Research On Mobile Internet Data Resource Management And Content Recommendation Technology

This topic is based on the mobile Internet resource management platform, which is based on the exchange of information between the network server and the mobile device. Users can use the mobile terminal of the APP can receive the server to accurately push the content, but also to the server to submit the use of mobile phones, geographical location, the use of APP applications and other information. To push the precise content to the user, this platform needs to quantify the analysis and management of mobile users to use the mobile device data characteristics, and calculate the conditions to push the content of the user.The contents of the main research topics is the users in mobile Internet environment using a mobile phone data analysis and classification of storage, then prepare content sequence database, finally according to the user use mobile phone automatically from the content library screened for user to push in content, and calculation of similar users selected to fit the contents of the group and automatically pushed to the user.First of all, according to the characteristics of the location and real-time of the data type, the data is divided into static characteristic data and dynamic characteristic data. The static characteristic data is divided into three dimensions:gender, age, and terminal type. Due to small changes in the data, can be unified modeling management. For dynamic feature data, it is divided into the user’s habit dimension, the preference dimension and the position dimension. In different dimensions, the data is divided into different categories, and the unified coding is carried out.Secondly, for the processed data, according to the label and the weight of the way, in order to form a monthly vector storage, the establishment of the user’s characteristic model. Using the top-N and user preferences personalized recommendation, association analysis content recommendation algorithms on the user’s behavior analysis, and according to the set the confidence threshold, screening out the contents need to be pushed to the user, finally to automatically push the form and the model choice of content pushed to the user.Finally, through the combination of the actual project to verify, so as to reflect the correct application and value of content recommendation. This process uses the Echart chart to display the process of filtering the content of the push content, making the content of the recommended process intuitive. And by way of timing task, automatically screening results pushed to the user, mobile display terminal app content and server side filtering content of a sequence corresponding to verify correctness of the push, the test proof, and prediction results are consistent.