Banca de DEFESA: PETER CHARRIE JANAMPA SARMIENTO
Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.DISCENTE : PETER CHARRIE JANAMPA SARMIENTO
DATA : 03/08/2018
HORA: 09:00
LOCAL: ANFITEATRO DO INSTITUTO DE ZOOTECNIA DA UFRRJ
TÍTULO:
Fitting nonlinear equations to the growth-out phase of commercial rainbow trout.
PALAVRAS-CHAVES:
Keywords: Logistic, Gompertz, Bertalanffy, Brody
PÁGINAS: 84
GRANDE ÁREA: Ciências Agrárias
ÁREA: Zootecnia
SUBÁREA: Produção Animal
RESUMO:
ABSTRACT
JANAMPA-SARMIENTO, Peter Charrie. Fitting nonlinear equations to the growth-out phase of commercial rainbow trout. 2018. 69p. Dissertation (Masters in Animal Science). Zootecnhy Institute, Federal Rural University from Rio de Janeiro, Seropédica, RJ, 2018.
Mathematical models, through non-linear equations, are an important tool to represent from the growth of an animal or even to estimate the best level of a given macro or micronutrient. The present work adjusted 04 nonlinear models (Gompertz, Von Bertalanffy, Logistic and Brody) to a data set of growth in weight and length of rainbow trout (Oncorhynchus mykiss) during 98 days of cultivation in the fattening stage under conditions of commercial cultivation. A total of 900 trout of average initial weight and length and post-hatching age of 122.11 ± 15.6g; 22.42 ± 0.71cm; and 273 days, respectively, in nine identical tanks, with three tanks being fed with three extruded diets with different levels of Digestible Energy. The fit was based on the Least Squares theory using the Marquardt iterative method. The computational procedures were performed by the NLIN from SAS. The evaluation of the adjusted models was made by adjusToneladasent criteria: convergence capacity, coefficient of determination (R2), mean square of the residue (MSR), Akaike criterion (AIC), mean absolute deviation of the residuals (MAD), mean percentage error (MPE), congruence and usefulness of the information generated by the adjusted model the biological growth of trout, and the examination and distribution of residues and studentized residues. The Logistico, Gompertz and Bertalanffy models were able to adjust to the data by weight and were sensitive to variations in growth because of the different diets provided; however, the parameters A (between 580.10 to 714.10), B (between 0.0196 to 0.0346) and T (between 311.80 to 341.40) obtained by the logistic model reached the best values of the evaluators of adjusToneladasent for this type of data (R2 ≤0.9460 - 0,8051≥; MSR ≤2889,60 - 1223,80≥; AIC ≤14062,06 - 1391,37≥ MAD ≤35,92 - 24,59 ≥), demonstrating a better capacity to describe the growth in weight of rainbow trout in the fattening stage in commercial growing conditions, even if its predictive values tend to overestimate and the presence of discrepant values; however, in order to contrast the predictive capacities of growth through these models as well as other alternative nonlinear models, it is recommended to verify the adjusToneladasent in a perspective of global growth, that is, to include the early growth phases of this fish species. Finally, data in length had difficulties in adjusting adequately in all models, recommending the search of other models with more flexible properties for this type of information.
MEMBROS DA BANCA:
Externo à Instituição - CLEBER FERNANDO MENEGASSO MANSANO - UNESP
Presidente - 066.467.316-32 - MARCELO MAIA PEREIRA - OUTRO
Interno - 1899630 - VINICIUS PIMENTEL SILVA