Comparison of Gompertz and artificial neutral network models of broiler growth received Artichoke extract in their drinking water

An experiment was carried out to estimate growth parameters of broiler chickens received Artichoke extract in their drinking water using Gompertz non-linear model and to compare Gompertz non-linear regression equation and artificial neural network modeling in terms of their ability to predict body weight of broiler chicks at day 42 of age. A total number of 250 one- day- old Ross 308 broiler chicks were used. The chicks received Artichoke extract in their drinking water at the doses of 0، 100، 200، 300 and 500 mg per liter. The birds were weighed at days 1، 2، 7، 14، 21، 24، 27، 30، 33، 35 and 42 of experimental period after 3 hours of fasting. Estimated mature body weight was significantly higher in birds received 200 mg/liter Artichoke extract in drinking water than the other birds (P<0. 05). The chicks received 200 mg/liter Artichoke extract in their drinking water had a significantly higher coefficient of relative growth compared to the control group and those received 300 mg/liter Artichoke extract (P<0. 05). The goodness of fit in terms of R2 values of the artificial neural network (ANN) model showed a higher accuracy of prediction for body weight of broiler chicks at day 42 of age than the equation established by Gompertz model (0. 998 vs. 0. 997، respectively). Because mean square error (MSE)، mean absolute deviation (MAD)، mean absolute percentage error (MAPE) was lower in the ANN than in Gompertz model، it estimated body weight of birds at day 42 of age better than did Gompertz model.