Ajuste e avaliação da tendência e da precisão de modelos de predição do peso final e da conversão alimentar de suínos em fases de crescimento e terminação
| Ano de defesa: | 2019 |
|---|---|
| Autor(a) principal: |
Oliveira, Anderson Luís Garla
 |
| Orientador(a): |
Carvalho, Paulo Levi de Oliveira
|
| Banca de defesa: |
Carvalho, Paulo Levi de Oliveira
,
Nunes, Ricardo Vianna
,
Pozza, Paulo Cesar
|
| Tipo de documento: | Dissertação |
| Tipo de acesso: | Acesso aberto |
| Idioma: | por |
| Instituição de defesa: |
Universidade Estadual do Oeste do Paraná
Marechal Cândido Rondon |
| Programa de Pós-Graduação: |
Programa de Pós-Graduação em Zootecnia
|
| Departamento: |
Centro de Ciências Agrárias
|
| País: |
Brasil
|
| Palavras-chave em Português: | |
| Área do conhecimento CNPq: | |
| Link de acesso: | http://tede.unioeste.br/handle/tede/4806 |
Resumo: | The aim of this work was to verify the performance of mathematical models in the prediction of zootechnical results in swine growth and finishing phases. The data were collected from 663 records of swine batches of average performance, between 2013 and 2015. The animals were with an average weight between 5.7 and 28.4 kg, from a company located in Western Paraná. By screening the mean performance values, the data were obtained on the formation of a database regarding the average variables of the analyzed batch: number of animals housed in the lot (NA), initial average body weight (IABW), initial average age (IAA), number of animals housed at the end of rearing (NAE), final average body weight (FABW), final average age (FAA), feeding intake per animal per day (FIA), feeding intake in the growing phase (FIG), mortality (MORT), daily weight gain (DWG) and feed conversion (FC). The data were classified into eight categories resulting from the combination of two periods of the year (P): from November to April (P1) and from May to October (P2), and four farm sizes (FS): NA ≤ 1000 (FS1); 1001 ≤ NA ≤ 2000 (FS2); 2001 ≤ NA ≤ 3000 (FS3) and NA > 3000 (FS4). Student's standardized residuals (RStudent) analysis was carried out. The estimated regression equation of the FABW and CA as a function of the performance regressors, P and FS were: FABW = -46.13 + 0.00214NA + 0.87467IABW - 0.85194IAA + 0.85184FAA + 23.72FIA - 0.02151DP*IAA - 0.73775DFS2*FIA - 0.00129DFS3*NA (R2 = 0.9083) and FC = 1.1762 - 0.00006NA + 0.00335IABW + 0.00091FAA + 0.50477FIA + 0.01481DP*FIA + 0.00002DFS2*NA + 0.000037DFS3*NA (R2 = 0.5794). After the substitution of binary data for the P and FS, eight daughter equations for FABW and eight daughter equations for FC were originated. The coefficient of determination (R2) showed that the estimated regression equation explained about 58% of the FC data variation as a function of regressors of performance, P and FS, suggesting median reliability. The FAA showed a higher correlation with the FABW, this variable can explain 71.25% of the variations in the FABW data. The FIA was more correlated with FC (rXY = 0.7087) than the other regressors. The FIA positively influenced FC, but this influence was stronger for lots from May to October (0.51958) than those from November to April (0.50477). The results indicated that observed FABW values (FABWO) of pig lots were similar to the predicted FABW values (FABWP) of pig lots using the models FABWE = -46.13 + 0.00214NA + 0.87467IABW - 0,87345IAA + 0.85184FAA + 23.72FIA, for farmhouses with up to 1,000 housed pigs from May to October and FABWH = -46.13 + 0.00085NA + 0.87467IABW - 0.87345IAA + 0.85184FAA + 23.72FIA, for farmhouses above 3,000 housed pigs from May to October. The FABWE and FABWH models were not considered precise, due to the estimated correlations between FABWPE and FABWOE (rXY = 89.83%) and FABWPH and FABWOH (rXY = 91.67%) being lower than their respective average hit index (1 – AREi), of 97.01% for FABWE and 97.53% for FABWH. |