Programmer Recommendation System Based On GitHub

With the development of market-oriented economy,human resources are gradually concerned by all sectors of society.In the Internet industry,programmers are not only the basis for enterprises to develop new products,but also the indispensable role to maintain the smooth operation of software.However,due to the rapid development of technology and the change of programmers from time to time,the demand for all kinds of programmers is increasing,and the competition for high-quality programmers among enterprises is becoming increasingly fierce.However,the traditional headhunting recruitment and company recruitment have been difficult to meet the current needs of enterprises.Therefore,based on GitHub,one of the favorite gathering places for programmers,this paper designs a programmer recommendation system based on GitHub,which can provide decision-making reference for enterprises to select excellent programmers to some extent.The programmer recommendation system based on GitHub regularly collects the activity information of registered programmers in GitHub,selects the available programmer information for recruitment through preprocessing,and extracts the keywords to build the programmer database.On the basis of this talent pool,the development ability of programmers is evaluated,and according to the recruitment Posts published in the system,excellent programmers who meet the post conditions are recommended to enterprisesFirst of all,the thesis calls the method of combining the rest API(representative state transfer application programming interface)and crawler of GitHub to collect the project profile of programmer's mailbox,company,repositories,daily contributions and project repository README.md Documents and other information.Using TF-IDF(term frequency-inverse document frequency)algorithm to extract project introduction and README.md Keyword are used as tags to define the technical direction of programmers,build a talent pool for programmers,and extract keywords from post details published by enterprises as tags to build a job pool.Then the paper uses the analytic hierarchy process to analyze the importance of the collected data and determine the data weight,so as to establish five kinds of evaluation system for the development ability of programmers.On this basis,the convolution neural network is used to complete the dynamic evaluation of historical development ability,and the short-term memory network is used to complete the prediction of ftiture development ability.Then,on the one hand,the system builds a vector space model to transform the position and the programmer's label in the talent pool into feature vector,and uses the content-based recommendation algorithm to calculate the cosine similarity of the two to get the candidate programmer list;On the other hand,based on the enterprise^ selection preference,it builds the PLSA(probabilistic latent semantic analysis) model.The collaborative filtering algorithm is used to predict the probability of recruitment of candidate programmers,so as to obtain the recommended value required by the collaborative filtering based recommendation algorithm,and the total recommended value obtained by the combination of the two is inverted to generate a recommendation list.Finally,recommender list is rearranged according to recommender list and programmer development ability level.On this basis,the thesis implements the programmer recommendation system based on GitHub,and obtains the recommended programmer information by simulating the post and recruitment behavior of enterprises in the system.Finally,the integrity of the system function is verified by the system test,and the evaluation index of the system is obtained by comparing the conformity of the recommended programmer and the post.