Research And Implementation Of Algorithm Evaluation Platform For Social Network Analysis

In recent years,the development of Social Network Analysis has been very rapid,which mainly due to social software user's explosive growth,resulting in a large number of social networks.This phenomenon has brought a lot to the researchers of the study sample,prompting some excellent and efficient algorithms have been born.In the study,people often need to analyze the comparative conventional Social Network Analysis algorithm at first,then present a new algorithm or improvement of some existing algorithms on this basis.But for a long time,there isn`t a general management,operation and evaluation of social network analysis algorithm platform for researchers.In this paper,by analyzing WebService technology,we designed a platform for a Social Network Analysis algorithms.The platform uses RESTful style Web application design,which directly expose the presence of resources to achieve the transparency and self-explanatory;stateless API design can not consider the context,greatly reducing the complexity of development;taking full advantage of the HTTP protocol semantic achieve a low coupling system.In addition,with the combination of Docker virtualization technology,this paper constructs a standard Docker image of the evaluation platform,which contains a number of dependent libraries,and it can be run directly to support most of the Social Network algorithms;on this basis,we design a distribution type of algorithm execution engine Worker,which can talk with the main server to interact,and accept tasks submitted by users to run.Finally,we use the Python programming language,combined with several excellent open source framework to achieve the development of the system.By running a number of examples of social network analysis algorithms to verify the feasibility and correctness of the system.