Design And Implementation Of Internet Technology Article Recommendation System Based On Community Discovery

With the advent of the Internet era,the number of Internet users in China is increasing day by day.While the Internet has brought great convenience to people,there is also information overload,which prevents people from obtaining the required information quickly and efficiently.Recommender systems have become an effective way to solve the problem of information overload.one of the methods.The rapid rise of the IT industry has made more and more people willing to work in the computer industry.At the same time,industry practitioners need to continuously learn new technical knowledge to cope with the rapid update and iteration of IT technology.Faced with this demand,the technical article recommendation platform came into being.However,there are still many problems in the domestic mainstream recommendation platforms,and the targeted recommendation for users is not obvious enough.Meanwhile,traditional recommendation algorithms are relatively simple.For example,content-based recommendation cannot capture changes in user interests in time,and the similarity calculation in collaborative filtering algorithms is relatively rough and computationally intensive.In view of the above problems,combined with the actual project during the internship of an enterprise,in order to develop a recommendation system for Internet technical articles that meets the needs,the main research contents of this paper are as follows:In order to improve the problem of rough similarity calculation,this paper combines the idea of network community with the collaborative filtering algorithm,and proposes a recommendation algorithm based on community discovery.First,a text model is built to calculate the similarity between articles.At the same time,a user behavior matrix is established.The similarity between users is calculated using the improved cosine formula,and the two are weighted to obtain the comprehensive similarity of users.Set the similarity threshold and establish a user relationship network model..Secondly,the model is divided into communities according to the Louvain algorithm.After the division,users in the same community have similar interests,and the collaborative filtering algorithm is used to recommend articles to users in the same community.When selecting the neighbor user set,it is no longer necessary to calculate the similarity between the users in the community instead of the entire network,which reduces the calculation range of neighbor users.Finally,through the experimental evaluation on Zhihu Rec and Caixin news datasets,it is proved that this method can effectively improve the accuracy of neighbor user set selection and article recommendation.In this paper,a detailed design of an Internet technical article recommendation system based on community discovery is also carried out.The system includes five modules: data crawling module,data preprocessing module,community model building module,community discovery and recommendation algorithm module and recommendation display module.The data crawling module provides the underlying data support for the system.It uses the Scrapy framework to crawl articles under each column of the CSDN blog platform,develop crawler scripts and implement GUI interfaces.The data preprocessing module is mainly used to extract the feature words of the article.The community model establishment module establishes the user community model according to the designed similarity calculation method and the set similarity threshold.The community discovery recommendation algorithm module uses the Louvain algorithm to divide the model into communities and filters out the neighbor user set,and then uses the collaborative filtering algorithm to generate a personalized recommendation list for users in the community.It is recommended that the presentation module is responsible for page rendering of the data.Finally,according to the above design and technology,this paper implements an Internet technical article recommendation system based on community discovery,and through system testing and verification,the function and performance of the system meet the expected requirements.