作者: Xiaojun Wan , Fang Liu
DOI: 10.1002/ASI.23083
关键词:
摘要: Literature citation analysis plays a very important role in bibliometrics and scientometrics, such as the Science Citation Index (SCI) impact factor, h-index. Existing methods assume that all citations paper are equally important, they simply count number of citations. Here we argue not some more than others. We use strength value to assess importance each propose regression method with few useful features for automatically estimating citation. Evaluation results on manually labeled data set computer science field show estimated values can achieve good correlation human-labeled values. further apply evaluating influence author influence, preliminary evaluation demonstrate usefulness