|
This experiment, in comparison slaughtering method Found SBM net energy (NE) based on the value, a comparative study of the near infrared, and chemical composition to establish the feasibility of the NE prediction model Fourier; expanded by changing the moisture content of a sample, and explored Fourier Transform Near Infrared establish the NE prediction model improved effect. (1) soybean meal NE value measured method using to maintain NE deposited NE. Maintain NE using multiple linear regression analysis, to set ad libitum and restricted feeding 30%, 50% and 70% of the 4 treatment groups. Determination of 21 kinds of soybean meal deposition NE sets of algorithms, each of soybean meal as a treatment, while collecting excreta Determination of soybean meal apparent metabolizable energy (AME). The trial period is 8-16 d of age. For each treatment six replicates, each 2 chicken, try chicken average weight of 66.5 ± 2.1g. (2) Determination of the the 21 conventional chemical composition of the samples of soybean meal, and the NE apparent metabolizable can (AME), crude protein (CP), starch (ST), crude fiber (CF), neutral detergent fiber (NDF), acid detergent fiber (ADF), crude ash (ASH), a multiple linear regression analysis. (3) to adjust the moisture content of the 21 known NE value of soybean meal samples were 11%, 12% and 13%, respectively, and were established the NE prediction model of the three kinds of water and its global NIRS. The results show that: (1) 0-3 weeks old Yellow Chicken 21 SBM net value of 6.045-7.829 MJ / kg DM, AME converted to the efficiency of the NE 55.24-62.78%. (2) to establish the chemical composition of soybean meal NE prediction equations R2 of 0.96, the RSD was 0.114 MJ / kg of the DM; chemical composition combined with AME to establish the best prediction equation R2 is 0.98 the RSD to 0.079 MJ / kg DM . (3) the three moisture range as well as the global soybean meal NE near infrared model calibration coefficient of determination (R2cal) were 0.96,0.98,0.97,0.94 correction standard deviation (RMSEE) were 0.100,0.072,0.069,0.105 MJ / kg. Cross-validation coefficient of determination (R2cv) were 0.92,0.95,0.95,0.93; the cross verification standard deviation (RMSECV) in were 0.131,0.096,0.089 and 0.116MJ/kg. To sum up: the the NE prediction equation (1) combined with AME chemical composition to establish superior to the NE prediction equations created using only the chemical composition. (2) by adjusting the moisture content to expand the sample to establish satisfactory soybean meal NE near infrared prediction model. (3) established with near infrared the NE prediction model with predictive models created using only the chemical composition effect considerable, than the effect of the chemical composition build predictive models combined with AME somewhat less.
|