Hagassou, Djangsou (2016) Application of support vector machines as support to early prediction of mastitis in Sarda dairy ewes. Doctoral Thesis.
|Full text disponibile come PDF Richiede visualizzatore di PDF come GSview, Xpdf o Adobe Acrobat Reader|
Mastitis is one of the major reasons for the reduction of both milk yield and quality with consequent strong economic losses.
The Sardinian dairy ewe industry consists of 2,400,000 ewes, which is of relevant economic importance for the Island (SRDP 2007 - 2013).
In normal standard hygienic conditions of ewe farm, the prevalence of clinical mastitis is about 5%. Unfortunately it represents the tip of the iceberg. The submerge part of the iceberg is represented by subclinical mastitis, which in some cases can reach 65% of the flock (Fthenakis and Jones, 1990).
The economic loss due to mastitis is estimated around 69 NZ$ (≈ 42 €) per infected half-udder per entire period of lactation (Cuccuru et al., 2011).
The current investigation presents an alternative approach to predict the presence of udder inflammation in Sarda dairy ewes through the application of Support Vector Machines (SVMs), a sub-discipline in the field of artificial intelligence. The milk lactose content (MLC) and milk electrical conductivity (MEC) are used as predictive variable and the milk somatic cell count (MSCC) as classifier. Data used was collected from a 10-years historical database of ARAS laboratory (Sardinian Regional Association of Farmers). The SVMs has shown a sensitivity and specificity of 62% and 75% respectively. This method could therefore to be suggested as first screening system for udder inflammation detection before carrying out expensive and time-consuming bacteriological analysis.
I documenti depositati in UnissResearch sono protetti dalle leggi che regolano il diritto d'autore
Repository Staff Only: item control page