The analysis of simulated sow herd datasets using decision tree technique
2004
Kirchner, K. | Tolle, K.H. | Krieter, J.
The ability of machine learning techniques, especially the decision tree method, to analyse pig production data was investigated. The C4.5-classification algorithm was used to detect threshold values of management decisions relating to sows' replacement. In order to generate side-effect-free data, three pig production herds, each on a different performance level, were created using a Monte Carlo simulation. The results of applying the C4.5-algorithm on simulated datasets showed different threshold values within the trees depending on sow herd performance. The evaluation parameters differed due to the adjusted number of instances per class and the dataset size. The sensitivity reached a value of 42.1-72.7%, the kappa value was 44.5-78.1%, and the error rate fluctuated between 6.0 and 31.3%. The overall classification accuracy ranged from 85.8 to 93.4% and the specificity reached a value of 93.7-98.8%. The generated decision trees, visualising the threshold values, varied their number of leaves between 2 and 31 and the number of nodes from 3 to 61.
Afficher plus [+] Moins [-]Mots clés AGROVOC
Informations bibliographiques
Cette notice bibliographique a été fournie par National Agricultural Library
Découvrez la collection de ce fournisseur de données dans AGRIS