Dataset: Horse Movement Data and Analysis of its Potential for Activity Recognition

Jacob Wilhelm Kamminga; Nirvana  Meratnia; Paul J.M. Havinga

doi:10.1145/3359427.3361908

Dataset: Horse Movement Data and Analysis of its Potential for Activity Recognition

Jacob Wilhelm Kamminga, Nirvana Meratnia, Paul J.M. Havinga

Pervasive Systems

Research output: Contribution to conference › Paper › peer-review

6 Citations (Scopus)

247 Downloads (Pure)

Abstract

We describe and analyze a dataset that comprises horse movement. Data was collected during horse riding sessions and when the horses freely roamed the pasture over 7 days. The dataset comprises 1.8 million 2-second data samples from 18 individual horses, of which 93303 samples from 11 subjects were labeled. Sensor devices were attached to a collar around the neck of the horses while the orientation was not fixed. The devices contained a 3-axis accelerometer, gyroscope, and magnetometer that were sampled at 100 Hz. To demonstrate how this dataset can be used, we evaluated a Naive Bayes classifier with leave-one-out validation. Our results show that a performance of 90% accuracy can be achieved using only the 3D acceleration vector as input. Furthermore, we demonstrate the effect of increased complexity, parameter tuning, and class balancing on classification performance and identify open research challenges.
The complete dataset is available online with open access at the 4TU.Centre for Research Data.

Original language	English
Pages	22-25
DOIs	https://doi.org/10.1145/3359427.3361908
Publication status	Published - 10 Nov 2019
Event	2nd Workshop on Data Acquisition To Analysis, DATA 2019 - Columbus University, New York, United States Duration: 10 Nov 2019 → 10 Nov 2019 Conference number: 2

Conference

Conference	2nd Workshop on Data Acquisition To Analysis, DATA 2019
Abbreviated title	DATA
Country/Territory	United States
City	New York
Period	10/11/19 → 10/11/19

Keywords

Animals, Horses, Activity Recognition, Accelerometer, Gyroscope, Compass, IMU, Orientation Independent, Neck

Access to Document

10.1145/3359427.3361908

Kamminga-2019-DatasetFinal published version, 2.33 MB

1 Participating in a conference, workshop, ...

2nd Workshop on Data Acquisition To Analysis, DATA 2019
Kamminga, J. W. (Participant)
10 Nov 2019
Activity: Participating in or organising an event › Participating in a conference, workshop, ...

Horsing Around -- A Dataset Comprising Horse Movement
Kamminga, J. W. (Creator), 4TU.Centre for Research Data, 14 Aug 2019
DOI: 10.4121/uuid:2e08745c-4178-4183-8551-f248c992cb14
Dataset

6 Citations
1 PhD Thesis - Research UT, graduation UT

Hiding in the Deep: Online Animal Activity Recognition using Motion Sensors and Machine Learning
Kamminga, J. W., 9 Sept 2020, Enschede: University of Twente. 225 p.
Research output: Thesis › PhD Thesis - Research UT, graduation UT

Open Access
File
832 Downloads (Pure)

Cite this

@conference{5bfd6a8af48149b18bc361a2c6b73f85,

title = "Dataset: Horse Movement Data and Analysis of its Potential for Activity Recognition",

abstract = "We describe and analyze a dataset that comprises horse movement. Data was collected during horse riding sessions and when the horses freely roamed the pasture over 7 days. The dataset comprises 1.8 million 2-second data samples from 18 individual horses, of which 93303 samples from 11 subjects were labeled. Sensor devices were attached to a collar around the neck of the horses while the orientation was not fixed. The devices contained a 3-axis accelerometer, gyroscope, and magnetometer that were sampled at 100 Hz. To demonstrate how this dataset can be used, we evaluated a Naive Bayes classifier with leave-one-out validation. Our results show that a performance of 90% accuracy can be achieved using only the 3D acceleration vector as input. Furthermore, we demonstrate the effect of increased complexity, parameter tuning, and class balancing on classification performance and identify open research challenges.The complete dataset is available online with open access at the 4TU.Centre for Research Data.",

keywords = "Animals, Horses, Activity Recognition, Accelerometer, Gyroscope, Compass, IMU, Orientation Independent, Neck",

author = "Kamminga, {Jacob Wilhelm} and Nirvana Meratnia and Havinga, {Paul J.M.}",

year = "2019",

month = nov,

day = "10",

doi = "10.1145/3359427.3361908",

language = "English",

pages = "22--25",

note = "2nd Workshop on Data Acquisition To Analysis, DATA 2019, DATA ; Conference date: 10-11-2019 Through 10-11-2019",

}

TY - CONF

T1 - Dataset: Horse Movement Data and Analysis of its Potential for Activity Recognition

AU - Kamminga, Jacob Wilhelm

AU - Meratnia, Nirvana

AU - Havinga, Paul J.M.

N1 - Conference code: 2

PY - 2019/11/10

Y1 - 2019/11/10

N2 - We describe and analyze a dataset that comprises horse movement. Data was collected during horse riding sessions and when the horses freely roamed the pasture over 7 days. The dataset comprises 1.8 million 2-second data samples from 18 individual horses, of which 93303 samples from 11 subjects were labeled. Sensor devices were attached to a collar around the neck of the horses while the orientation was not fixed. The devices contained a 3-axis accelerometer, gyroscope, and magnetometer that were sampled at 100 Hz. To demonstrate how this dataset can be used, we evaluated a Naive Bayes classifier with leave-one-out validation. Our results show that a performance of 90% accuracy can be achieved using only the 3D acceleration vector as input. Furthermore, we demonstrate the effect of increased complexity, parameter tuning, and class balancing on classification performance and identify open research challenges.The complete dataset is available online with open access at the 4TU.Centre for Research Data.

AB - We describe and analyze a dataset that comprises horse movement. Data was collected during horse riding sessions and when the horses freely roamed the pasture over 7 days. The dataset comprises 1.8 million 2-second data samples from 18 individual horses, of which 93303 samples from 11 subjects were labeled. Sensor devices were attached to a collar around the neck of the horses while the orientation was not fixed. The devices contained a 3-axis accelerometer, gyroscope, and magnetometer that were sampled at 100 Hz. To demonstrate how this dataset can be used, we evaluated a Naive Bayes classifier with leave-one-out validation. Our results show that a performance of 90% accuracy can be achieved using only the 3D acceleration vector as input. Furthermore, we demonstrate the effect of increased complexity, parameter tuning, and class balancing on classification performance and identify open research challenges.The complete dataset is available online with open access at the 4TU.Centre for Research Data.

KW - Animals, Horses, Activity Recognition, Accelerometer, Gyroscope, Compass, IMU, Orientation Independent, Neck

U2 - 10.1145/3359427.3361908

DO - 10.1145/3359427.3361908

M3 - Paper

SP - 22

EP - 25

T2 - 2nd Workshop on Data Acquisition To Analysis, DATA 2019

Y2 - 10 November 2019 through 10 November 2019

ER -

Dataset: Horse Movement Data and Analysis of its Potential for Activity Recognition

Abstract

Conference

Keywords

Access to Document

Fingerprint

Activities

2nd Workshop on Data Acquisition To Analysis, DATA 2019

Datasets

Horsing Around -- A Dataset Comprising Horse Movement

Research output

Hiding in the Deep: Online Animal Activity Recognition using Motion Sensors and Machine Learning

Cite this