Interface: DynamicDiscoveryRequest

@kortexya/reasoninglayer / Statistical / DynamicDiscoveryRequest

Interface: DynamicDiscoveryRequest

Defined in: src/types/statistical.ts:271

Request for dynamic causal discovery with configurable strategy.

Remarks

Supports constraint-based (PC), Bayesian (MCMC), and hybrid strategies for learning causal structure from observational data.

Properties

alpha?

optional alpha: number | null

Defined in: src/types/statistical.ts:273

Significance level for independence testing (default: 0.05).

edgeThreshold?

optional edgeThreshold: number | null

Defined in: src/types/statistical.ts:275

Hybrid-specific: edge confidence threshold (default: 0.5).

mcmcBurnIn?

optional mcmcBurnIn: number | null

Defined in: src/types/statistical.ts:277

MCMC-specific: burn-in period (default: 100).

mcmcSamples?

optional mcmcSamples: number | null

Defined in: src/types/statistical.ts:279

MCMC-specific: number of samples (default: 500).

minSamples?

optional minSamples: number | null

Defined in: src/types/statistical.ts:281

Minimum samples before discovery (default: 30).

strategy?

optional strategy: DiscoveryStrategy | null

Defined in: src/types/statistical.ts:283

Structure learning strategy (default: “pc”).

temporalTiers?

optional temporalTiers: string[][] | null

Defined in: src/types/statistical.ts:289

Temporal tiers for background knowledge (tPC algorithm). Each tier is a list of variables that belong to that temporal tier. Tiers are ordered from past to future (tier 0 = earliest).

useActiveLearning?

optional useActiveLearning: boolean | null

Defined in: src/types/statistical.ts:291

Enable active learning recommendations (default: true).

useGes?

optional useGes: boolean | null

Defined in: src/types/statistical.ts:293

Enable GES refinement (default: true).