Table 5. Detailed comparison of various models and algorithms used to monitor and assess pig behavior and farm conditions, including specific applications like tail biting detection, piglet crushing prevention, and environmental condition tracking

Detection/Classification	Classification technique	Feature Extraction Technique	Accuracy	Reference
Behavior/activity	Power spectrum density	Average peak frequency, fundamental frequency, duration	81.12	[25]
	Rule-based classifier	FFT, CGD, FFC	87.00	[38]
	Polynomial adjustment	FT	-	[46]
	Decision tree	Acoustic response	93.0	[78]
Tail biting/piglet crushing	GLMM+LME	Pitch frequency	52.9	[82]
	Vocalization pattern	Pitch frequency, maximum amplitude, intensity	78.20	[43]
	Mean maximum frequency	Acoustic properties	25-75	[44]
	Vocalizations pattern	Intensity	-	[77]
Environmental factors	PCA + SVM	PMFCCs	95.00	[80]
	ANOVA	Noise level	-	[75]

FFT, fast Fourier transform; CGD, chirp group delay; FFC, fundamental frequency calculation; FT, Fourier transform; GLMM, generalized, linear mixed-effects model; LME, linear mixed-effects; PCA, principal component analysis; SVM, support vector machine; PMFCC, principal mel-frequency cepstrum coefficient.