An automatic literature knowledge graph and reasoning network modeling framework based on ontology and natural language processing
Hainan Chen,Xiaowei Luo
DOI: https://doi.org/10.1016/j.aei.2019.100959
IF: 8.8
2019-10-01
Advanced Engineering Informatics
Abstract:<p>With the advancement of scientific and engineering research, a huge number of academic literature are accumulated. Manually reviewing the existing literature is the main way to explore embedded knowledge, and the process is quite time-consuming and labor intensive. As the quantity of literature is increasing exponentially, it would be more difficult to cover all aspects of the literature using the traditional manual review approach. To overcome this drawback, bibliometric analysis is used to analyze the current situation and trend of a specific research field. In the bibliometric analysis, only a few key phrases (e.g., authors, publishers, journals, and citations) are usually used as the inputs for analysis. Information other than those phrases is not extracted for analysis, while that neglected information (e.g., abstract) might provide more detailed knowledge in the article. To tackle with this problem, this study proposed an automatic literature knowledge graph and reasoning network modeling framework based on ontology and Natural Language Processing (NLP), to facilitate the efficient knowledge exploration from literature abstract. In this framework, a representation ontology is proposed to characterize the literature abstract data into four knowledge elements (background, objectives, solutions, and findings), and NLP technology is used to extract the ontology instances from the abstract automatically. Based on the representation ontology, a four-space integrated knowledge graph is built using NLP technology. Then, reasoning network is generated according to the reasoning mechanism defined in the proposed ontology model. To validate the proposed framework, a case study is conducted to analyze the literature in the field of construction management. The case study proves that the proposed ontology model can be used to represent the knowledge embedded in the literatures' abstracts, and the ontology elements can be automatically extracted by NLP models. The proposed framework can be an enhancement for the bibliometric analysis to explore more knowledge from the literature.</p>
engineering, multidisciplinary,computer science, artificial intelligence