Physical Therapy Exercises
Donated on 5/7/2022
This dataset contains wearable inertial and magnetic sensor data during the execution of physical therapy exercises. There are eight types of physical therapy exercises, each of which has three execution types (correct, fast, and low-amplitude). Each execution type of each type of exercise was performed multiple times by five subjects. The subjects wore five MTx sensor units manufactured by XSens. Each unit contains three tri-axial sensors: an accelerometer, a gyroscope, and a magnetometer, sampled at 25 Hz.
Dataset Characteristics
Time-Series
Subject Area
Engineering
Associated Tasks
Classification, Clustering, Other
Feature Type
-
# Instances
268881
# Features
-
Dataset Information
For what purpose was the dataset created?
The dataset was created within a study on automated detection of the executions of physical therapy exercises using wearable motion sensors. The exercises and the experimental procedure are determined in collaboration with a physical therapy specialist.
What do the instances in this dataset represent?
The dataset contains time series data recorded during training sessions (with time indices of the selected templates) and simulated exercise sessions (with multiple exercise executions). The sessions exhibit pre-determined patterns, which are explained in the Description.pdf document.
Are there recommended data splits?
The dataset is already split into training and test sets. The former contains template recordings and the latter contains the simulated physical therapy exercise sessions.
Was there any data preprocessing performed?
The time series data are filtered by the software of the sensor manufacturer (XSens Technologies) using the default settings during the data acquisition. No additional processing is applied.
Additional Information
Additional information can be found in the Description.pdf file and in the papers listed in the Citation Request field.
Has Missing Values?
No
Introductory Paper
By Aras Yurtman, B. Barshan. 2014
Published in Comput. Methods Programs Biomed.
Dataset Files
| File | Size |
|---|---|
| s3/e3/u3/test.txt | 638.9 KB |
| s3/e8/u3/test.txt | 638.8 KB |
| s3/e2/u3/test.txt | 637.3 KB |
| s2/e2/u3/test.txt | 636.2 KB |
| s3/e4/u3/test.txt | 633.9 KB |
0 to 5 of 441
Reviews
There are no reviews for this dataset yet.
pip install ucimlrepo
from ucimlrepo import fetch_ucirepo # fetch dataset physical_therapy_exercises = fetch_ucirepo(id=730) # data (as pandas dataframes) X = physical_therapy_exercises.data.features y = physical_therapy_exercises.data.targets # metadata print(physical_therapy_exercises.metadata) # variable information print(physical_therapy_exercises.variables)
Yurtman, A. & Barshan, B. (2014). Physical Therapy Exercises [Dataset]. UCI Machine Learning Repository. https://doi.org/10.24432/C5JK60.
Keywords
Creators
Aras Yurtman
aras.yurtman@kuleuven.be
KU Leuven
Billur Barshan
billur@ee.bilkent.edu.tr
Bilkent University
DOI
License
This dataset is licensed under a Creative Commons Attribution 4.0 International (CC BY 4.0) license.
This allows for the sharing and adaptation of the datasets for any purpose, provided that the appropriate credit is given.