CATTLE DISEASE DIAGNOSIS SYSTEM USING RANDOM FOREST CLASSIFICATION METHOD

Abstract

Cattle farming plays an important role in Indonesia's economy, but its productivity can decline due to livestock health issues. To address this, this study develops a cattle disease diagnosis system based on machine learning using the Random Forest classification method. The system helps farmers identify diseases independently based on input symptoms. The model is built using the Random Forest algorithm, trained on 1745 primary data obtained from the Barru Regency Department of Agriculture. The data undergoes a comprehensive pre-processing stage, including cleaning to remove inconsistencies, One-Hot Encoding for categorical feature transformation, and class balancing using the Synthetic Minority Over-sampling Technique (SMOTE) to ensure fair representation of all disease categories. Model evaluation using a Confusion Matrix demonstrates a high accuracy of 91%, indicating strong predictive performance. Based on the model, a mobile application based on Android is developed to assist farmers in the early detection of cattle diseases.