/** A vector given either as a scalar (1-D case) or an array (k-D case). */
export type MvnVector = number | readonly number[];
/** A covariance given either as a scalar variance (1-D case) or a k×k matrix. */
export type MvnCov = number | readonly (readonly number[])[];
/**
 * Multivariate-normal probability density, handling both the 1-D case
 * (`x`/`mean` scalars, `cov` a scalar variance) and the general k-D case
 * (`x`/`mean` length-k arrays, `cov` a k×k matrix). Reduces to the ordinary
 * normal PDF when k=1.
 *
 * @example
 * mvnPdf([0, 0], [0, 0], [[1, 0], [0, 1]]); // ~0.15915494 (scipy multivariate_normal.pdf)
 * mvnPdf(0, 0, 1); // ~0.3989422804 (standard normal density at 0)
 */
export declare function mvnPdf(x: MvnVector, mean: MvnVector, cov: MvnCov): number;
/** Options for {@link mvnSample}. */
export interface MvnSampleOptions {
    /** Seed for the deterministic RNG (reproducible draws). Omit for a
     * time-seeded, non-reproducible generator. */
    seed?: string | number;
}
/**
 * Draw `n` samples from a multivariate normal `N(mean, cov)` via `x = mean +
 * L·z`, where `Σ = LLᵀ` (Cholesky) and `z` is a vector of independent standard
 * normals (Box-Muller). Handles the 1-D case the same as {@link mvnPdf}.
 * Returns an `n`-length array of length-k row vectors.
 *
 * @example
 * mvnSample([1, 2], [[2, 0.5], [0.5, 1]], 20000, { seed: 42 });
 * // empirical mean ~= [1, 2], empirical covariance ~= [[2, 0.5], [0.5, 1]]
 */
export declare function mvnSample(mean: MvnVector, cov: MvnCov, n: number, opts?: MvnSampleOptions): number[][];
//# sourceMappingURL=mvn.d.ts.map