Isometric Plantarflexion Moment Prediction SVR

Published: 3 August 2021| Version 3 | DOI: 10.17632/hzjsr56fkb.3
Contributor:
Qiang Zhang

Description

1. Data collection during isometric ankle joint plantarflexion at 5 different postures, incluidng sEMG raw signals, ultrasound imaging videos, joint moment measurements from nine able-bodied participants. 2. sEMG signal time-domain features extraction and ultrasound imaging structural and functional features extraction results. 3. Correlation analysis results between each neuromuscular feature and measured ankle joint plantarflexion moment. 4. Illustrative support vector machine regresion (SVR) and deep feedforward neural network (FFNN) models training and prediction code and results, including the saved results in .mat files, root mean square error and R-square values between trained and measured joint moments, as well as between predicted and measured joint moment. 5. SVR and FFNN regression models training and prediction results summary for individual participant.

Files