Method
for the determination and interpretation of multivariate response functions
Gy. Biczók Ministry of Environmental Protection and Regional
Development,
L. Tolner, University of Agricultural Sciences,
Gy. Simán, University of Agricultural Sciences,
ABSTRACT
Soil fertility is influenced by the joint effect of a
number of factors. The vastly increasing number of experimental treatments
required as the number of variables increases causes problems when multivariate
response functions are estimated, as does the interpretation of the large
number of parameters in the correlation. Interpretation is complicated by the
fact that the more parameters the correlation contains, the greater the problem
caused by the interdependence of the estimated parameters.
It is possible to estimate the parameters independently
and to reduce the number of experiments to a realistic level if the experiments
are set up according to an orthogonal factor design, though these designs are
not suitable for calculating the parameters of Baule-Mitscherlich function
containing combinations of exponential terms.
The multivariate extension of the parabolic response
function recommended here leads to a quadratic polynomial form, the parameters
of which can be reasonably estimated by applying an orthogonal factor design.
The interpretation of the model parameters was investigated considering the
optimum values of the individual factors
and the possible characteristics of the
available nutrient forms in the soil.