Mass Mortality Risk Prediction and Fish Loss Simulation in Mariculture

This study aimed to reduce fish loss in marine aquaculture (mariculture) in Indonesia. In Indonesia, mariculture is an important industry that is expected to grow in the future. However, production is unstable and mass mortality of fish often occurs because appropriate technology has yet to be established. To solve these problems, we examined a machine learning model to predict the occurrence of mass mortality. Results showed the average accuracy of the proposed model was 0.699. Subsequently, we simulated fish loss based on the prediction model and found that the number of fish lost was up to 59.7% lower than the actual loss. Thus, the models were effective in predicting the occurrence of mass mortality and reducing fish loss.