PoMoS: Polynomial (ordinary differential equation) Model Search
PoMoS-package aims to determine from a set of N time
series the optimal polynomial structure of a model built on
first-order ordinary differential equations. The core of the
package is based on the poMoS function: an evolutionary
algorithm combined with a least square fitting. Optimality is
estimated with AIC (Akaike, 1974) or AIC-like criterions.
Although efficient in its selection, the identification of the
optimal structure cannot be guaranteed. Therefore, both
selected and rejected models are reconsidered after optimal
solutions are obtained from the evolutionary algorithm for
another analysis. This analysis is based on a statistical
evaluation of the regressors' quality.
||R (≥ 2.10), RGtk2 (≥ 2.12.18), cairoDevice (≥ 2.0), igraph (≥ 0.5.4)
||Mangiarotti S., Coudret R., Drapeau L.
||Drapeau L. <laurent.drapeau at ird.fr>
||Gtk+ (>= 2.16.0)