2016, vol. 66, br. 3, str. 317-335
|
|
Identifikacija osetljivosti krava na mastitis zasnovana na odabranim genotipovima korišćenjem sistema odluke i generalizovanog linearnog modela
Identification of cows susceptible to mastitis based on selected genotypes by using decision trees and a generalized linear model
aDepartment of Ruminants Science, West Pomeranian University of Technology, Szczecin, Poland bDepartment of Genetics and Animal Breeding, West Pomeranian University of Technology, Szczecin, Poland
e-adresa: daniel.zaborski@zut.edu.pl
Projekat: Project of the Polish Ministry of Science and Higher Education, no. 517-01-028-3962/17
Sažetak
Cilj ispitivanja je bio da se: 1) proveriti mogućnost detektovanja, u ranoj fazi eksploatacije, krava prijemčivih na mastitis na osnovu odabranih genotipova i proizvodnih rezultata u prve tri laktacije, i uz upotrebu metoda objedinjavanja podataka (boostedclassification trees - BT and random forest - RF), 2) proveri da li uključivanje nekih drugih proizvodnih promenljivih veličina tokom kasnijih laktacija može da poboljša detekciju performansi modela, 3) identifikuju najznačajnije osobine koje mogu da predvide prijemčivost na mastitis, 4) uporede rezultati dobijeni upotrebom BT i RF, sa onima koji su dobijeni tradicionalnom linearnom metodom (GLZ). Ukupno je analiziran 801 podatak koji se odnosio na poljsko frizijsko crno-belo goveče. Maksimalna osetljivost, specifičnost i tačnot testa bila je za BT: 72.13%, 39.73%, 55.90%, za RF 86.89%, 17.81% i 59.49% i za GLZ 90.16%, 8.22% i 58.97%. Obuhvatanje drugih promenljivih veličina nije značajno povećalo značajnost efekta modela. Najznačajniji elementi predviđanja prijemčivosti na mastitis su bili: količina mleka, broj dana u laktaciji, ocena bika i procenat holštajn-frizijskih gena. Sa druge strane, sezona telenja kao i genotipovi (laktoferin, α-faktor nekroze tumora, lizocim i defenzini) su daleko manje bili značajni. Primenjeni modeli, su pokazali nizak nivo tačnosti u detekciji krava koje su bile prijemčive na mastitis. Može se zaključiti da je potrebno da se obuhvate neki drugi kriterijumi i elementi na osnovu kojih bi se uspešnije mogla predvideti povećana osetljivost na mastitis kod krava.
Abstract
The aim of the present study was to: 1) check whether it would be possible to detect cows susceptible to mastitis at an early stage of their utilization based on selected genotypes and basic production traits in the first three lactations using ensemble data mining methods (boosted classification tress - BT and random forest - RF), 2) find out whether the inclusion of additional production variables for subsequent lactations will improve detection performance of the models, 3) identify the most significant predictors of susceptibility to mastitis, and 4) compare the results obtained by using BT and RF with those for the more traditional generalized linear model (GLZ). A total of 801 records for Polish Holstein-Friesian Black-and-White cows were analyzed. The maximum sensitivity, specificity and accuracy of the test set were 72.13%, 39.73%, 55.90% (BT), 86.89%, 17.81%, 59.49% (RF) and 90.16%, 8.22%, 58.97% (GLZ), respectively. Inclusion of additional variables did not have a significant effect on the model performance. The most significant predictors of susceptibility to mastitis were: milk yield, days in milk, sire's rank, percentage of Holstein-Friesian genes, whereas calving season and genotypes (lactoferrin, tumor necrosis factor alpha, lysozyme and defensins) were ranked much lower. The applied models (both data mining ones and GLZ) showed low accuracy in detecting cows susceptible to mastitis and therefore some other more discriminating predictors should be used in future research.
|