TargetedLearning.Parameters

Exported

TargetedLearning.Parameters.applyparam

Computes the parameter estimate and influence curve for a particular Parmeter given an estimate of Q and g.

Arguments

source: TargetedLearning/src/Parameters.jl:124

TargetedLearning.Parameters.regimen

Converts it's argument a to a StaticRegime if a is a scalar, a DynamicRegime if a is a vector, or returns a if a is already a Regime.

source: TargetedLearning/src/Parameters.jl:60

TargetedLearning.Parameters.ATE{T<:AbstractFloat}

The ATE parameter is (E_0[E_0(Ymid A=d1, W) - E_0(Ymid A=d0, W)]).

Under causal assumptions, this can be interpreted as the difference in mean counterfactual outcome under regimens d1 and d0. When d1 is the static regimen setting treatment to 1 and d0 is the static regimen setting treatment to 0, this is called the averate treatment effect (ATE).

source: TargetedLearning/src/Parameters.jl:92

TargetedLearning.Parameters.DynamicRegimen{T<:AbstractFloat}

A DynamicRegimen sets treatment for each observation to a particular value. Constructed with DynamicRegimen(a) where a a vector of floating point is 0 or 1s.

source: TargetedLearning/src/Parameters.jl:44

TargetedLearning.Parameters.Mean{T<:AbstractFloat}

The Mean parameter is (E_0E_0(Ymid A=d, W)).

Under causal assumptions, this can be interpreted as the mean of the counterfactual outcome under regimen d.

source: TargetedLearning/src/Parameters.jl:78

TargetedLearning.Parameters.Regimen{T<:AbstractFloat}

Represents a particular single time point treatment regimen.

source: TargetedLearning/src/Parameters.jl:23

TargetedLearning.Parameters.StaticRegimen{T<:AbstractFloat}

A StaticRegimen sets treatment to a single value for all observations. Constructed with StaticRegimen(a) where a is a floating point 0 or 1.

source: TargetedLearning/src/Parameters.jl:30