A Stratification Model For Physical Fitness In The Elderly Based On Walking Speed

With the worsening of the aging society, the health of the elderly had become animportant factor that restricts the development of the family and society. The goodphysical fitness is vital for the aged staying health and maintaining self-care. Physicalperformance is the most popular way to evaluate physical fitness and it needs to test atmany activities but without considering the important feature that related to physical fitness.In this context, a stratification model for physical fitness in the elderly based on walkingspeed was built and it can provide technical support for evaluating physical fitness in theelderly.For complex features relate to walking speed for aged, a method has been proposedfor extracting the important features. Firstly, the population was stratified and four groupswhich are the whole group without stratification, fast speed group (we call it A), usualspeed group (B) and slow speed group (C) were obtained. Multiple regression, randomforest and neural networks were used to select the important features for each group andthen integrate the important features. At last, health status, age, arthritis, cognitive,carnivorous frequency, physical test, weight, height, waist, hip circumference and systolicblood pressure are selected as the important features of walking speed.A stratification model for physical fitness in the elderly was proposed based on fast–usual speed logistic regression and usual–slow speed logistic regression with11importantfeatures and the three groups were divided to seven fitness grades that are A1, A2, B1, B2,B3, C1and C2. Results shows that there is statistically significant difference onquantitative attributes and disability status in the seven grades and the trends with sevengrades are the same. This indicates the rationality of the model so that it can provide amethod for evaluating the physical fitness and selecting intervention targets...