Evaluating ROCKET and Catch22 features for calf behaviour classification from accelerometer data using Machine Learning models - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Computers and Electronics in Agriculture Année : 2024

Evaluating ROCKET and Catch22 features for calf behaviour classification from accelerometer data using Machine Learning models

Résumé

Monitoring calf behaviour continuously would be beneficial to identify routine practices (e.g., weaning, transport, dehorning, etc.) that impact calf welfare in dairy farms. In that regard, accelerometer data collected from neck collars can be used along with Machine Learning models to classify calf behaviour automatically. However, further development is needed to classify a broad spectrum of behaviours with good genericity from one animal to another. While Hand-Crafted features are typically used in the field as inputs for Machine Learning models, feature sets designed explicitly for time-series classification problems have been developed in related fields, such as ROCKET and Catch22 features. This study aims to compare the performance of ROCKET and Catch22 features to Hand-Crafted features commonly used in the field. 30 Irish Holstein Friesian and Jersey pre-weaned calves were equipped with an accelerometer sensor for several weeks, and their behaviours were annotated, allowing for 27.4 hours of observation aligned with the accelerometer time-series. Additional time-series were computed from the raw X, Y and Z-axis and split into 3-second time windows. ROCKET, Catch22 and Hand-Crafted features were calculated for each time window, and the dataset was then split into the train, validation and test sets. Each set of features was used to train three Machine Learning models (Random Forest, eXtreme Gradient Boosting, and RidgeClassifierCV) to classify six behaviours indicative of pre-weaned calf welfare (drinking milk, grooming, lying, running, walking and other). Models were tuned with the validation set, and the performance of each feature-model combination was evaluated with the test set. The best performance across the three models was obtained with ROCKET [average balanced accuracy ± standard deviation] (0.70 ± 0.07), followed by Catch22 (0.69 ± 0.05), well ahead of Hand-Crafted (0.65 ± 0.034). The best balanced accuracy (0.77) was obtained with ROCKET and RidgeClassifierCV, followed by Catch22 and Random Forest (0.73). Thus, tailoring these approaches for specific behaviours and contexts will be crucial in advancing precision livestock farming and enhancing animal welfare on a larger scale.
Fichier principal
Vignette du fichier
COMPAG.pdf (18.1 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04622072 , version 1 (24-06-2024)

Identifiants

  • HAL Id : hal-04622072 , version 1

Citer

Oshana Dissanayake, Sarah E. Mcpherson, Joseph Allyndrée, Emer Kennedy, Padraig Cunningham, et al.. Evaluating ROCKET and Catch22 features for calf behaviour classification from accelerometer data using Machine Learning models. Computers and Electronics in Agriculture, In press. ⟨hal-04622072⟩
0 Consultations
0 Téléchargements

Partager

Gmail Mastodon Facebook X LinkedIn More