|
frederick, MD | These companies that are collecting this data need to figure out how to filter it. Much like USDA and the bull studs have done with bull proofs.
For example when gathering data to prove a bull , the daughters deviation from herd mates is huge. Statistically it has more meaning then just the raw number from the daughter. Also the proof process limits data from one herd. It also weighs the probability that the data from a herd is accurate.
In the old days the bull studs paid good money to dairy farmers for the information. We had to be on test and keep accurate records. When I took breeding school at a large dairy we where told they used all young sires. Because the owner wanted the check for the info the stud farm would get when the daughter completed a lactation.
On the final day of the breeding school the herds people brought in about 30 cows for use to breed. We asked what semen do we use? The herds people told us it doesn't matter. Just breed them. So 7 or 8 of us students grabbed whatever out of the can and bred whatever cow. Then a man came in and wrote the cows numbers down. But he never recorded or asked who we bred them too. I assume that later someone made up the sire numbers when these where entered into the computer.
So my point is, the data on bull daughters from this large farm where completely bogus. I guess that is why now if a bull looks good the stud company will come out and pull DNA samples of the daughters. To confirm that they are in fact who we say they are. I have been told that since they started confirming the daughters they need much less data to prove the bull. | |
|
Why are farmers so enamored with “Big Data”
