InhomogeneousBernoulliProcess

multinomial.InhomogeneousBernoulliProcess(
    self
    alpha
    rho
    hypothesis
    weights
    k=100
)

Conversion rate optimization test from Lindon and Malek (2022).

• Parent class: Multinomial
• Example

Parameters

Name	Type	Description	Default
alpha	float	Probability of Type I error \(\alpha\).	required
rho	np.ndarray	Assignment probabilities \(\mathbf{\rho}\).	required
hypothesis	Callable[[cp.Variable], List[cp.Constraint]]	Function to generate hypothesis constraints.	required
weights	np.ndarray	Contrast weights \(W\).	required
k	float	Concentration for Dirichlet prior parameters \(\mathbf{\alpha}_0 = k \mathbf{\rho}\).	100

Attributes

Name	Description
contrasts	Estimate of contrasts \(\hat{W \mathbf{\delta}}\).
hypothesis	Function to generate hypothesis constraints.
weights	Contrast weights \(W\).

References

Lindon, Michael, and Alan Malek. 2022. “Anytime-Valid Inference for Multinomial Count Data.” In Advances in Neural Information Processing Systems, edited by Alice H. Oh, Alekh Agarwal, Danielle Belgrave, and Kyunghyun Cho. https://openreview.net/forum?id=a4zg0jiuVi.