Prediction of Feed Intake of Pregnant Sows Based on GRU

Abstract

With the development of artificial intelligence technology, deep learning is widely used in industry, agriculture and other fields. In large-scale farms, the production level of pregnant sows directly affects the economic benefits of farms. In order to improve the feeding level of pregnant sows, this paper takes the backfat thickness, parity, pregnancy date and feeding information of pregnant sows as the data set, and builds a feed intake prediction model based on Gated Recurrent Unit (GRU) network to predict the feed consumption of pregnant sows. Through precise feeding, it can not only reduce the feeding cost of the farm, but also maintain the best posture of sows, and improve the litter size and healthy litter size of sows. At the same time, establish the Long Short Term Memory (LSTM) network, the recurrent neural network (RNN) and the deep neural network (DNN) to compare with GRU. The experimental results show that the GRU model has faster training speed and higher prediction accuracy than other models, which is in line with the feeding law of pregnant sows, and has higher application value, which is conducive to the accurate feeding of pregnant sows.