/adwin

Adwin is an adaptive windowing algorithm. It is from `Learning from time-changing data with adaptive windowing, Bifet, Albert, and Ricard Gavalda; Proceedings of the 2007 SIAM international conference on data mining. Society for Industrial and Applied Mathematics, 2007`.

Primary LanguageGoMIT LicenseMIT

Adwin and Adwin2 Actions Status

Adwin is an adaptive windowing algorithm. It is from Learning from time-changing data with adaptive windowing, Bifet, Albert, and Ricard Gavalda; Proceedings of the 2007 SIAM international conference on data mining. Society for Industrial and Applied Mathematics, 2007.

Adwin2 improves time and memory requirements by using Exponential Histograms.

See also:

Usage

Adwin

	delta := 0.01
	adwin := NewAdwin(delta)
	adwin.Conservative(true) // if you wants

	// Add stream data
	adwin.Add(x)

	// Show status
	adwin.Size()
	adwin.Sum()
	adwin.Detected()

Adwin2

	delta := 0.01
	adwin2 := NewAdwin2(delta)
	adwin2.Conservative(true) // if you wants

	// Add stream data
	adwin2.Add(x)

	// Show status
	adwin2.Size()
	adwin2.Sum()
	adwin2.Detected()

Installation

$ go get github.com/monochromegane/adwin

License

MIT

Author

monochromegane