Digital construction of higher education management based on multimodal machine database

In this paper, we first study the multimodal interaction activities in higher education management from the two points of multimodal interaction types and multimodal interaction activity evaluation and propose the perfect way of digital construction of multimodal higher education management from the perspective of students, teachers and higher education institutions. Then, the higher education management system is constructed based on semi-supervised fusion feature learning and homogeneous multimodal features of multimodal machines, and the system architecture and database design are explained in detail. Next, the research literature statistics and system performance of ‘Digitalization of Higher Education Management’ are analyzed through experimental research. The results show that the periodicals’ literature statistics show a rapid increase in articles, mainly from 2003 to 2004 and 2006-2007. The higher education management system maintained an accuracy rate of 80.49% regarding system performance. It is confirmed that the management system constructed in this study can predict the learners’ status more accurately and predictably, contributing to the development of intelligent education management in higher education institutions.