API Reference

fastloess

---

fastloess / SmoothOptions

Interface: SmoothOptions

Defined in: index.d.ts:253

Configuration options for LOESS smoothing.

Properties

auto_converge?

optional auto_converge?: number

Defined in: index.d.ts:269

Auto-convergence tolerance. Default: None.

---

boundary_degree_fallback?

optional boundary_degree_fallback?: boolean

Defined in: index.d.ts:292

Reduce polynomial degree to linear at boundary vertices (default true).

---

boundary_policy?

optional boundary_policy?: string

Defined in: index.d.ts:265

Boundary handling (“extend”, “reflect”, “zero”, “noboundary”). Default: “extend”.

---

cell?

optional cell?: number

Defined in: index.d.ts:288

Interpolation cell size (default 0.2). Smaller = more vertices, higher accuracy.

---

cv?

optional cv?: CvOptions

Defined in: index.d.ts:273

Grouped cross-validation configuration for Batch smoothing.

---

degree?

optional degree?: string

Defined in: index.d.ts:278

Polynomial degree (“constant”, “linear”, “quadratic”, etc.). Default: “linear”.

---

dimensions?

optional dimensions?: number

Defined in: index.d.ts:280

Number of predictor dimensions. Default: 1.

---

distance_metric?

optional distance_metric?: string

Defined in: index.d.ts:282

Distance metric (“normalized”, “euclidean”, “manhattan”, “chebyshev”, “minkowski:p”, “weighted”). Default: “normalized”.

---

fraction?

optional fraction?: number

Defined in: index.d.ts:255

Smoothing fraction (0 < fraction <= 1). Default: 0.67.

---

interpolation_vertices?

optional interpolation_vertices?: number

Defined in: index.d.ts:290

Maximum number of interpolation vertices.

---

intervals?

optional intervals?: IntervalsOptions

Defined in: index.d.ts:274

---

iterations?

optional iterations?: number

Defined in: index.d.ts:257

Number of robustness iterations. Default: 3.

---

missing?

optional missing?: string

Defined in: index.d.ts:296

Policy for non-finite (NaN/Inf) values in input data (“error”, “drop”). Default: “error”.

---

outputs?

optional outputs?: string[]

Defined in: index.d.ts:271

Optional output components: diagnostics, residuals, weights, gradient (or derivative), se, sorted.

---

parallel?

optional parallel?: boolean

Defined in: index.d.ts:276

Enable parallel execution. Default: true.

---

retain_model?

optional retain_model?: boolean

Defined in: index.d.ts:298

Retain the fitted model’s training data, enabling LoessResult.predict(). Default: false.

---

robustness_method?

optional robustness_method?: string

Defined in: index.d.ts:261

Robustness method (“bisquare”, “huber”, “talwar”). Default: “bisquare”.

---

scaling_method?

optional scaling_method?: string

Defined in: index.d.ts:267

Scaling method (“mad”, “mar”, “mean”). Default: “mad”.

---

seed?

optional seed?: number

Defined in: index.d.ts:294

Non-negative JavaScript safe-integer seed for reproducible K-fold cross-validation splits.

---

surface_mode?

optional surface_mode?: string

Defined in: index.d.ts:286

Surface mode (“interpolation” or “direct”). Default: “interpolation”.

---

weight_function?

optional weight_function?: string

Defined in: index.d.ts:259

Weight function (“tricube”, “epanechnikov”, “gaussian”, “uniform”, “biweight”, “triangle”, “cosine”). Default: “tricube”.

---

weighted_metric_weights?

optional weighted_metric_weights?: number[]

Defined in: index.d.ts:284

Per-dimension weights for the “weighted” distance metric.

---

zero_weight_fallback?

optional zero_weight_fallback?: string

Defined in: index.d.ts:263

Fallback strategy when weights are zero (“use_local_mean”, “return_original”, “return_none”). Default: “use_local_mean”.