Research On Intelligent Feeding Strategy Of Laying Hen Based On Data Mining Technology

Eggs are important agricultural products in the lives of residents.As people's food safety and nutritional needs continue to increase,the quality of eggs is constantly increasing.But the quality of eggs is closely related to the welfare level of laying hens.The most important factor is the intake of feed and water.However,the current monitoring of laying hen information still relies on experienced breeders for inspections,which is inefficient;Feeding of laying hens is still based on free feeding.Although the current automatic feeding system can feed regularly,the amount of feed and feeding time are still measured according to the experience of the breeder,and the intelligent feeding requirements are not met,so the utilization rate of feed is low,feed waste is serious,and the production performance of laying hens is difficult to achieve full play.Therefore,it is of great practical significance to monitor and collect the information in the process of laying hens and to study the intelligent feeding strategy of laying hens.In this paper,a research method of intelligent feeding strategy for laying hens based on data mining technology is proposed for laying hens.It mainly includes: a laying hen information collection system is designed to collect the feeding and drinking behavior data and production performance data,and the reliability analysis of the collected data;By analyzing the time series of feed weight and water weight,the time series is segmented and fitted by the method of perceptually important point,so as to quantitatively extract,description statistics and visual analysis of the feeding and drinking behavior of laying hen,and use video monitoring result for error analysis;Using association rule mining technology to discover the relationship between feeding and drinking behavior and production performance,and discover the implicit relationship between data;Finally,an LSTM-based behavior prediction model is designed to predict the feeding behavior in order to obtain specific feeding time and the amount of feed,and established a system platform for intelligent feeding decision-making in laying hen.The experimental results show that by verifying the collected data,it is found that the relative error between collected data and actual data is less than 2%;Through the video monitoring result,it can be found that the feeding time error obtained by the quantization algorithm is within half a minute,and feed intake has almost no large error;Through the association rule mining,it is found that there is a great correlation between feeding behavior and production performance.There is a positive correlation between feed intake and production performance.When the intake rate is between0.012-0.014g/s,the egg weight and feed utilization rate are at a high level;Through the LSTM-based behavior prediction model,the feeding behavior of laying hens is predicted,and the probability that the true value of the behavioral quantitative data is within the prediction range is more than 70%.