Oral vs Poster Publication Influence Analysis

Many top CS conferences in Artificial Intelligence assign accepted papers into different presentation types, mainly orals and posters. There is a perception that oral papers are better, however some paper with high influence later were also accepted as poster. To end this discussion, in this paper, we proposed a framework to analysis the influence of presentation type and other factors on academic works. To be specific, statistic analysis and significant test are done based on publication’s citation count. Further, a better incluence representation was proposed based on the ranking methods in [15]. In which influence level was calculated based on the influence of other works and authors in its citation network, instead of simple applying citation count as an influence measure. According our analysis on conference, we show that oral are statistically better than poster in most conferences. Also our ranking method can better represent paper’s influence comparing to simply using citation count.