Extraction knowledge objects in scientific web resource for research profiling

Autor: Jian Xu, Jian-hua Liu, Zhixiong Zhang, Na Hong, Qi Zhao, Si-Zhu Wu, Dai-Qing Yang
Rok vydání: 2009
Předmět:
Zdroj: 2009 International Conference on Machine Learning and Cybernetics.
DOI: 10.1109/icmlc.2009.5212770
Popis: Research profiling is a large-scale analysis method based on the literature information to depict the state-of-the-art scientific researches wisely. Much research profiling work has been carried out based on formal, structured scientific publications (such as Chemical Abstracts) while there is much information contained in abundant, unstructured scientific web resources. In order to profiling research based on those unstructured scientific web resources, the authors bring forth a solution that tries to extract useful knowledge objects from them. Three kinds of knowledge objects have been extracted: (1) research related objects (research objects), such as institutes, scientists, projects, conferences etc.; (2) the relationships between the research objects which reflected in scientific web resources, such as one scientist worked in one institute; (3) the terms which indicate the topic of the research areas. The authors implemented the knowledge objects extraction system and did some experiments to test and evaluate the effect of this system.
Databáze: OpenAIRE