摘要：Open-source has become a very important topic in this era. the number of open-source projects on github Shows a huge growth trend. Facing so many open-source projects, it’s not easy to find the projects and topics that the developers interested in. so, it is necessary to model the user's behavior data,So as to automatically recommend projects to developers. to explore this problem, we constructed a dataset of 90w users and 461w projects based on github log and did a lot of cleaning work on the data. finally, we model the data through the improvement of the Light-GCN model to recommend relevant open-source projects to users. The experimental results show that the accuracy of our model is more than 15%.
