A Neural Network Ensemble Models for Pig Size Prediction

Abstract

The paper presents a neural network ensemble models for pig size prediction in swine fattening farm. This approach has been developed based on three methods: Multilayer perceptron with back-propagation neural network, Radial basis function and Generalize regression neural network. The solutions from three methods, then are combined using the Simple averaging, Weighted averaging and Median approaches. The results show that the developed model provide reasonable prediction solutions with the root-mean square error (RMSE) lower than 9.0 for all sizes of pigs.