The Pantanal has been facing consecutive years of extreme drought, with an impact on the quantity and quality of available pasture. However, little is known about how locally adapted breeds respond to the distribution of forage resources in this extreme drought scenario. This study aimed to evaluate the movement of free-grazing Pantaneiro sheep using a low-cost GPS to assess the main grazing sites, measure the daily distance traveled, and determine the energy requirements for walking with body weight monitoring. In a herd of 100 animals, 31 were selected for weighing, and six ewes were outfitted with GPS collars. GPS data collected on these animals every 10 m from August 2020 to May 2021 was analyzed using the Python programming language. The traveled distance and activity energy requirements (ACT) for horizontal walking (Mcal/d of NE_m) were determined. The 31 ewes were weighed at the beginning and end of each season. The available dry matter (DM) and floristic composition of the grazing sites were estimated at the peak of the drought. DM was predicted using power regression with NDVI (normalized difference vegetation index) (R² = 0.94). DM estimates averaged 450 kg/ha, ranging from traces to 3830 kg/ha, indicating overall very low values. Individual variation in the frequency of use of grazing sites was observed (p < 0.05), reflecting the distances traveled and the energetic cost of the activity. The range of distances traveled by the animals varied from 3.3 to 17.7 km/d, with an average of 5.9 km/d, indicating low energy for walking. However, the traveled distance and ACT remained consistent over time; there were no significant differences observed between seasons (p > 0.05). On average, the ewes’ initial weight did not differ from the weight at the drought peak (p > 0.05), indicating that they maintained their initial weight, which is important for locally adapted breeds as it confers robustness and resilience. This study also highlighted the importance of the breed’s biodiverse diet during extreme drought, which enabled the selection of forage for energy and nutrient supplementation. The results demonstrated that precision tools such as GPS and satellite imagery enabled the study of animals in extensive systems, thereby contributing to decision-making within the production system.

One of the main challenges in machine learning tasks is unlabeled data. Consequently, semi-supervised learning is an approach commonly used in labeling data tasks. The use of diffusion kernels is one of the powerful methods in this set of learning. This paper introduces Phylogram based Data Labeling (PhyDL), an innovative method that combines the benefits of phylogram analysis with Learning with Local and Global Consistency (LLGC). PhyDL leverages phylogram analysis to act as an aggregator of information from labeled samples, capable of correcting possible labeling errors. PhyDL constructs a phylogram using community detection algorithms, FastGreedy and Edge Betweenness, to identify and leverage the structure of the data, with FastGreedy emerging as the best-performing algorithm among those evaluated. We implemented PhyDL and evaluated its performance on the TwoMoons, MNIST, MUSK, HEART, and SONAR datasets. PhyDL outperformed LLGC on 4 out of 5 evaluated datasets, highlighting its effectiveness in semi-supervised learning scenarios. On the TwoMoons, MNIST, MUSK, and SONAR benchmarks, the method showed steady accuracy gains, with consistent improvements as the number of labeled samples increased. However, on the HEART dataset, standard LLGC achieved better results, indicating that PhyDL’s community-based diffusion can be limited by certain underlying spatial structures. Furthermore, the experiments revealed that the Edge Betweenness algorithm is unsuitable for these specific high-dimensional tasks. Ultimately, this study showcases PhyDL’s potential, when equipped with the FastGreedy algorithm, to significantly improve data labeling performance in semi-supervised scenarios, effectively leveraging the structural topology of complex datasets.