Machine Learning for Increasing Productivity in Shrimp Hatchery

Poor quality in the shrimp hatchery stage increases susceptibility to infections during the shrimp grow-out stage, leading to lower survival rates and daily gains. This study focuses on leveraging machine learning to enhance productivity in shrimp hatcheries, particularly targeting the critical hatchery stage of shrimp cultivation in Thailand. The aim of this research is to develop and compare four classifier models for shrimp hatchery. Data was collected from various sources, preprocessed, and analyzed using the four models including logistic regression, random forest, support vector machine, and neural networks. The study emphasizes the importance of feature selection in optimizing machine learning models for shrimp hatchery, improving predictive capabilities and decision-making for enhanced productivity in shrimp farming operations. The results revealed that the Random Forest model achieved the best performance, with perfect scores across all evaluation metrics, including an accuracy of 100%. In contrast, logistic regression, while not achieving perfect scores, showed the greatest improvement after applying feature selection, with an accuracy rate of 95% and enhancements in all result metrics, including recall, precision, specificity, and F1-score.