MovingWindowFilter Class

Class implementing data filtering by replacing data points f(i) with a linear combination of the data points immediately to the left and right of f(i). The user provides the coefficients to use in the linear combination. Static class methods are provided for generating coefficients to implement a moving average filter and a Savitzky-Golay smoothing filter.

Inheritance Hierarchy

SystemObject
CenterSpace.NMath.CoreCorrelationFilter
CenterSpace.NMath.CoreMovingWindowFilter

Namespace: CenterSpace.NMath.Core
Assembly: NMath (in NMath.dll) Version: 7.3

Syntax

C++

Copy

[SerializableAttribute]
public class MovingWindowFilter : CorrelationFilter

<SerializableAttribute>
Public Class MovingWindowFilter
	Inherits CorrelationFilter

[SerializableAttribute]
public ref class MovingWindowFilter : public CorrelationFilter

[<SerializableAttribute>]
type MovingWindowFilter =  
    class
        inherit CorrelationFilter
    end

The MovingWindowFilter type exposes the following members.

Constructors

	Name	Description
	MovingWindowFilter	Constructs a MovingWindowFilter instance which implements a moving average. Number of points to the left and right both default to two.
	MovingWindowFilter(Int32, Int32, DoubleVector)	Constructs a MovingWindowFilter instance with the specified parameters.

Top

Properties

	Name	Description
	Coefficients	Gets the filter coefficients.
	NumberLeft	Gets the number of points left for the filter window.
	NumberOfCoefficients	Gets the number of filter coefficients. Effectively the filter width. (Inherited from CorrelationFilter.)
	NumberRight	Gets the number of points right for the filter window.
	WindowBoundaryOption	Gets / Sets the current boundary option.
	WindowWidth	Gets the width of the moving window.

Top

Methods

	Name	Description
	Correlate	Does the correlation and takes care of internal correlation and work objects. (Inherited from CorrelationFilter.)
	ExponentiallyWeightedMovingAverageCoefficients	Returns a vector of exponentially weighted moving average (EWMA) coefficients of length n.
	Filter(DoubleVector)	Filters Copy data with the PadWithZeros option. (Overrides CorrelationFilterFilter(DoubleVector).)
	Filter(DoubleVector, DoubleVector)	Filters Copy data with the PadWithZeros option. (Overrides CorrelationFilterFilter(DoubleVector, DoubleVector).)
	Filter(DoubleVector, MovingWindowFilterBoundaryOption)	Applies the filter to the given data using the given boundary option. The given boundary option sets the current boundary option.
	Filter(DoubleVector, MovingWindowFilterBoundaryOption, DoubleVector)	Applies the filter to the given data using the given boundary option and places the output in a given vector. The given boundary options sets the current boundary option.
	MovingAverageCoefficients	Constructs the coefficient vector that implements a moving average filter when used with the MovingWindowFilter class.
	SavitzkyGolayCoefficients	Constructs the coefficient vector that implements a Savitzky-Golay smoothing filter when used with the MovingWindowFilter class. The algorithm is also known by the terms, least-squares, or DIgital Smoothing POlynomial (DISPO). The filter coefficients, c(n) are chosen so as to approximate the underlying function in the window [i - nL, i + nR] with a polynomial, typically quadratic or quartic, and replace the point f(i) with the value of the approximating polynomial at i. The polynomial is fit using a least squares algorithm.
	SetFilterParameters	Sets the parameters for this filter.

Top

Fields

	Name	Description
	coefficients_	Filtering coefficients for correlation operation. (Inherited from CorrelationFilter.)
	numberLeft_	The number of points to the left.
	numberRight_	The number of points to the right.
	work_	Provides a working vector for correlation results. (Inherited from CorrelationFilter.)

Top

Remarks

The filter replaces data points f(i) with a linear combination g(i) of itself and some number of neighboring points. The neighboring points are determined by by parameters nL, the number of points to the left, and nR, the number of points to the right: g(i) = sum(n = -nL, n = nR)c(n)*f(i+n) See also

Copy

SavitzkyGolayFilter

for more boundary options and Savitzky-Golay derivatives.

Reference

CenterSpace.NMath.Core Namespace