Estimation of infection in dairy cows with different categories by using artificial neural network

The dairy industry has been encountered to the problem with the determination of the oestrus in milking cows. This significantly impairs the productivity of high yielding dairy cows which in turn increases the cost of milk and meat with negative implications on the affordoble of these animal products by the consumer. In this study, it was aimed to determine the estrus detection of holstein frisien milking cows reared in an enterprise with herd management system with the least error rate by using artificial intelligence methods. In order to achieve this, the rate of change in the number of movements, change in the number of ruminantion, rate of change in the milk yield and conductivity during milking of the cattle taken at two hour intervals as input data were processed. In order to provide the most reliable information, artificial neural networks have been modeled in different categories. As a result of the output data obtained, the status of dairy cattle was determined as eostrus or not. In this study, estrus accuracy ratio was determined by using artificial neural network models in different categories and it was observed that these accuracy rates ranged between 85% and 93% for each training of the data set.