holt_winters

Documentation

Index

• type HoltWinters
  • func New(n, s int, withFit bool, alloc memory.Allocator) *HoltWinters
  • func (r *HoltWinters) Do(vs *array.Float) (*array.Float, float64)
• type Optimizer
  • func NewOptimizer(alloc memory.Allocator) *Optimizer
  • func (o *Optimizer) Optimize(objfunc func(*mutable.Float64Array) float64, start *mutable.Float64Array, ...) (float64, *mutable.Float64Array)

Types

type HoltWinters

type HoltWinters struct {
	// contains filtered or unexported fields
}

HoltWinters forecasts a series into the future. This is done using the Holt-Winters damped method.

1. The initial values are calculated using a SSE.
2. The series is forecast into the future using the iterative relations.

func New

func New(n, s int, withFit bool, alloc memory.Allocator) *HoltWinters

New creates a new HoltWinters. HoltWinters uses the given allocator for memory tracking purposes, and in order to build its result.

func (*HoltWinters) Do

func (r *HoltWinters) Do(vs *array.Float) (*array.Float, float64)

Do returns the points generated by the HoltWinters algorithm given a dataset.

type Optimizer

type Optimizer struct {
	// Maximum number of iterations.
	MaxIterations int
	// Reflection coefficient.
	Alpha,

	Beta,

	Gamma float64
	// contains filtered or unexported fields
}

Optimizer is an implementation of the Nelder-Mead optimization method. Based on work by Michael F. Hutt: http://www.mikehutt.com/neldermead.html

func NewOptimizer

func NewOptimizer(alloc memory.Allocator) *Optimizer

NewOptimizer returns a new instance of Optimizer with all values set to the defaults.

func (*Optimizer) Optimize

func (o *Optimizer) Optimize(
	objfunc func(*mutable.Float64Array) float64,
	start *mutable.Float64Array,
	epsilon,
	scale float64,
) (float64, *mutable.Float64Array)

Optimize applies the Nelder-Mead simplex method with the Optimizer's settings. Optimize returns a new Float64Array, it is responsibility of the caller to Release it.