Objective:To predict the feeding amount of pregnant sows,in order to accurately control the amount of feed required for pregnant sows,and to help feed sows accurately and save breeding costs.Methods:By combining the advantages of ARIMA and LSTM algorithms,and utilizing the ARIMA-LSTM optimization algorithm that integrated ARIMA and LSTM,the feeding amount of pregnant sows was accurately predicted to control the precise feeding of intelligent feeders.Results:The ARIMA-LSTM optimization algorithm has been experimentally verified to have the highest prediction accuracy for sow feeding volume.Compared with ARIMA and LSTM algorithms,the root mean square error has been reduced by 48.74％and 17.22％,respectively,and the average absolute deviation has been reduced by 48.70％and 27.37％,respectively.Conclusion:The ARIMA-LSTM optimization algorithm used in this article improved the prediction accuracy of feeding amount during pregnant sows,and could control the precise feeding of intelligent feeders,providing a good basis for predicting feeding amount in pregnant sows.
维普
万方数据