<原著>生理的年齢予測の重回帰分析

概要

論文の詳細を見る
The purpose of this study was to establish the multiple regression equation to predict females' physiological age. Data for this study were collected during the course of survey of various indices on the degree of health and fitness and life-style of healthy women who aged 23 through 59 years old. Collected information included the followings :systolic blood pressure (=SYSBP), plasma creatinine (=CREAT), plasma albumin (=ALBUM), sum of skinfold thickness of upperarm and subscapulars (=FAT), grip strength (=GRIP), back strength (=BACK), standing trunk flexion (=FLEX), vital capacity (=VITAL), one-foot blind balance (=BALANCE), maximal aerobic power (=MAP, by Margaria's method), and chronological age (=AGE; dependent variables). Forward stepwise regression analysis examined the reliability of the physiological and internal medicine and anthropometric measurements as the indicator of age. At the each step, the statistics of degree of fitness, that is. Akaike's AIC, Mallows' Cp and R* (adjusted for d.f. R) adding to multiple correlation coefficient R and SEE (standard error of estimates) were calpulated. The results obtained were as follows. 1) In terms of following variables order, all variables were entered into equation ; SYSBP, BALANCE, VITAL. FAT, ALBUM, CREAT, FLEX, BACK, GRIP. The multiple correlation coefficient R, and standard error of estimates (SEE) were 0.738, 6.1 (age of years), respectively. 2) Each equation obtained at each step, which included m independent variables at m step, in general, were evaluated by several statistics such as Akaike's AIC, Mallow's Cp. R* and R. According to Cp, the equation III which had SYSBP, BALANCE and VITAL was evaluated as the best equation. According to AIC, the equation IV which was added FAT to preceding equation (III) was evaluated as best. 3) By the forward stepwise regression analysis, of which stopping rule for variable selection was F-statistics=2.0, equation IV was selected as the best equation. Equation IV was descrived as follows and R=0.721, SEE=5.9 (age of years). , respectively. Y=24.1＋0.182×SYSBP−0.064×BALANCE−0.0030×VITAL＋0.111×FAT 4) The high reliability was verified by analysis of residuals. And, the validity of this equation was verified by the results of preceding studies. 5) When MAP was added to these independent variables, at the 1st step MAP entered into the equation by the forward stepwise method (FENTER-2.0). SYSBP at the 2nd step, and BALANCE at the 3rd step were respectively entered. In this case, multiple correlation coefficient was 0.773. However, the calculated MAP was unequal to other variables as for the indicator of age changes because this variables were essentially designed to correlate to chronological age.
九州大学の論文
1985-03-30

<原著>生理的年齢予測の重回帰分析

スポンサーリンク

概要

著者

関連論文

スポンサーリンク