experiment,this paper proves that the time and semantics of the stop-point extraction algorithm is able to obtain the accurate information of the user's stay point and stop point.The user behavior characteristic model based on semantic stop points constructed by analysis can show the characteristics of users' daily behaviors.