Interface: BayesianPredictRequest

@kortexya/reasoninglayer

---

@kortexya/reasoninglayer / Inference / BayesianPredictRequest

Interface: BayesianPredictRequest

Defined in: src/types/inference.ts:338

Request for Bayesian prediction.

Remarks

Dual input mode: Provide EITHER goal OR goal_id.

Unlike other inference requests, Bayesian prediction requires an effect_var naming the variable that holds the FuzzyNumber effect.

Properties

conformal?

optional conformal: number | null

Defined in: src/types/inference.ts:350

Optional conformal coverage (e.g., 0.9 for 90% coverage).

---

effectVar

effectVar: string

Defined in: src/types/inference.ts:340

Name of the variable that holds the FuzzyNumber effect.

---

goal?

optional goal: TermInputDto | null

Defined in: src/types/inference.ts:342

Inline goal (should bind a FuzzyNumber variable). Mutually exclusive with goal_id.

---

goalId?

optional goalId: string | null

Defined in: src/types/inference.ts:344

Saved goal reference. Mutually exclusive with goal.

---

maxDepth?

optional maxDepth: number

Defined in: src/types/inference.ts:346

Maximum proof depth (default: 100).

---

minDegree?

optional minDegree: number

Defined in: src/types/inference.ts:348

Minimum degree threshold.

---

robust?

optional robust: boolean

Defined in: src/types/inference.ts:352

Use robust aggregation (Student-t/Huber) to downweight outliers.

---

saveGoal?

optional saveGoal: boolean

Defined in: src/types/inference.ts:354

If true, save the goal for later reuse (returns goal_id in response).