keyword association rule, keyword network analysis, data mining, future signal
Measurements of physical activity taken in a valid and reliable way are essential in characterizing the relationship between physical activity and health outcomes. Given the steadily growing interest in the physical activity measurement and the lack of research to identify current trends, this study investigated the research trend of physical activity measurement by applying four text data mining techniques (i.e., future signal, keyword network analysis, keyword trend, and keyword association rule). A total of 54,670 publications from 1982 to 2021 were collected from PubMed. As a result, the current study 1) confirmed two weak signal topics (i.e., “validity of physical activity instrument” and “classification of physical activity patterns using machine learning algorithms”) that are likely to affect future research trends, 2) identified keywords (e.g., “youth,” “adult,” “woman,” “survey,” “questionnaire,” and “monitor”) from the perspective of populations and measurement tools, 3) examined that the relative importance of keyword, “senior” increased rapidly, and 4) indicated that new keywords (i.e., “smartphone,” “wearable device,” “GPS,” “tracker,” and “app”) appeared in the early 2000s. The findings of this study provided implications for the selection of research topics and the use of text mining techniques in physical activity measurement research.
Lee, Seungbak and Kang, Minsoo
"Trend Analysis of Physical Activity Measurement Research Using Text Mining in Big Data Analytics,"
International Journal of Physical Activity and Health: Vol. 1:
1, Article 5.
Available at: https://scholarworks.boisestate.edu/ijpah/vol1/iss1/5