Construction and Application of Innovation and Entrepreneurship Sharing Platform Based on Attention-Mechanism Graph Convolutional Network

Abstract

To promote the efficient development of innovation and entrepreneurship (IE), the study calculates the node attention weights by introducing the attention mechanism to learn the node embedding representation and combines the Django-Pycharm technology to design the function, architecture, and database of the IE sharing platform. The data show that the proposed recommendation model performs better with 98.5% accuracy, 97.6% recall, and 98.2% F1 value. Compared with the graph neural network recommendation model, the F1 value of the model is improved by 3.1% with 160 users, the system response time of this IE sharing platform is in the range of 8–298 s, and the throughput is about 91 per minute. The overall performance of the sharing platform meets the expected requirements, with good reliability and stability, and is able to satisfy the users’ needs. The overall performance evaluation of the shared platform is higher than 91 points, and the overall performance and stability are affirmed by users. The study shows that the sharing platform provides an efficient, convenient, and safe environment for IE, helps to improve the success rate of IE, and provides a research hotspot and cooperation opportunity for academia and industry.