Rapid evaluation of poultry manure content using artificial neural networks (ANNs) method

With increasing concern over the potential pollution from farm wastes, there is a need for rapid and robust methods that can analyse animal manure. In order to evaluate rapid testing methods based on the relationship between layer manure composition (ammonium nitrogen, total potassium, total nitrogen, total phosphorus, iron, copper, zinc, magnesium and sodium) and physicochemical properties (specific gravity, electrical conductivity, pH), diverse layer manure samples (n = 105) were used. Relationships were investigated using linear regression and artificial neural networks (ANNs). The performance of a neural network-based model was compared with a linear regression-based model using the same observed data. It was found that ANN model consistently gives better predictions. Based on the results of this study, ANNs appear to be a promising technique for predicting layer manure composition.