Authors: D. Balasubramanyam, M. Babu, M. Sakthivel, Karu. Pasupathi, T. V. Raja and S. N. Sivaselvam
In the present study the connectionist model also known as Artificial Neural Network (ANN) was used for prediction of body weight using linear body measurements in Madras Red sheep of Tamil Nadu, India. Data collected from 6561 sheep (2403 males and 4158 females) in the age group of 0-12 months were randomly divided into two sets namely training set comprising of 75 per cent data (1802 and 3119 records in males and females, respectively) to build the neural network model and test set comprising of remaining 25 per cent data points (601 and 1039 records in males and females, respectively) to test the model. Three different morphometric measurements, viz., chest girth, body length and height at withers were used as input variables and body weight was considered as output variable. A multilayer feed forward back propagation network was developed to predict the body weight. The prediction efficiency of the ANN models was compared with that of the traditional Multiple Regression Analysis (MRA) using the coefficient of determination (R2 value) and Root Mean Square error (RMSE). The correlation coefficients between the actual and predicted body weights in case of ANN model were found to be positive and highly significant and ranged from 86.81 to 90.20%. The ANN model was found to be better than the conventional MRA in terms of lower RMSE value (2.2461 in males and 2.4365 in females) and higher R2 value (81.33% in males and 77.60% in females) revealing the potential effectiveness of ANN models in predicting the body weight in Madras Red sheep.
Keywords: Artificial Neural Network, body measurements, body weight prediction, Madras Red sheep, multiple regression analysis, prediction efficiency