点击次数：

论文类型：会议论文

发表时间：2015-05-18

收录刊物：EI、CPCI-S、Scopus

页面范围：591-596

关键字：Science Navigation Map; Interactive data mining; Multi-view non negative matrix factorization

摘要：With the advances of all research fields and web 2.0, scientific literature has been widely observed in digital libraries, citation databases, and social media. Its new properties, such as large volume, wide exhibition, and the complicated citation relationship in papers bring challenges to the management, analysis and exploring knowledge of scientific literature. in addition, although data mining techniques have been imported to scientific literature analysis tasks, they typically requires expert input and guidance, and returns static results to users after process, which makes them inflexible and not smart. Therefore, there is the need of a tool, which highly reflects article level -metrics and combines human users and computer systems for analysis and exploring knowledge of scientific literature, as well as discovering and visualizing underlying interesting research topics. We design an online tool for literature navigation, filtering, and interactive data mining, named Science Navigation Map (SNM), which integrates information from online paper repositories, citation databases, etc. SNM provides visualization of article level metrics and interactive data mining which takes advantage of effective interaction between human users and computer systems to explore and extract knowledge from scientific literature and discover underlying interesting research topics. We also propose a multi-view non-negative matrix factorization and apply it to SNM as an interactive data mining tool, which can make better use of complicated multi-wise relationships in papers. In experiments, we visualize all the papers published at the journal of PLOSI3iology from 2003 to 2012 in the navigation map and explore six relationship in papers for data mining. From this map, one can easily filter, analyse and explore knowledge of the papers through an interactive way.