Development of an Expert System for Diagnosing Respiratory Diseases in Animals Using the Bayesian Network and Rule-Based System Methods

Respiratory diseases in animals are a common health issue affecting both livestock and pets. These conditions can lead to significant economic losses in the livestock sector and reduce the quality of life for pets if left untreated. Early and accurate diagnosis is crucial to identify diseases promptly, prevent further spread, and minimize negative impacts on animals and their owners. Therefore, a system capable of providing fast, accurate, and data-driven diagnoses is essential. This study aims to develop an expert system specifically designed to diagnose respiratory diseases in animals using two main approaches: Bayesian Network and Rule-Based System. The Bayesian Network models uncertainties by analyzing probabilistic relationships between observed symptoms and potential diseases, while the Rule-Based System supports decision-making based on predefined rules. The combination of these methods is expected to yield more accurate and informative diagnostic results. Symptom data for this study were obtained from various sources, including relevant medical literature and animal health databases. The system was developed using Python programming language, leveraging libraries such as pgmpy for constructing Bayesian Network models and experta for implementing the Rule-Based System. The development and testing processes were conducted on the Google Colab platform, enabling efficient data processing, simulation, and visualization. The expert system was evaluated using simulated symptom data, with performance parameters including diagnosis probability and overall accuracy. The results indicate that the expert system effectively provides diagnoses based on user-input symptoms. The probability information included in the diagnostic results aids veterinarians and livestock owners in making more precise, data-driven medical decisions