pharmpy.parameter module

class pharmpy.parameter.Parameter(name, init, lower=None, upper=None, fix=False, unit=1)

Bases: object

A single parameter

Example

>>> from pharmpy import Parameter
>>> param = Parameter("TVCL", 0.005, lower=0)
>>> param.fix = True

Parameters

• name (str) – Name of the parameter

• init (number) – Initial estimate or simply the value of parameter.

• fix (bool) – A boolean to indicate whether the parameter is fixed or not. Note that fixing a parameter will keep its bounds even if a fixed parameter is actually constrained to one single value. This is so that unfixing will take back the previous bounds.

• lower (number) – The lower bound of the parameter. Default no bound. Must be less than the init.

• upper (number) – The upper bound of the parameter. Default no bound. Must be greater than the init.

• unit (str or unit expression) – Unit of parameter. Default to unitless

copy()

Create a copy of this Parameter

property init

Initial parameter estimate or value

is_close_to_bound(value=None, zero_limit=0.01, significant_digits=2)

Check if parameter value is close to any bound

Parameters

• value (number) – value to check against parameter bounds. Defaults to checking against the parameter init

• zero_limit (number) – maximum distance to 0 bounds

• significant_digits (int) – maximum distance to non-zero bounds in number of significant digits

Examples

>>> from pharmpy import Parameter
>>> par = Parameter("x", 1, lower=0, upper=10)
>>> par.is_close_to_bound()
False

>>> par.is_close_to_bound(0.005)
True

>>> par.is_close_to_bound(0.005, zero_limit=0.0001)
False

>>> par.is_close_to_bound(9.99)
True

>>> par.is_close_to_bound(9.99, significant_digits=3)
False

property lower

Lower bound of the parameter

property parameter_space

The parameter space set

A fixed parameter will be constrained to one single value and non-fixed parameters will be constrained to an interval possibly open in one or both ends.

Examples

>>> import sympy
>>> from pharmpy import Parameter
>>> par = Parameter("x", 1, lower=0, upper=10)
>>> sympy.pprint(par.parameter_space)
[0, 10]

>>> par.fix = True
>>> sympy.pprint(par.parameter_space)
{1}

property symbol

Symbol representing the parameter

unconstrain()

Remove all constraints from this parameter

Example

>>> from pharmpy import Parameter
>>> par = Parameter("x", 1, lower=0, upper=2)
>>> par.unconstrain()
>>> par
Parameter("x", 1, lower=-oo, upper=oo, fix=False)

property upper

Upper bound of the parameter

class pharmpy.parameter.Parameters(params=None)

Bases: collections.abc.MutableSequence

A collection of parameters

Class representing a group of parameters. Usually all parameters in a model. This class give a ways of displaying, summarizing and manipulating more than one parameter at a time.

Specific parameters can be found using indexing on the parameter name

Example

>>> from pharmpy import Parameters, Parameter
>>> par1 = Parameter("x", 0)
>>> par2 = Parameter("y", 1)
>>> pset = Parameters([par1, par2])
>>> pset["x"]
Parameter("x", 0, lower=-oo, upper=oo, fix=False)

>>> "x" in pset
True

>>> pset.append(Parameter("z", 0.1))
>>> len(pset)
3

>>> del pset["x"]
>>> len(pset)
2

copy()

Create a deep copy of this Parameters

property fix

Fixedness of parameters as dict

property inits

Initial estimates of parameters as dict

insert(ind, value)

S.insert(index, value) – insert value before index

is_close_to_bound(values=None, zero_limit=0.01, significant_digits=2)

Logical Series of whether values are close to the respective bounds

Parameters

values (pd.Series) – Series of values with index a subset of parameter names. Default is to use all parameter inits

Returns

pd.Series – Logical Series with same index as values

Example

>>> from pharmpy import Parameters, Parameter
>>> par1 = Parameter("CL", 1, lower=0, upper=10)
>>> par2 = Parameter("V", 10, lower=0, upper=100)
>>> pset = Parameters([par1, par2])
>>> pset.is_close_to_bound()
CL    False
V     False
dtype: bool

property lower

Lower bounds of all parameters as a dictionary

property names

List of all parameter names

property nonfixed_inits

Dict of initial estimates for all non-fixed parameters

remove_fixed()

Remove all fixed parameters

simplify(expr)

Simplify expression given constraints of parameters

property symbols

List of all parameter symbols

to_dataframe()

Create a dataframe with a summary of all Parameters

Returns

DataFrame – A dataframe with one row per parameter. The columns are value, lower, upper and fix Row Index is the names

Example

>>> from pharmpy import Parameters, Parameter
>>> par1 = Parameter("CL", 1, lower=0, upper=10)
>>> par2 = Parameter("V", 10, lower=0, upper=100)
>>> pset = Parameters([par1, par2])
>>> pset.to_dataframe()
    value  lower  upper    fix
CL      1      0     10  False
V      10      0    100  False

property upper

Upper bounds of all parameters as a dictionary

Inheritance Diagram

pharmpy.model_factory module pharmpy.parameter_sampling module