Design And Implementation Of Video Personalized Push System Based On Lbs On Android Platform

The rapid development of mobile Internet makes Internet users transfer from PC to mobile terminal,more and more users are keen to browse video at mobile terminal,which requires the video data push of mobile terminal can be more targeted than the web page,so the mobile terminal of video data mining is particularly important.First,the variability of geographical position of mobile user's is an important characteristic different from the web page,so the data mining of location information can be used as support of mobile video data pushing;second,the current video applications mainly focus on video content aggregation and video search,video recommendation is mostly just from the relationship between video and video,but the data mining of the relationship between users and videos,users and users is not enough.Third,simply rely on users taking the initiative to search or the video app recommended is difficult to take the initiative to remind users concerned about the video updates.We can active to the mobile user of video information by video push system,and increase the interaction between video applications and users.In order to fully exploit the characteristics of the video mobile terminal,this paper combines the video information with geographical location elements,user attributes,designing the video personalized push system based on LBS on the Android platform,to actively attract current user's attention to the video which around people have watched and the current user is interested in.This paper mainly includes the following contents:Firstly,video recommendation based on LBS.This paper collects user location data based on mobile terminal positioning monitoring.Then through the rectangle algorithm,spatial database algorithms and GeoHash algorithms to calculate nearby users.Then it combines the data of nearby users with video click history record to draw near video recommendation results.According to the video number which nearby users have clicked,and the similarity between the current video type and the video type which current user's are interested in,system sorts the results of nearby video recommendation.Secondly,video user clustering.this system collects the mobile user's video clicked data,according to the video classification,video types,year and area of the video produced,and other factors,then uses hierarchical clustering algorithm for video user clustering.Thirdly,video active pushing,this paper designs and implements the video personalized push system's server and client,including disconnection reconnection,offline push,and packet push based on the result of user clustering and other functions.Above targets have implemented in the video push system based on LBS on the Android platform.The related algorithm has been analysed by contrast,the details of various functions has been tested to ensure the availability and reliability of the system.