A new strategy for developing scientific community model based on scientific article publication

摘要

The study was a literature review to develop a new strategy to clustering data. This study is part of our ongoing main project to develop scientific community model based on scientific article publication. The concept of this strategy is utilizing scientific article metadata elements that has been published electronically to map scientists community respect to the topics that are developing. This paper introduces TermRank approach that is adopted from PageRank to determine the strength value of a node and implements clustering technique to limit the number of computation in the topic modeling. Based on preliminary results, it is shown that in the stage of article data extraction, it turns out that it must be readjusted by adding the number of process to solve the problems on inconsistency inputting metadata elements, the use of two languages in abstract, and non-standard references format.