의사결정나무분석을 이용한 노인 낙상관련 체력요인의 기여도 예측 모형

Abstract

This study measured the factors of fall that Brown (1999) suggested targeting 39 aged people of over 65 years old with experience of fall and 68 with no experience of fall in order to develop a contribution level prediction model of physical fitness factors related with fall, and conducted the senior fitness test of Rikli & Jones (2001). Danger factors of fall was put into a diagram using decision tree analysis based on the measured data, and a prediction model was established by classifying few influential subgroups. As a result, one leg standing, extensor endurance, flexibility, reaction time, age, TGT, cadence had important effects when establishing a prediction model without classification setting. As a result of establishing a model after setting age as the first classification standard in order to build a prediction model of danger factors of fall according to gender, one leg standing of measuring static balance ability was shown as a variable for both men and women as in the model that was established without classification setting. With one leg standing, which is the static balance ability, as the first factor, reaction time, step length, 2 minute gait, and TGT were taking a great ratio among danger factors of fall for men. One leg standing, extensor endurance, flexibility, and reaction time were taking an important ratio as prediction variables of danger factors of fall for women.