Collaborative Filtering of Call for Papers

Call for papers (CFPs) are notifications of academic events that invite researchers to submit their works. Traditionally, CFPs are handed out to researchers by mailing lists and web pages. With the number of conferences increasing, finding, reading and filtering out relevant CFPs become time consuming and need the assistance from information retrieval techniques. In this paper, we employ collaborative filtering to match relevant CFPs to researchers. Non-personalized, neighborhood-based and class-based methods are applied in CFP recommendation. We also propose a hybrid approach that utilizes conference series and submission deadlines of CFPs. The experiments on Wiki CFP data set show that the class-based method outperforms both neighborhood-based and non-personalized methods, whereas the proposed hybrid approach has the best overall performance.