Answers are grounded in this paper and cite their sources.
This research developed a two-stage approach involving the use of machine learning models for predicting molting and ovarian maturation in Penaeus monodon broodstock. The unique and innovative aspect of this research lies in its application of machine learning techniques to improve the efficiency, reduce human error, and increase sustainability in shrimp aquaculture. By achieving accurate predictions of molting and ovarian maturation, this study offers significant potential for enhancing aquaculture management, ensuring optimal timing of spawning. The proposed approach involves eight machine learning models: decision tree, random forest, k-nearest neighbors, support vector machine, adaptive boosting, extreme gradient boosting, long short-term memory, and convolutional neural network. These models were evaluated through 10-fold cross-validation by using data collected from the International Center for the Scientific Development of Shrimp Aquaculture. In the two-stage approach, the first stage involves predicting the next molting event for P. monodon broodstock, and the second stage entails using the prediction results to further predict ovarian maturation. The evaluation results revealed that the random forest model delivered strong performance, achieving an accuracy rate of 88.5% and a receiver operating characteristic area under the curve of 94.3%. Early identification of ovarian maturation is essential to optimizing breeding and harvesting in shrimp aquaculture to increase efficiency and promote sustainability. The use of data-driven decision-making through machine learning not only holds value for the practice of shrimp aquaculture but also provides a strong foundation for future research and innovation. • Two-stage machine learning framework predicts both molting events and ovarian maturation in Penaeus monodon, enhancing reproductive management. • Random Forest model achieved top performance with 88.5% accuracy and 94.3% AUC in predicting ovarian maturation. • Combines physiological, environmental, and historical data using eight machine learning models for robust prediction and comparison. • Demonstrates potential for automated, data-driven decision-making in shrimp aquaculture to reduce labor and human error. • Provides preliminary results supporting the framework's applicability to broader aquaculture scenarios, including other species and diverse farming systems.
No takes yet. Share an insight, caveat, or question.
Yang et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: