bestlevt

Best level tree wavelet packet analysis

collapse all in page

    Syntax

    B = bestlevt(T)
    [B,E] = bestlevt(T)

    Description

    bestlevt is a one- or two-dimensional wavelet packet analysis function that computes the optimal complete subtree of an initial tree with respect to an entropy type criterion. The resulting complete tree may be of much smaller depth than the initial tree.

    B = bestlevt(T) computes the wavelet packet tree corresponding to the best level tree decomposition.

    example

    [B,E] = bestlevt(T) returns the best entropy value, E.

    Examples

    collapse all

    Open Live Script

    Save the current extension mode, and then change to the zero-padding extension mode. Then load the noisy Doppler signal.

    origMode = dwtmode("status","nodisp");
dwtmode("zpd","nodisp")
load noisdopp

    Obtain the wavelet packet tree down to level 3 using the Haar wavelet and the default (Shannon) entropy.

    wpt = wpdec(noisdopp,3,"haar");

    Decompose the packet (3,0) and plot the result.

    wpt = wpsplt(wpt,[3 0]);
plot(wpt)

    Figure contains 2 axes objects and other objects of type uimenu. Axes object 1 with title Tree Decomposition contains 33 objects of type line, text. Axes object 2 with title data for node: 0 or (0,0). contains an object of type line.

    Compute the best level tree and plot the result.

    blt = bestlevt(wpt);
plot(blt)

    Figure contains 2 axes objects and other objects of type uimenu. Axes object 1 with title Tree Decomposition contains 29 objects of type line, text. Axes object 2 with title data for node: 0 or (0,0). contains an object of type line.

    Restore the DWT extension mode to the original setting.

    dwtmode(origMode,"nodisp")

    Input Arguments

    collapse all

    Wavelet packet tree, specified as a wptree object.

    Output Arguments

    collapse all

    Best level tree, returned as a wptree object. B may be much smaller than T.

    Optimal entropy of the node, returned as a vector. The optimal entropy of the node, whose index is j-1, is E(j).

    Algorithms

    See besttree algorithms section. The only difference is that the optimal tree is searched among the complete subtrees of the initial tree, instead of among all the binary subtrees.

    Version History

    Introduced before R2006a

    See Also

    | | |