Can we detect patterns in behavioral time series of cows using cluster analysis?

Area of Science:

• Animal behavior and welfare science
• Time series analysis in ethology
• Dairy cattle management

Background:

• Assessing animal welfare through behavioral patterns is crucial.
• Dairy cows' daily activity and area use patterns require investigation.
• Understanding individual behavioral consistency is key to welfare assessment.

Purpose of the Study:

• To determine if dairy cows exhibit daily, individual patterns in activity and barn area usage.
• To test if behavioral patterns are more consistent within than between cows.
• To evaluate if specific area categorization and time of day influence pattern detection.

Main Methods:

• Utilized IceTag pedometers and SMARTBOW sensor systems for data collection over 55 days.
• Analyzed activity and area use data at 1-minute intervals using hierarchical clustering.
• Compared daily behavioral time series within and between 20 lactating cows.

Main Results:

• Hierarchical clustering did not group individual cows' days more closely than different cows'.
• More specific area categorization yielded slightly better grouping but not during specific times.
• Average distances between days were consistently smaller within cows than between cows.

Conclusions:

• Dairy cows display weak individual preferences in area use and daily activity patterns.
• Current cluster analysis approaches may be unsuitable for detecting patterns in behaviorally plastic animals.
• Further exploration of time series methods accounting for temporal fluctuations is recommended for accurate behavioral assessment.