Differences between the Quantitative Features of Articles from Biomedical Engineering and the Medical Sciences Aimed at Effective Retrieval of Literature

Motoki Sakai

The field of biomedical engineering (BM) is concerned with solving medical problems in diverse ways, for example, by using information technology (information technology-based BM is the focus of this study.) BM research is linked to both information technology and medical (ME) research, and both BM and ME research papers contain common terms such as disease names. However, novice BM researchers have difficulty in searching for and retrieving BM articles because of this. This study attempted to overcome this difficulty by adopting a text mining-based method for classifying BM and ME articles to investigate the quantitative features of these articles. The classification of BM and ME articles was examined by collecting five types of articles containing the common terms obstructive sleep apnea syndrome (OSAS), T-wave alternans (TWA), late potential (LP), epilepsy (EPY), or event-related potential (ERP). Gathered BM and ME articles were quantitatively converted into document-term matrices (D-T matrices), following which the term-term matrices and the two feature values, the norm of the term-term matrix and the first principal component of the D-T matrix, were determined for the classification. The classification was conducted through the use of the fact that two of the feature values of the BM articles were smaller than those of ME. As a consequence, BM and ME articles could be distinguished with 96% averaged sensitivity and with 82% averaged specificity, thereby verifying the effectiveness of the proposed method.

Select Volume / Issues:

• Volume 6,Issue 6, Nov. 2015

Year:

2015

Type of Publication:

Article

Keywords:

Biomedical Engineering; Medical; Classification; Text Mining

Journal:

IJECCE

Volume:

6

Number:

6

Pages:

664-670

Month:

November