Other language title :
Comparison of Artificial Neural Network and Multiple Regression Analysis for Prediction of Fat Tail Weight of Sheep
Title of article :
مقايسه كارآيي شبكه عصبي مصنوعي و رگرسيون چندگانه در پيشبيني وزن دنبه گوسفند
Author/Authors :
Norouzian، M.A. نويسنده College of Abouraihan,Department of Animal Science,University of Tehran,Tehran,Iran نوروزيان, م.ع. , Vakili Alavijeh، M. نويسنده Faculty of Mathematical Science,Department of Mathematics,Shahid Beheshti University,Tehran,Iran وكيلي علويجه, م.
Issue Information :
فصلنامه با شماره پیاپی سال 2016
Abstract :
در اين مطالعه ارتباط بين وزنهاي تولد، از شيرگيري و پايان پروار با وزن دنبه 69 رأس گوسفند بلوچي توسط روشهاي شبكه عصبي مصنوعي و رگرسيون چندگانه بررسي شد. هر دو روش با دقت بالايي وزن دنبه را پيشبيني كردند. هر چند كه ميانگين خطا به صورت معنيداري در روش شبكه عصبي مصنوعي كمتر از رگرسيون چندگانه بود. ضريب تعيين برآورد شده در روش شبكه عصبي مصنوعي (93/0) بالاتر از رگرسيون چندگانه (81/0) به دست آمد. استفاده از شبكه عصبي مصنوعي ميانگين خطاي استاندارد را 59 و ضريب تعيين را 15 درصد بهبود داد. به نظر ميرسد كه بتوان با استفاده از شبكه عصبي مصنوعي وزن دنبه را از صفات وزن بدن پيشبيني كرد.
Abstract :
A comparative study of artificial neural network (ANN) and multiple regression is made to predict the fat tail weight of Balouchi sheep from birth, weaning and finishing weights. A multilayer feed forward network with back propagation of error learning mechanism was used to predict the sheep body weight. The data (69 records) were randomly divided into two subsets. The first subset is the training set comprising of 75 percent data (52 records) to build the neural network model and test data set comprising of 25 percent (17 records), which is not used during the training and is used to evaluate performance of different models. The mean relative error was significantly (P<0.01) lower for ANN than the MLR model. The coefficient of determination (R^2) values computed for the body measurements were generally higher (0.93) using ANN model than the multiple linear regression (MLR) model (0.81). The ANN model improved the mean squared error (MSE) of the MLR model by 59% and R^2 by 15% that the ANN represents a valuable tool for predicting of lamb fat tail weight from birth, weaning and finishing weights.
Keywords :
fat tail , Sheep , multiple linear regression , Artificial neural network
Journal title :
Iranian Journal of Applied Animal Science
Journal title :
Iranian Journal of Applied Animal Science