Disease detection in pigs based on feeding behaviour traits using machine learning

Kavlak, A.T.; Pastell, Matti; Uimari, Pekka

doi:10.1016/j.biosystemseng.2023.01.004

Cited by 8 publications

(6 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A detailed description of the data has been given in the study by Kavlak and Uimari (2019) , Kavlak et al (2021) , and Kavlak et al (2023) . Briefly, the data consisted of feeding behavior traits of 7,347 pigs (Finnish Yorkshire: 3,059, Finnish Landrace: 3,320, and their F1-crosses: 968) raised at the central test station Figen Oy (Längelmäki, Finland) between 2011 and 2016.…”

Section: Methodsmentioning

confidence: 99%

“…The recorded information included pig ID, observed symptom(s), and date. The identified symptoms were categorized as a cough, limp, loss of appetite (for pigs consuming <600 g/d), skin damage, and a bitten tail ( Kavlak et al, 2023 ). The average number of sick days for a single pig was calculated by dividing the total number of days the specific pig was observed with symptoms by the total duration of the test period.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Inheritance of feed intake-based resilience traits and their correlation with production traits in Finnish pig breeds

Kavlak,

Uimari

2024

Journal of Animal Science

Self Cite

View full text Add to dashboard Cite

High resilience against diseases, changing environmental conditions, and other stress factors and the ability to efficiently recover to normal status, is becoming increasingly important in pig production. Finding new phenotypes that relate to resilience is a crucial step for improving the resilience of pigs through selection. The objective of this study was to extract resilience-related phenotypes based on fluctuations in daily feed intake (DFI, g) and time spent in feeding per day (TPD, min) and to estimate the heritability of these traits and genetic correlations with production traits (PT). Resilience-related traits with high enough heritability and with either favorable or neutral genetic correlation with PT could be used in the selection program to improve the productivity and welfare of pigs. In this study, we used data from 7,347 Finnish Yorkshire, Landrace and crossbred pigs raised at the test station. Six pig-specific resilience-related phenotypes were extracted from the individual DFI and TPD: root mean square error (RMSE), quantile regression (QR), and coefficient of variation (CV). Root mean square error was calculated from the differences between the actual DFI (or TPD) and the pig-specific predicted values. Quantile regression was based on the number of days that a pig belonged to the group with the lowest 5% of pigs based on DFI (or TPD), and CV was calculated over the daily observations of DFI (or TPD). Production traits included average daily gain (ADG, g), backfat thickness (BF, mm), and feed conversion rate (FCR, g/g). The heritability estimates for resilience-related traits varied between 0.07 ± 0.02 (QRDFI) and 0.20 ± 0.03 (RMSETPD). The genetic correlations between resilience-related traits and PT were mostly neutral, but for example, RMSEDFI had a favorable genetic correlation with FCR and BF but an unfavorable correlation with ADG. Lastly, we observed that pigs belonging to the lowest 10% group based on their breeding value for QRTPD had a lower proportion (10% incidence) of sick days compared to the highest 10% breeding value group (30% incidence). Therefore, pigs exhibiting small TPD variation (related to high resilience) tend to be less susceptible to sickness than pigs with large TPD variation (related to low resilience). Given its moderate heritability, neutral genetic correlation with PT, and positive effect on health, QRTPD can be considered the most promising resilience-related trait in the Finnish production system.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Inheritance of feed intake-based resilience traits and their correlation with production traits in Finnish pig breeds

Kavlak,

Uimari

2024

Journal of Animal Science

Self Cite

View full text Add to dashboard Cite

show abstract

“…They are based on the detection of physicochemical and biological changes in milk or udders or on the assessment of milk or blood biomarkers associated with mastitis (Džermeikaitė et al, 2023). To detect lameness, animal limb movement is assessed, along with imaging, using the Fog computing, Classification, and regressive tree (CART) XGBoost algorithm (Coşkun et al, 2023;Kavlak et al, 2023). Swine flu, on the other hand, is assessed using the Optical flow algorithm by measuring animals' mobility and direction (Neethirajan, 2020).…”

Section: Precision Livestock Farming and Crop Productionmentioning

confidence: 99%

Agriculture in the face of new digitization technologies

Wrzecińska,

Czerniawska-Piątkowska,

Kowalewska

et al. 2023

UBSRAS

View full text Add to dashboard Cite

Agriculture plays a vital role in food production, resource utilization, and employment but faces challenges from population growth, climate change, and food shortages. The development of information technology has significantly contributed to the industry's development, and modern technologies such as artificial intelligence, the Internet of Things, computer vision, and machine learning have revolutionized agricultural practices. The purpose of this review is to explore the adoption of digital technologies in agriculture, with a specific focus on their application in livestock breeding. Through the examination of current literature and the utilization of various research methods, this review contributes to the existing knowledge in the field. It is established that the latest information tools allow collecting, analysing data, automating tasks and supporting decision-making, which leads to increased agricultural efficiency, resource management and sustainable development. It has been proven that modern technologies play a crucial role in increasing agricultural production, improving the efficiency of livestock and crop production. These technologies include devices and sensors, data analytics and decision support systems, as well as systems for overall farm productivity assessment. Precision technologies in agriculture, thanks to automation, sensors and machine learning, allow farmers to monitor animal health, optimise feed consumption, detect diseases at early stages and increase overall productivity. IT solutions in agriculture facilitate data processing, visualisation and decision-making, leading to lower costs, greater efficiency and improved food security. The study provides practical insights for farmers and other agricultural stakeholders who can benefit from accurate information, real-time monitoring and automated processes through the integration of modern technologies, ultimately improving agricultural practices and sustainability

show abstract

Section: Precision Livestock Farming and Crop Productionmentioning

confidence: 99%