An Innovative Concept for a Multivariate Plausibility Assessment of Simultaneously Recorded Data

Area of Science:

• Veterinary Medicine
• Animal Science
• Data Science

Background:

• Simultaneously recorded physiological data in livestock often exhibit interdependencies.
• Identifying physiologically implausible or extreme data is crucial for accurate health monitoring and management.

Purpose of the Study:

• To develop and demonstrate an innovative multivariate plausibility assessment (MPA) algorithm.
• To differentiate between normal, physiologically extreme, and implausible observations in complex datasets.

Main Methods:

• The MPA algorithm leverages the physiological linkage between multiple measurable parameters.
• Applied to time-series data from 100 cows across 10 dairy farms, including climate, pH, temperature, behavior, milk yield, and blood parameters.

Main Results:

• The MPA algorithm successfully identified implausible observations across various parameters.
• Intra-reticular pH data showed the highest proportion of implausible values (approx. 5%), with other traits showing up to 2.5% implausibility.
• The MPA demonstrated effectiveness in improving data quality and detecting extreme physiological conditions.

Conclusions:

• The developed MPA algorithm is a foundational concept for enhancing data quality in animal health monitoring.
• It offers a method to detect implausible data and understand extreme physiological states within complex, multi-parameter datasets.
• Further development and validation are recommended for its application as a management tool.