0% Complete
صفحه اصلی
/
ورود به سیستم
/
سوالات متداول
/
تماس با ما
صفحه اصلی
/
هفتمین کنفرانس ملی و اولین کنفرانس بین المللی محاسبات توزیعی و پردازش داده های بزرگ
کاربرد تجزیه و تحلیل داده های تابعی ناپارامتری در بیومکانیک ورزشی پردازش داده های بزرگ
An Application of Nonparametric Functional Data Analysis in the Sport Biomechanics Big Data Processing
نویسندگان :
Mohammad Fayaz ( دانشگاه علوم پزشکی شهید بهشتی ) , Seyed Mehran Hosseini ( دانشگاه علوم پزشکی گلستان )
کلید واژه ها :
Big data،Functional data analysis،Curve registration،Sport biomechanics،Open dataset
چکیده مقاله :
Abstract— Big data has 5Vs (Volume, Variety, Velocity, Veracity, and Value) that are present in many fields like sports biomechanics with devices such as wearables, markers and etc. Among different statistical methods for big data analysis, functional data analysis (FDA) provides a methodology to analyze the dataset by considering underlying functions and curves. The open dataset includes 57 healthy subjects with 54 markers among the full-body and three-dimensional ground reaction force (GRF) that is a functional predictor. The GRF curves have registered with three methods: Group-wise function alignment and PCA Extractions, Square Root Velocity Function (SRVF), and Bayesian-SRVF and compare their result with two indices: The synchronization (Sync) coefficient and the inverse of pairwise correlation (IPC). And four functional prediction methods considering weight, height, and body mass index (BMI) as continuous responses are used including regression, median, mode, and multimethod (nonlinear) operator. Two kernel functions (Triangle and Quadratic), three semi-metrics (first and second derivative, PCA) with a local selection of the optimal numbers of the neighbors for kernel estimation are compared to each other. The result showed that the Mean Square Error (MSE) in group-wise function alignment for BMI in the second derivative semi-metric with the quadratic kernel are lowest: regression (1.03,4.32 - 1.36,5.69), conditional mode (1.72,7.02 - 0.97,4.07), conditional median (1.47,6.15-1.83,7.65), multimethod (4.43,5.64- 4.98,4.53) for train, test and left-right GRF dataset. We conclude that the nonparametric FDA and FDA methods provide comprehensive methods for studying the complex relationship in big data and they are popular in the sport biomechanics literature.
لیست مقالات این رویداد
Samin Hamayesh
