conkit.misc.bandwidth module

A collection of bandwidth estimators for Kernel Density Estimation

class AmiseBW(data, niterations=25, eps=0.001)

Bases: conkit.misc.bandwidth.BandwidthBase

Asymptotic Mean Integrated Squared Error (AMISE)

This particular choice of bandwidth recovers all the important features whilst maintaining smoothness. It is a direct implementation of the method used by [1].

[1]	Sadowski, M.I. (2013). Prediction of protein domain boundaries from inverse covariances.

bandwidth

class BandwidthBase

Bases: abc.ABC

Abstract class for bandwidth calculations

bandwidth

bw

class BowmanBW(data)

Bases: conkit.misc.bandwidth.BandwidthBase

Bowman & Azzalini [2] bandwidth calculation

\[\sqrt{\frac{\sum{X}^2}{n}-(\frac{\sum{X}}{n})^2}*(\frac{(d+2)*n}{4})^\frac{-1}{d+4}\]

[2]	Bowman, A.W. & Azzalini, A. (1997). Applied Smoothing Techniques for Data Analysis.

bandwidth

class LinearBW(data, threshold=15)

Bases: conkit.misc.bandwidth.BandwidthBase

Linear [3] implementation

\[\frac{N_{max}}{t}\]

[3]	Sadowski, M.I. (2013). Prediction of protein domain boundaries from inverse covariances.

bandwidth

class ScottBW(data)

Bases: conkit.misc.bandwidth.BandwidthBase

Scott’s [4] implementation

\[1.059*\sigma*n^\frac{-1}{d+4}\]

[4]	Scott, D.W. (1992). Multivariate Density Estimation: Theory, Practice, and Visualization.

bandwidth

class SilvermanBW(data)

Bases: conkit.misc.bandwidth.BandwidthBase

Silverman’s [5] implementation

\[0.9*\sigma*(n*\frac{d+2}{4})^\frac{-1}{d+4}\]

[5]	Silverman, B.W. (1986). Density Estimation for Statistics and Data Analysis.

bandwidth

bandwidth_factory(method)

Obtain the bandwidth as defined by user method