Work

Research Category Classification for Scientific Literature on Human Health Risk of Electromagnetic Fields

Year: 2020

Type: article

Abstract: This study presents a research category classification method for scientific literature on the human health risk of electromagnetic fields. The proposed method uses the document embedding model, Doc2Vec, to convert the research article text data into a vector, through which it trains three binary classifiers to determine whether the article belongs to animal experiments, cell experiments, or epidemiological studies. Finally, by using the trained binary classifiers, the research articles were automatically classified into one of three research categories. The proposed method was implemented for performance evaluation with 120 research articles on the human health risk of electromagnetic fields. Among the 120 research articles, 90 were used as a training dataset, while 30 were used as a test dataset. The implementation results showed that the proposed method can classify the research articles on the human health risk of electromagnetic fields with 88% accuracy on average.

Source: The Journal of Korean Institute of Electromagnetic Engineering and Science

DOI: 10.5515/kjkiees.2020.31.10.839

Authors: Sang-Woo Lee, Jung-Hyok Kwon, Nam Kim, Hyung-Do Choi, Eui-Jik Kim

Institutions: Hallym University, Electronics and Telecommunications Research Institute, Chungbuk National University

Language: English

Topic: Text and Document Classification Technologies

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: diamond

Funder: Ministry of Science ICT and Future Planning

Award: G7287621566