Design And Implementation Of Resource Release System Based On Personalized Recommendation

In recent years,with the development of technologies such as computers and information,the data generated by users online has increased day by day.On the one hand,it is difficult for users to directly filter out the data information they need from a large amount of information;on the other hand,users have higher and higher requirements for client product experience.To improve the user’s experience of the client product requires constant updates to the client.It is very important to quickly update some static resources of the client and extract data and resources that users are more likely to be interested in from a large amount of data resources personalized for the client users.The system includes a resource allocation platform,resource distribution,and resource data analysis.According to the relevant software development process,the system has been subjected to related demand analysis,system design,implementation and testing.In the requirements analysis stage,the descriptions are made from functional requirements,nonfunctional requirements,and feasibility.Then the system is divided into R&D configuration module,resource clustering module,scoring prediction module,resource issuing module and resource cleaning module and detailed design is carried out.In the system implementation stage,screenshots of the core operations and the implementation process are given.The distribution part provides resources to pull resources according to the latest compliance rules,which can quickly update resources to the client,thereby solving the problems caused by the long release cycle of the client.In addition,TF-IDF and K-Means are used to perform clustering analysis for texts,and provide resource distribution by category,which is convenient for the client to use resources in some scenarios.In addition,combining matrix decomposition to analyze the feedback data when users use resources,predict the scores of unused resources for users,so as to deliver resource data with higher scores,which can achieve the effect of filtering information for users.Finally,a black box test and performance test were performed on each module of the system to ensure the reliability and stability of the system.The system can quickly configure resources and send them to the client,which solves some problems caused by the long release cycle of the client.At the same time,through text clustering and matrix decomposition,the system provides R&D personnel with a personalized way to obtain resources,which greatly improves the client user experience.