heatFilter {sccore}R Documentation

Graph filter with the heat kernel: f(x)=exp(βx/λmab)f(x) = exp(-\beta |x / \lambda_m - a|^b)

Description

Graph filter with the heat kernel: f(x)=exp(βx/λmab)f(x) = exp(-\beta |x / \lambda_m - a|^b)

Usage

heatFilter(x, l.max, order = 1, offset = 0, beta = 30)

Arguments

x

numeric Values to be filtered. Normally, these are graph laplacian engenvalues.

l.max

numeric Maximum eigenvalue on the graph (λm\lambda_m in the equation)

order

numeric Parameter bb in the equation. Larger values correspond to the sharper kernel form (default=1). The values should be positive.

offset

numeric Mean kernel value (aa in the equation), must be in [0:1] (default=0)

beta

numeric Parameter β\beta in the equation. Larger values provide stronger smoothing. β=0\beta=0 corresponds to no smoothing (default=30).

Value

smoothed values for 'x'

See Also

Other graph smoothing: computeChebyshevCoeffs(), smoothChebyshev(), smoothSignalOnGraph()


[Package sccore version 1.0.5 Index]