Àá½Ã¸¸ ±â´Ù·Á ÁÖ¼¼¿ä. ·ÎµùÁßÀÔ´Ï´Ù.
KMID : 1023920140160010031
Journal of the Korean Academy of Kinesiology
2014 Volume.16 No. 1 p.31 ~ p.39
Development of O2max Estimation and Validity Based on Body Index Variables for Adults
Jeon Yoo-Joung

Lee Byung-Kun
Im Jae-Hyeng
Abstract
INTRODUCTION: The purpose of this study is to develop O2max estimation model with body index variables such as gender, age, weight, height, BMI, and resting HR, and to verify prediction model bases on verification group.

METHOD: The subjects are consisted of 837 male and female subjects aged from 18 to 60, and they were separated into two groups randomly: 649 for sample group and 188 for cross-validation group. They went through maximal exercise testing using Bruce protocol and we ran applied multiple regression analysis to the sample group.

RESULT: Estimation prediction initially used gender, age, weight, height, BMI, and resting heart rate as independent variable, but only gender, age, weight, and height were chosen to be input variables. As the result, 2 models were developed, and both of them had good r value(0.66, 0.68, p<.01), low SEE%(14.32, 14.59, P<.01), and low %TE(5.89, 5.99, p<.01). In addition, multicolinearity didn't show on 2 models. Model 1 was O2max(ml/kg/min) = 27.703 - 7.423 * (gender) - 0.067 * (age) - 0.232 * (weight) + 0.221 * (height). Model 2 was O2 max(ml/kg/min) = 64.543 - 9.397 * (gender) - 0.110 * (age) - 0.167 * (weight). For 2 models, Cross-validation also showed correlation between predicted and measured O2max(r=0.56, 0.58, p<.01), and there was no significant difference(p=0.443, p=0.847) with very low %error(-18.8~12.2) and %TE(5.89, 5.99, p<.01).

CONCLUSION: O2max estimation 2 model for adults were developed based on body index variables. As a result of verification, these model show valuable cross-validation. In the future studies, the higher prediction power model with questionnaire or convenient scientific methods should be developed.
KEYWORD
estimation of O2max, multiple regression analysis, cross-validation
FullTexts / Linksout information
Listed journal information
ÇмúÁøÈïÀç´Ü(KCI)