Menu
Home
People
Places
Arts
History
Plants & Animals
Science
Life & Culture
Technology
Reference.org
Kernel density estimation
open-in-new
<p>In <a href="/facts/Statistics/DNPsKGYU">statistics</a>, kernel density estimation (KDE) is the application of <a href="/facts/Kernel_smoothing/JaGaXr6s">kernel smoothing</a> for <a href="/facts/Probability_density_estimation/NmznyEW8">probability density estimation</a>, i.e., a <a href="/facts/Non-parametric_statistics/svTDJHTb">non-parametric</a> method to <a href="/facts/Estimation/dsztTYcy">estimate</a> the <a href="/facts/Probability_density_function/zvfybna4">probability density function</a> of a <a href="/facts/Random_variable/TwTBXnLT">random variable</a> based on <i><a href="/facts/Kernel_(statistics)/otWLlBbA">kernels</a></i> as <a href="/facts/Weight_function/YNJL7o99">weights</a>. KDE answers a fundamental data smoothing problem where inferences about the <a href="/facts/Statistical_population/KAILs8E3">population</a> are made based on a finite data <a href="/facts/Statistical_sample/URCwKNge">sample</a>. In some fields such as <a href="/facts/Signal_processing/Npaja6zb">signal processing</a> and <a href="/facts/Econometrics/7U52gzHt">econometrics</a> it is also termed the Parzen–Rosenblatt window method, after <a href="/facts/Emanuel_Parzen/QMNbEJ0j">Emanuel Parzen</a> and <a href="/facts/Murray_Rosenblatt/xY83C0a6">Murray Rosenblatt</a>, who are usually credited with independently creating it in its current form. One of the famous applications of kernel density estimation is in estimating the class-conditional <a href="/facts/Marginal_distribution/U9XBWAd1">marginal densities</a> of data when using a <a href="/facts/Naive_Bayes_classifier/FmNHI688">naive Bayes classifier</a>, which can improve its prediction accuracy. </p>