Attractive community detection in academic social network

Academic social network analysis has attracted significant attention. For each researcher in such a network, he/ she has several research interests. We regard these researchers sharing common interests as a research community. For each community, it may be attractive or not to researchers from other communities. In this paper, we study a new and interesting problem: which is the most attractive research community in the academic social network? Here, attractive research communities are those potentially valuable and increasingly popular communities, which are different from hot communities. To address this problem, we first extract both of the internal and external features of attractive research communities. The internal feature refers to the novelty of the topic in the research community and the external feature refers to the researchers? transition among the communities. Intuitively, a community with a novel topic attracts the researchers from other research communities can be considered as the attractive community. Based on the extracted features, we design a novel Attractive Research community Ranking (ARTRank) algorithm to rank the research communities. The core idea of this algorithm lies in two measurements for each community: a positiveness score and a negativeness score, which measure the attractiveness of a community from the in-attention aspect and the out-attention aspect, respectively. Similar to HITS, these two scores are calculated in an iterative way until convergence. Through extensive experiments, we show that our proposed algorithm significantly outperforms the state-of-the-art algorithms in terms of the recommendation intensity.