Detection of broiler heat stress by using the generalised sequential pattern algorithm

Abstract

The sequence pattern mining method aims to identify frequent sequences that exceed a user-specified support threshold. The present study uses the same approach based on sequential standards to estimate the heat stress of broilers from a resulting behavioural pattern. Experimental data were recorded in a climate chamber where the behaviour of broilers was recorded under thermoneutral (comfort) conditions, set as standard, and when exposed to thermal stress (cold and heat). The Generalised Sequential Patterns (GSP) algorithm was used to evaluate the heat stress of broilers in the third and fourth week of growth. The results indicated that the mining of pattern sequences is a useful and straightforward technique to estimate the welfare of broiler chickens, allowing the identification of temporal relations between thermal stress and the consequent behaviour of the broiler. Temperature 8 °C below the standard thermoneutral conditions showed that the broiler remained lying down most of the time, walking only to the drinker and feeder trough. Broilers exposed to temperatures 8 °C above the standard thermoneutral conditions () tend to decrease locomotor activities, showing lower welfare status.