A Study of Innovative Entrepreneurial Behavior of College Students under Algorithmic Recommendation

In this paper, for a large amount of text content such as college students’ innovation and entrepreneurship education practice, considering the problem of directionality, based on the idea of a recommendation algorithm, we use word similarity and TextRank algorithm to extract the keywords of innovation and entrepreneurship behaviors as the first-knowledge labels of the textual resources, and introduce complex labeling network based on this. On the premise of semantic-based TextRank keywords and tag network, the fusion-gated graph attention group recommendation algorithm is used to construct college students’ innovation and entrepreneurship practice platform, and the simulation analysis of college students’ innovation and entrepreneurship practice platform is carried out. The results show that the survival rate of virtual startups can be calculated as 5.37% by using the college students’ innovation and entrepreneurship practice platform, which can accurately analyze that college students’ innovation and entrepreneurship behaviors are stronger, but the survival rate of the enterprises founded by college students for the first time needs to be improved. This study can provide theoretical knowledge guidance for the study of college students’ innovative entrepreneurial behavior and has a facilitating effect on the development of innovation and entrepreneurship education in colleges and universities.