# A Supervised Deep Learning Model Was Developed to Classify Nelore Cattle (Bos indicus) with Heat Stress in the Brazilian Amazon

## TL;DR

A deep learning model was developed to classify Nelore cattle in Brazil as either comfortable or under heat stress using biotic and abiotic variables.

## Contribution

A supervised deep learning model was developed for non-invasive heat stress classification in cattle using real-time data.

## Key findings

- The model achieved 72% accuracy and 72% recall but only 42% specificity.
- Rectal temperature was identified as a reliable predictor variable.
- The model's performance suggests potential for precision livestock farming with improvements in data balance and features.

## Abstract

Intelligent technologies that do not require the handling of animals have been used to monitor agricultural production systems in real time using captured images, enabling efficient decision-making and, consequently, minimizing animal stress. In this way, we used a deep learning model to classify Nelore cattle (Bos indicus) into two groups (comfortable and above thermal comfort) based on data collection at four different times, from June to December 2023, considering abiotic and biotic variables. After analysis, it was possible to see that the model used showed excellent accuracy, high precision and recall but low specificity in animals above comfort. Canonical analysis indicated that rectal temperature (RT) was a reliable predictor variable. Thus, it is concluded that the model has potential, but it is recommended that more variables be adopted to improve its classification capacity.

Non-invasive and intelligent technologies have been utilized to monitor agricultural systems in real time, facilitating expedient decision-making and the reduction in animal stress in diverse climatic conditions. The objective of this study was to develop a deep learning supervised model to classify Nelore cattle (Bos indicus) into two groups: those in comfort and those under thermal stress. Thirty cattle, aged between 18 and 20 months, were evaluated between June and December 2023, resulting in 676 samples collected across four daily periods (6:00, 12:00, 18:00, and 24:00). Biotic variables included rectal temperature (RT) and respiratory rate (RR), while abiotic variables included air temperature (AT) and relative humidity (RH). The neural network model exhibited an accuracy and recall of 72% but a low specificity of 42%. These metrics indicate that while the model is effective in detecting stress situations, it faces challenges in correctly identifying animals in thermal comfort, likely due to class imbalance and the need for additional input features to capture environmental adaptability. Consequently, it can be posited that supervised learning models are valuable tools for precision livestock farming, provided that discriminatory limitations are mitigated by refining input characteristics and data balancing.