University of Guilan Animal Production Research 2252-0872 8 1 1970 01 01 Modeling output energy in the dairy and beef cattle farms using the methods of Artificial Neural Network and ANFIS (Case Study: Mazandaran Province, Iran) Modeling output energy in the dairy and beef cattle farms using the methods of Artificial Neural Network and ANFIS (Case Study: Mazandaran Province, Iran) 17 28 3418 10.22124/ar.2019.10377.1313 FA R. Loghmanpour Zarini Academic Staff, Department of Agricultural Machinery, Agricultural College of Sari, Technical and Vocational University, Mazandaran, Iran N. Ghavi Hossein-Zadeh Associate Professor, Department of Animal Science, Faculty of Agricultural Sciences, University of Guilan, Rasht, Iran H. Nabipour Afrouzi Academic Staff, Department of Animal Science, Agricultural College of Sari, Technical and Vocational University, Mazandaran, Iran Journal Article 1970 01 01 In this research, the artificial neural networks (ANNs) and multi-layered non-fuzzy inductive inference system (ANFIS) were used to model the output energy in dairy and beef cattle farms. For this purpose, according to Cochran's relation, 105 beef and dairy farms were randomly selected from five townships which were the main producers of this sector in Mazandaran province from 2016-2017. Using the energy balance of inputs and outputs, the input and output energy averages in beef production farms were calculated to be 16994.76 and 3449.85 MJcow^-1 and for dairy production farms were equal to 100100 and 58277 MJcow^-1, respectively. Also, ER (Energy Ratio), EP (Energy Productivity), SE (Special Energy) and NE (Net Energy) indices for dairy production farms were 0.58, 0.08 KgMJ^-1, 12.5 MJKg^-1 and -41825.93 MJcow^-1, respectively and for beef production farms were calculated as 0.2, 0.02 KgMJ^-1, 50 MJKg^-1and 13544.91 MJcow^-1, respectively. Using the results of statistical analysis of the data, modeling of the output energy for each unit of input energy was performed by two above methods (ANNs and ANFIS). The results showed that the model of nervous- fuzzy inference comparative multi-layered system with the correlation coefficient of 0.9899 for steer farms and 0.9933 for dairy farms, had better performance and accuracy than that of the artificial neural network with the correlation coefficient of 0.8118 and the structure of 6-16-1 for beef farms and correlation coefficient of 0.9837 and the structure of 5-12-1 for dairy farms. In this research, the artificial neural networks (ANNs) and multi-layered non-fuzzy inductive inference system (ANFIS) were used to model the output energy in dairy and beef cattle farms. For this purpose, according to Cochran's relation, 105 beef and dairy farms were randomly selected from five townships which were the main producers of this sector in Mazandaran province from 2016-2017. Using the energy balance of inputs and outputs, the input and output energy averages in beef production farms were calculated to be 16994.76 and 3449.85 MJcow^-1 and for dairy production farms were equal to 100100 and 58277 MJcow^-1, respectively. Also, ER (Energy Ratio), EP (Energy Productivity), SE (Special Energy) and NE (Net Energy) indices for dairy production farms were 0.58, 0.08 KgMJ^-1, 12.5 MJKg^-1 and -41825.93 MJcow^-1, respectively and for beef production farms were calculated as 0.2, 0.02 KgMJ^-1, 50 MJKg^-1and 13544.91 MJcow^-1, respectively. Using the results of statistical analysis of the data, modeling of the output energy for each unit of input energy was performed by two above methods (ANNs and ANFIS). The results showed that the model of nervous- fuzzy inference comparative multi-layered system with the correlation coefficient of 0.9899 for steer farms and 0.9933 for dairy farms, had better performance and accuracy than that of the artificial neural network with the correlation coefficient of 0.8118 and the structure of 6-16-1 for beef farms and correlation coefficient of 0.9837 and the structure of 5-12-1 for dairy farms. https://ar.guilan.ac.ir/article_3418_1beaac9afab1bcb84501f4aa2798ef38.pdf