README
¶
An implementation of B.P. Welford's algorithm to maintain a running variance over a stream of observations.
This code is based on a John D Cook's blogpost.
Example usage
s := stats.NewSink()
s.Push(-1.1813782)
s.Push(-0.2449577)
s.Push(0.8799429)
fmt.Printf("Mean = %v\n", s.Mean()) // -0.182131...
fmt.Printf("Standard Deviation = %v\n", s.StandardDeviation()) // 1.032095...
Todos
- Add a method that returns the running median.
- Make the implementation multi-goroutine safe.
License
- Code is released under MIT license.
Documentation
¶
Index ¶
Constants ¶
This section is empty.
Variables ¶
var ErrInvalidValue = errors.New("NaN or +-Inf are not allowed")
ErrInvalidValue means that the item you supplied is a NaN or an Inf which is considered illegal.
Functions ¶
This section is empty.
Types ¶
type Variability ¶
type Variability interface {
Min() float64
Max() float64
Mean() float64
StandardDeviation() float64
Count() int
}
Variability is a measure of how spread out is your dataset.
type WelfordSink ¶
type WelfordSink struct {
Sink
Variability
// contains filtered or unexported fields
}
WelfordSink is an implementation of a Sink and Variability @TODO make this goroutine-safe
func NewSink ¶
func NewSink() *WelfordSink
NewSink returns a pointer to an implementation of interfaces Sink & Variability
func (*WelfordSink) Count ¶
func (s *WelfordSink) Count() int
Count returns the number of observations that was seen thus far.
func (*WelfordSink) Max ¶
func (s *WelfordSink) Max() float64
Max returns the largest observation seen.
func (*WelfordSink) Mean ¶
func (s *WelfordSink) Mean() float64
Mean returns the arithmetic mean of all the observations.
func (*WelfordSink) Min ¶
func (s *WelfordSink) Min() float64
Min returns the smallest observation seen.
func (*WelfordSink) Push ¶
func (s *WelfordSink) Push(x float64) error
Push accepts an observation. On failure, it returns an ErrInvalidValue.
func (*WelfordSink) StandardDeviation ¶
func (s *WelfordSink) StandardDeviation() float64
StandardDeviation returns the running sd of the observations
Source Files