Research of Estrus Detection Models in Dairy Cows by Activity
Science Discovery
Volume 6, Issue 2, April 2018, Pages: 102-109
Received: Jun. 20, 2018; Published: Jun. 22, 2018
Views 545      Downloads 30
Authors
Daoerji Fan, College of Electronic Information Engineering, Inner Mongolia University, Hohhot, China
Huijuan Wu, College of Electronic Information Engineering, Inner Mongolia University, Hohhot, China
Article Tools
Follow on us
Abstract
The detection of estrus in large-scale ranch is an extremely labor-intensive task. Accurate and efficient detection can shorten the tires spacing of cow and increase economic income of ranch.The change in activity is one of the most important external features of dairy cows' estrus.In this paper, the activity of cows was collected remotely using activity collectors, wireless transmission networks and cloud servers, and the data was framed, labeled and preprocessed. The logistic regression, MLPs and SVMs model were trained by differential activity data of current and historical dates within the same hour.The experimental results show that the detection model by cow activity has a higher accuracy in predicting the estrus of cows. The SVMs in the three models have the best prediction effect. The recall and accuracy rate reach 90.12% and 93.74%, respectively.In the actual ranch test, the estrus detection system using the SVM model has more than double the exposure rate than the manual disclosure.
Keywords
Cows Estrus Prediction, Activity, Logistic Regression, Multilayer Perceptions, Support Vector Machine
To cite this article
Daoerji Fan, Huijuan Wu, Research of Estrus Detection Models in Dairy Cows by Activity, Science Discovery. Vol. 6, No. 2, 2018, pp. 102-109. doi: 10.11648/j.sd.20180602.15
ADDRESS
Science Publishing Group
1 Rockefeller Plaza,
10th and 11th Floors,
New York, NY 10020
U.S.A.
Tel: (001)347-983-5186