Sistema de alerta temprana de labor de parto en el hato porcino empleando técnicas de redes neuronales convolucionales y LSTM

The objective of this titration work was to develop an early warning system to detect farrowing patterns in sows and issue an early alert to the caretakers of the pig herd indicating that the sow is close to labor using artificial intelligence techniques. such as convolutional neural networks and LSTM (Long Short Term Memory) models. For the development of this work, two CRISP-DM methodologies were used, focused on planning and monitoring, which has phases such as: Understanding the problem, understanding data, data preparation, modeling, model implementation; The SABOR IBM methodology was also used, focused purely on the code development of the project in question. Finally, once the model was trained, a telegram bot was created to issue the prediction alert. After performing the respective hot test, it was confirmed that the system issues effective alerts to caregivers when the probability of birth exceeds 70%. This ensures timely intervention by the team in charge.

• Main Author: Álava Zambrano, Karen Barbarita (author)
• Other Authors: Basurto Vidal, Willians Eduardo (author)
• Format: bachelorThesis
• Language: spa
• Published: 2024
• Subjects: Redes neuronales; CRISP-DM; Parto porcino; Patrón de comportamiento; Predicción de parto
• Online Access: http://repositorio.espam.edu.ec/handle/42000/2424