Modeling the Test Day Milk Yields via Time Series Method

Creative Commons License

Karaman E., FIRAT M. Z.

KAFKAS UNIVERSITESI VETERINER FAKULTESI DERGISI, cilt.19, sa.4, ss.659-664, 2013 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 19 Sayı: 4
  • Basım Tarihi: 2013
  • Doi Numarası: 10.9775/kvfd.2013.8609
  • Dergi Adı: KAFKAS UNIVERSITESI VETERINER FAKULTESI DERGISI
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, TR DİZİN (ULAKBİM)
  • Sayfa Sayıları: ss.659-664
  • Anahtar Kelimeler: Test day, Milk yield, Time series, ARIMA, Forecast, LACTATION CURVES, HOLSTEIN COWS
  • Akdeniz Üniversitesi Adresli: Evet

Özet

The aim of this study is to model the test day milk yields via time series methodology and to determine the number of test days which provide the most accurate forecasts. For this purpose, 10700 test day records belonging to 1070 dairy cattle were used. Data were divided into two groups of 5350 records in each. One set of observations was used to model parameters, while the remaining was used for evaluating the forecast power of the model and for determining the number of test day records which provide the most accurate forecasts. ARIMA(2,0,0)(1,1,1)(10) model was determined to be suitable and it was used to obtain the forecast values. The expression of the model using estimated parameter values is (1-B-10)(1)y(t) = [(1-0.99129B(10))/(1-0.36889B-0.06934B(2))(1-0.08352B(10))]a(t). Statistically significant and high correlations were determined between the actual and forecast values. The results indicated that the time series approach can be useful for prediction of milk yields.

The aim of this study is to model the test day milk yields via time series methodology and to determine the number of test days which provide the most accurate forecasts. For this purpose, 10700 test day records belonging to 1070 dairy cattle were used. Data were divided into two groups of 5350 records in each. One set of observations was used to model parameters, while the remaining was used for evaluating the forecast power of the model and for determining the number of test day records which provide the most accurate forecasts. ARIMA(2,0,0)(1,1,1)(10) model was determined to be suitable and it was used to obtain the forecast values. The expression of the model using estimated parameter values is (1-B-10)(1)y(t) = [(1-0.99129B(10))/(1-0.36889B-0.06934B(2))(1-0.08352B(10))]a(t). Statistically significant and high correlations were determined between the actual and forecast values. The results indicated that the time series approach can be useful for prediction of milk yields.