Agriculture Reference
In-Depth Information
Inputs
Daily feed intake (DFI)
Body weight (BW)
Genetic potential
Empirical component
Modelling techniques
recursive
BW and
∂
BW
DFI
Protein
deposition
NE intake
Factorial equation
Mechanistic component
Nutrient requirements
g/day
Optimal concentration in the diet
g/kg
Fig. 12.3.
General outline of the model with the empirical and mechanistic model components used to
estimate daily nutrient requirements for each individual from one pig population according to their actual
growth and feed intake patterns. (From Hauschild
et al
., 2012.)
The empirical component of the model
is used to estimate the starting day (
t
+ 1)
daily feed intake (DFI), BW and DG from the
information measured in each individual
animal up to the current day (
t
). The object-
ive of this approach is to follow in real time
the dynamic DFI and BW trajectories of each
pig from the herd. Short-term variation over
time and temporary drops in performance
are not taken into account in this model. For
this, forecasting methods using exponential
smoothing techniques, which weigh recent
and past observations in different ways, are
appropriate because they reduce the fluctu-
ations from irregular observations in the ob-
served time series (Claycombe and Sullivan,
1977). The double exponential smoothing
forecasting time series method was chosen
in this model because DFI and BW of pig
populations show evident long-run trends
