Abstract: As an important platform for colleges and universities to absorb social resources, university education foundation undertakes the important responsibility of attracting social donations and expanding social cooperation. How to expand the financing channels, gather the resources of society and alumni, and improve the long-term mechanism of fundraising and fund management are the key issues to enhance the sustainable development ability of university education foundations. In this paper, big data mining technology and artificial intelligence method are used to explore various relevant information of existing and potential donor companies from the open-source data of the Internet. We propose the framework of alumni resource mining and the prediction model of donation intention of listed companies based on distant supervision. This model is different from the traditional qualitative research method and the traditional text mining method suitable for small sample data. It can effectively support the donation information mining in the big data environment, and accurately identify the key potential donation enterprises that are more likely to make donations in the future, so as to provide insights for university foundations to take the advantages of alumni resources more effectively. In the case study of the University of Chinese Academy of Sciences, we use the proposed framework to explore alumni resources and predict the willingness of alumni enterprises to donate. Combined with the actual donation cases, the effectiveness and application value of the proposed method are verified.

Key words: big data technology, university foundation, donation intention prediction model, distant supervision