Skip to content
Kortexya Kortexya Reasoning Layer SDK

Interface: AteEstimateRequest

@kortexya/reasoninglayer

@kortexya/reasoninglayer / Causal / AteEstimateRequest

Interface: AteEstimateRequest

Defined in: src/types/causal.ts:724

Request to estimate the ATE (or ATT) from a discrete-covariate sample.

Properties

identification?

optional identification: IdentificationRefDto | null

Defined in: src/types/causal.ts:730

Gate this estimation on a persisted identification certificate (CausalClient.identifyCausalEffect). Absent means legacy behavior, honestly labeled caller_asserted in the response.

knownPropensity?

optional knownPropensity: number | null

Defined in: src/types/causal.ts:736

Declare a known constant treatment mechanism pi: this value is used for every unit instead of estimating it. Must be in (0, 1). Ignored by the ATT endpoints.

observations

observations: ObservationDto[]

Defined in: src/types/causal.ts:738

The observed sample: covariate X, treatment A, outcome Y.

randomized?

optional randomized: boolean | null

Defined in: src/types/causal.ts:744

Declare the treatment randomized (A independent of X): the constant marginal propensity is used and the covariate-adjusted EIF is selected. Defaults to false. Ignored by the ATT endpoints.