Design And Implementation Of A Overseas Operating System Based On Django

As Chinese companies gradually move overseas,more and more domestic apps are being promoted and operated overseas,but there are also many problems in the process of operation,such as unfamiliarity with the language and culture of overseas countries,lack of data drive,etc.This time,in response to problems in specific operating scenarios,an operating system for overseas operators of an App was created.The thesis starts with the problems that arise in the operation of overseas markets,and conducts a detailed analysis of the needs of operators.Then the overall system architecture is designed according to the requirements.The system adopts the B/S architecture mode and uses UML object-oriented modeling technology to give the system use case diagram.Next,the system function modules are divided into a video search module that crawls videos from different channels in real time,a video scoring module that provides score references for crawled videos,a predictive video tag and column module,and a display article clicks distribution and operation by country The data visualization module and the word generation test module are subdivided into different sub-modules,and the important and difficult points are designed in detail.Among them,the key design of the video scoring module is carried out.The module design chapter describes in detail the complete process of data modeling.By extracting data from actual business scenarios,preprocessing,feature construction and feature screening are performed,and then the limits are used respectively.Forest algorithm,GBDT algorithm,and Stacking model fusion algorithm are used for data modeling,and the stacking model fusion algorithm changes the previous method of fusion using homogenization algorithms,and uses distinct algorithms for fusion,so that in the case of small data volume There is a better learning effect in the next step.After that,the three models are compared and analyzed using the evaluation index R2,and the best model is selected for video scoring.In the video tag column prediction module,the TF-IDF algorithm is used to extract the video text information,and the cosine similarity is used to calculate the similarity value of the video information,and the KNN algorithm is used to predict the video tags and columns.Finally,the system uses Python language and its Web development framework Django to complete the development of the above functional modules,reducing the coupling between levels and making the system more scalable.At present,each main functional module has been initially realized,and has passed the system function test and performance test,and the trial operation is good.In general,this project has initially achieved the effects of improving the work efficiency of operating personnel,reducing labor costs,and raising operating indicators,and it has important reference significance for the operating process of overseas companies.