regression

Documentation

Overview

Package regression provides online linear regression calculation.

Example

r := New(7)

r.Add(1.5, 4.4)
r.Add(2.9, 1.56)

slope, intercept, stdError := r.CalculateWithStdError()
fmt.Printf("slope %f\n", slope)
fmt.Printf("intercept %f\n", intercept)
fmt.Printf("standard error %f\n", stdError)

r.Add(7.2, 10.5)
r.Add(9, 7.6)

slope, intercept, stdError = r.CalculateWithStdError()
fmt.Printf("slope %f\n", slope)
fmt.Printf("intercept %f\n", intercept)
fmt.Printf("standard error %f\n", stdError)

Output:

slope -2.028571
intercept 7.442857
standard error NaN
slope 1.188768
intercept -1.015158
standard error 3.720348

Index

• type Regression
  • func New(xDelta float64) *Regression
  • func (r *Regression) Add(x, y float64)
  • func (r *Regression) Calculate() (slope, intercept float64)
  • func (r *Regression) CalculateWithStdError() (slope, intercept, stdError float64)

Examples

• Package

Types

type Regression

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

Regression represents a queue of past points. Use New() to initialize.

func New

func New(xDelta float64) *Regression

New returns a Regression that keeps points back as far as xDelta from the last added point.

func (*Regression) Add

func (r *Regression) Add(x, y float64)

Add adds the new x and y as a point into the queue. Panics if given an x value less than the last.

func (*Regression) Calculate

func (r *Regression) Calculate() (slope, intercept float64)

Calculate returns the slope, intercept and standard error of a best fit line to the added points. Returns a cached value if called between adds. Deprecated in favor of CalculateWithStdError.

func (*Regression) CalculateWithStdError

func (r *Regression) CalculateWithStdError() (slope, intercept, stdError float64)

Calculate returns the slope, intercept and standard error of a best fit line to the added points. Returns a cached value if called between adds.