Download PDFOpen PDF in browser

Predicting Cow's Delivery Using Movement and Position Data Based on Machine Learning

7 pagesPublished: March 13, 2019

Abstract

One of the major problem farmers face is that of a parturition accident. A parturition accident result in the death of the calf when the cow gives birth. In addition, it reduces the milk yield. The farmer must keep the cow under close observation for the last few days of pregnancy.
A novel method to predict a cow’s delivery time automatically using time-series acceleration data and global position data by machine learning is proposed. The required data was collected by a small sensor device attached to the cow’s collar. An inductive logic programming (ILP) method was employed for a machine learning model as it can generate readable results in terms of a formula for first-order logic (FOL). To apply the machine learning technique, the collected data was converted to a logical form that includes predefined predicates of FOL. Using the obtained results, one can classify whether the cows are ready for delivery.
Data was collected from 31 cows at the NAMIKI Dairy Farm Co. Ltd. Using the method described above, 130 readings were obtained. The five-fold cross-validation process verified the accuracy of the model at 56.79%.

Keyphrases: acceleration data, global position data, inductive logic programming, machine learning, prediction of event

In: Gordon Lee and Ying Jin (editors). Proceedings of 34th International Conference on Computers and Their Applications, vol 58, pages 310-316.

BibTeX entry
@inproceedings{CATA2019:Predicting_Cows_Delivery_Using,
  author    = {Yusuke Ono and Ryo Hatano and Hayato Ohwada and Hiroyuki Nishiyama},
  title     = {Predicting Cow's Delivery Using Movement and Position Data Based on Machine Learning},
  booktitle = {Proceedings of 34th International Conference on Computers and Their Applications},
  editor    = {Gordon Lee and Ying Jin},
  series    = {EPiC Series in Computing},
  volume    = {58},
  publisher = {EasyChair},
  bibsource = {EasyChair, https://easychair.org},
  issn      = {2398-7340},
  url       = {/publications/paper/NBDT},
  doi       = {10.29007/bksq},
  pages     = {310-316},
  year      = {2019}}
Download PDFOpen PDF in browser