Research And Implementation Of POI Recognition And Recommendation Algorithm Based On Mobile Data Analysis

With the popularity of mobile devices,mobile phones built-in expansion of the growing functional components,the values of data in the device are attracting people's close attention.The obtained data,from portable devices such as mobile phone,are more real-time and accurate.However,trajectory behavior mining is not properly associated with context for users.Moreover,combined with the location information recommended services are not well applied.In this paper,we first perform the behavior mining on the user's trajectory,use the time window to identify the user's intensive time interval,and then carry on the user's POI identification for each time-intensive interval.During the identification of POI,the second-order velocity collaborative clustering POI recognition algorithm is proposed.In this algorithm,an adaptive strategy is utilized to compute the weights,which are weights of preselected in previous clustering algorithms when different dimensions are mixed.While in the process of finding POI to consider the speed factor,to avoid low-value POI.Secondly,we propose a personalized recommender algorithm based on timeliness in this paper.According to the set of POI generated by user history trajectory,the label scored is firstly carried out,and then the association degree between labels is extracted.After the final consideration,in accordance with the label level after sorting,according to the relevant content of the label recommendations.This algorithm simplifies the calculation of the association degree between labels,and improves the accuracy of the recommended content by making real-time recommendations in conjunction with the timeliness.The two algorithms mentioned above are compared with other methods of interest discovery and personality recommendation,and are mainly compared according to the accuracy,timeliness and other dimensions.The results show that the method proposed in this paper has certain advantages.The accuracy of the method is also taken into account in the analysis of the timeliness of the results and the specific preferences of the user,and combined with the user's own dense time interval to a great extent preserves the context in which the user is located.