Evolutionary Analysis of International Student Mobility Based on Complex Networks and Semi-Supervised Learning

With the expansion of globalization, the internationalization of education has become an essential strategy for developing in various countries. To obtain higher education, more and more students decide to study abroad. Many countries have been recruiting talented people to promote education and technology innovation in recent years. In particular, emerging and developing countries have improved systems on talent introduction. Meanwhile, the internationalization of education has become one of the important factors affecting the economic development of a country. Thus, the environment for talent development has been improving, and the international talent flow mode has become increasingly diversified, and it is of certain guiding significance to analyze and formulate corresponding laws according to the behavior of foreign students. This article constructs a weighted directed network to analyze the network's evolution by global student mobility data from 2007 to 2016. The results are as follows: first, the network has a small average path length and significant average clustering coefficient, showing small-world characteristics. Second, the density shows a trend of first reducing and then rising. It explains that emerging countries also improve the ability to attract talents, and interactions between countries are becoming more frequent. Moreover, degree and degree centrality gradually increase, indicating that the number of students is rising, and the study abroad path is also growing in the network. Finally, the division of community based on the label propagation and semi-supervised learning to demonstrate the club's change reflects the communications of these communities during the decade.