A predictive analytics of production planning model for yield performance in organic broiler industry using ensemble learning

Abstract

The Organic broiler production requires farmers to use organic herbs without growth promoters or other chemical contents. The problem is uncertain production yield performance results. On the other hand, farmers want good performance results to increase income and minimize loss of production. Therefore, machine learning can support organic broiler production decisions. Meanwhile, a single prediction model has a lower level of accuracy, so a combination of various prediction models is needed. The paper aims to find an accurate performance of the yield production prediction model and has a high degree of accuracy using ensemble learning techniques. The attributes that affect performance as input include corn, rice bran, soybean meal, coconut cake, fish meal, and herbs. Then we evaluate the model using the confusion matrix. The result shows that ensemble learning has succeeded in predicting the yield performance of organic broilers with an accuracy of 98.96%. The evaluation model showed that accuracy was 87.98% in the data train and 94.23% in the data test.