Dynamics
Models predicting the changes in variant abundances over time.
Import as
covvfit.dynamics.JointLogisticGrowthParams (tuple)
This is a model of logistic growth (selection dynamics)
in K cities for V competing variants.
We assume that the relative growth advantages do not change between the cities, however we allow different introduction times, resulting in different offsets in the logistic growth model.
This model has V-1 relative growth rate parameters
and K*(V-1) offsets.
Attrs:
relative_growths: relative growth rates, shape `(V-1,)`
relative_offsets: relative offsets, shape `(K, V-1)`
n_cities: int
property
readonly
Number of cities.
n_params: int
property
readonly
Number of all parameters in the model.
n_variants: int
property
readonly
Number of variants.
__getnewargs__(self)
special
Return self as a plain tuple. Used by copy and pickle.
__new__(_cls, relative_growths, relative_offsets)
special
staticmethod
Create new instance of JointLogisticGrowthParams(relative_growths, relative_offsets)
__repr__(self)
special
Return a nicely formatted representation string
from_vector(theta, n_variants)
classmethod
Wraps a vector with parameters of shape (dim,) to the model.
Note that dim should match the number of parameters.
predict_log_abundance(self, timepoints)
Predicts the abundances at the specified time points.
to_vector(self)
Wraps all the parameter into a single vector.
Note
This function is useful for optimization purposes, as many optimizers accept vectors, rather than tuples.