Exploring the specific features of 24-h behavioural pattern of dairy cows

L. Shi*, A. Kok, B. Kemp, A.T.M. van Knegsel

*Corresponding author for this work

Research output: Contribution to conferenceAbstract

Abstract

Circadian rhythm, physiological status and farm management shape the 24-h behavioural pattern of dairy cows. In turn, the changes of 24-h behavioural patterns can reflect the physiological status of the cow and conditions on the farm, such as milking time, feeding regime, the occurrence of diseases and heat stress. Furthermore, 24-h behavioural patterns are more informative than the daily behavioural patterns which have been the focus of most previous research. Because the 24-h patterns can capture those behavioural features which only exist in higher data granularity and eclipse in the process of data amalgamation. These Features can be used to summarize and quantify 24-h behavioural patterns, and deviations in feature values can further be used to detect of the occurrence of disturbances in the animals. We hypothesize that features subtracted from 24-h behavioural patterns can better classify and quantify the behavioural pattern of dairy cows.
In this project, multiple years of behavioural sensor data, including lying and standing time, steps and transition changes per 15-min interval per cow were collected from 8 farms with different types of management. In the current step of the study, we are focusing on quantifying and comparing 24-h patterns of the herd of different periods within and across farms. Through comparing with 4 previous and 4 later intervals, the number and height of step counts peaks per day were determined and indicated that the activity patterns were similar between pasture period and non-pasture period of farms with same type of management. The hierarchical clustering showed that filtered features had better performance on classifying pasture access, farm and month than daily value. Meanwhile, the monthly behavioural patterns of 8 farms were more similar within farm rather than month. These filtered features will give new insight into the behavioural patterns of dairy cows and will be associated with heat stress and diseases in future study.
Original languageEnglish
Publication statusPublished - 24 Oct 2023
EventADP Science Day 2023 - Landgoed Welderen, Elst, Netherlands
Duration: 24 Oct 202324 Oct 2023

Other

OtherADP Science Day 2023
Country/TerritoryNetherlands
Period24/10/2324/10/23

Fingerprint

Dive into the research topics of 'Exploring the specific features of 24-h behavioural pattern of dairy cows'. Together they form a unique fingerprint.

Cite this