With the development of Internet technology, the phenomenon of information\noverload is becoming more and more obvious. It takes a lot of time for users to\nobtain the information they need. However, keyphrases that summarize document\ninformation highly are helpful for users to quickly obtain and understand\ndocuments. For academic resources, most existing studies extract keyphrases\nthrough the title and abstract of papers. We find that title information in\nreferences also contains author-assigned keyphrases. Therefore, this article\nuses reference information and applies two typical methods of unsupervised\nextraction methods (TF*IDF and TextRank), two representative traditional\nsupervised learning algorithms (Na\\"ive Bayes and Conditional Random Field) and\na supervised deep learning model (BiLSTM-CRF), to analyze the specific\nperformance of reference information on keyphrase extraction. It is expected to\nimprove the quality of keyphrase recognition from the perspective of expanding\nthe source text. The experimental results show that reference information can\nincrease precision, recall, and F1 of automatic keyphrase extraction to a\ncertain extent. This indicates the usefulness of reference information on\nkeyphrase extraction of academic papers and provides a new idea for the\nfollowing research on automatic keyphrase extraction.\n