<?xml version="1.0"?>
<!-- name="generator" content="blosxom/2.0" -->
<!DOCTYPE rss PUBLIC "-//Netscape Communications//DTD RSS 0.91//EN" "http://my.netscape.com/publish/formats/rss-0.91.dtd">

<rss version="0.91">
  <channel>
    <title>Notebooks   </title>
    <link>http://bactra.org/notebooks</link>
    <description>Cosma's Notebooks</description>
    <language>en</language>

  <item>
    <title>Density Estimation</title>
    <link>http://bactra.org/notebooks/2012/03/21#density-estimation</link>
    <description>
&lt;P&gt;Yet Another Inadequate Placeholder, spun off from &lt;a href=&quot;statistics.html&quot;&gt;statistics&lt;/a&gt;.

&lt;P&gt;Two topics of particular interest: estimating &lt;em&gt;conditional&lt;/em&gt;
densities, and estimating the densities of short subsequences from
&lt;a href=&quot;time-series.html&quot;&gt;time series&lt;/a&gt;.

&lt;P&gt;See also:
	&lt;a href=&quot;density-estimation-on-graphs.html&quot;&gt;density estimation on
graphical models&lt;/a&gt;;
	&lt;a href=&quot;independence-and-dependence.html&quot;&gt;independence tests
and dependence measures&lt;/a&gt;

&lt;ul&gt;Recommended:
	&lt;li&gt;Luc Devorye and Gabor Lugosi, &lt;cite&gt;Combinatorial Methods in Density
Estimation&lt;/cite&gt; [&lt;a href=&quot;../weblog/algae-2009-11.html#combinatorial-methods&quot;&gt;comments&lt;/a&gt;]
	&lt;li&gt;Peter Hall, Jeff Racine and Qi Li, &quot;Cross-Validation and the Estimation of Conditional Probability Densities&quot;, &lt;cite&gt;Journal of the American Statistical Association&lt;/cite&gt; &lt;strong&gt;99&lt;/strong&gt; (2004): 1015--1026 [&lt;a href=&quot;http://www.ssc.wisc.edu/~bhansen/workshop/QiLi.pdf&quot;&gt;PDF&lt;/a&gt;]
	&lt;li&gt;Bruce E. Hansen
		&lt;ul&gt;
		&lt;li&gt;&quot;Nonparametric Conditional Density
Estimation&quot; [&lt;a href=&quot;http://www.ssc.wisc.edu/~bhansen/papers/ncde.pdf&quot;&gt;PDF preprint&lt;/a&gt;, 2004]
		&lt;li&gt;&quot;Nonparametric Estimation of Smooth Conditional Distributions&quot; [&lt;a href=&quot;http://www.ssc.wisc.edu/~bhansen/papers/cdf.html&quot;&gt;Preprint&lt;/a&gt;]
		&lt;/ul&gt;
	&lt;li&gt;Jeffrey S. Racine, &quot;Nonparametric Econometrics: A Primer&quot;,
&lt;a href=&quot;http://dx.doi.org/10.1561/0800000009&quot;&gt;&lt;cite&gt;Foundations and Trends in Econometrics&lt;/cite&gt;
&lt;strong&gt;3&lt;/strong&gt; (2008): 1--88&lt;/a&gt; [Good primer of nonparametric techniques
for regression, density estimation and hypothesis testing; next to no economic
content (except for examples).  Presumes reasonable familiarity with parametric
statistics.  &lt;a href=&quot;http://socserv.mcmaster.ca/racine/ECO0301.pdf&quot;&gt;PDF
reprint&lt;/a&gt;]
	&lt;li&gt;Jeffrey S. Simonoff, &lt;cite&gt;Smoothing Methods in Statistics&lt;/cite&gt;
	&lt;li&gt;&lt;a href=&quot;http://www.stat.cmu.edu/~larry/&quot;&gt;Larry Wasserman&lt;/a&gt;
		&lt;ul&gt;
		&lt;li&gt;&lt;cite&gt;All of Statistics&lt;/cite&gt;
		&lt;li&gt;&lt;cite&gt;All of Nonparametric Statistics&lt;/cite&gt;
		&lt;/ul&gt;
	&lt;/ul&gt;

&lt;ul&gt;Modesty forbids me to recommend:
	&lt;li&gt;The lecture on density estimation (currently no. 28) in my &lt;a href=&quot;http://www.stat.cmu.edu/~cshalizi/350/&quot;&gt;data mining
class notes&lt;/a&gt;
	&lt;/ul&gt;

&lt;ul&gt;To read:
	&lt;li&gt;Andrew R. Barron and Chyong-Hwa Sheu, &quot;Approximation of Density
Functions by Sequences of Exponential Families&quot;, &lt;a href=&quot;http://projecteuclid.org/euclid.aos/1176348252&quot;&gt;&lt;cite&gt;Annals
of Statistics&lt;/cite&gt; &lt;strong&gt;19&lt;/strong&gt; (1991): 1347--1369&lt;/a&gt;
	&lt;li&gt;Alain Berlinet, G&amp;eacute;rard Biau and Laurent Rouvi&amp;egrave;re,
&quot;Optimal L1 Bandwidth selection for variable kernel density estimates&quot;,
&lt;a href=&quot;http://dx.doi.org/10.1016/j.spl.2005.04.036&quot;&gt;&lt;cite&gt;Statistics and
Probability Letters&lt;/cite&gt; &lt;strong&gt;74&lt;/strong&gt; (2005): 116--128&lt;/a&gt; [&quot;[O]ne can
improve performance of kernel density estimates by varying the bandwidth with
the location and/or the sample data at hand. Our interest in this paper is in
the data-based selection of a variable bandwidth...  an automatic selection
procedure inspired by the combinatorial tools developed in Devroye and
Lugosi... the expected L1 error of the corresponding selected estimate is up to
a given constant multiple of the best possible error plus an additive term
which tends to zero under mild assumptions&quot;]
	&lt;li&gt;Z. I. Botev, J. F. Grotowski, and D. P. Kroese, &quot;Kernel density estimation via diffusion&quot;, &lt;a href=&quot;http://projecteuclid.org/euclid.aos/1281964340&quot;&gt;&lt;cite&gt;Annals of Statistics&lt;/cite&gt;
&lt;strong&gt;38&lt;/strong&gt; (2010): 2916--2957&lt;/a&gt;
	&lt;li&gt;Serge Cohen, Erwan Le Pennec, &quot;Conditional Density Estimation by Penalized Likelihood Model Selection&quot;, &lt;a href=&quot;http://arxiv.org/abs/1103.2021&quot;&gt;arxiv:1103.2021&lt;/a&gt;
	&lt;li&gt;Tilman M. Davies, Martin L. Hazelton, Jonathan. C Marshall, &quot;sparr: Analyzing Spatial Relative Risk Using Fixed and Adaptive Kernel Density Estimation in R&quot;, &lt;a href=&quot;http://www.jstatsoft.org/v39/i01&quot;&gt;&lt;citE&gt;Journal of Statistical Software&lt;/cite&gt; &lt;strong&gt;39:1&lt;/strong&gt; (2011)&lt;/a&gt;
	&lt;li&gt;Sam Efromovich
		&lt;ul&gt;
		&lt;li&gt;&quot;Distribution estimation for biased data&quot;,
&lt;a href=&quot;http://dx.doi.org/10.1016/S0378-3758(03)00202-7&quot;&gt;&lt;cite&gt;Journal of
Statistical Planning and Inference&lt;/cite&gt; &lt;strong&gt;124&lt;/strong&gt; (2004):
1--43&lt;/a&gt;
		&lt;li&gt;&quot;Conditional density estimation in a regression
setting&quot;, &lt;cite&gt;Annals of Statistics&lt;/cite&gt; &lt;strong&gt;35&lt;/strong&gt; (2007): 2504--2535, &lt;a href=&quot;http://arxiv.org/abs/0803.2984&quot;&gt;arxiv:0803.2984&lt;/a&gt;
		&lt;li&gt;&quot;Dimension Reduction and Adaptation in Conditional Density
Estimation&quot;, &lt;a href=&quot;http://dx.doi.org/10.1198/jasa.2010.tm09426&quot;&gt;&lt;cite&gt;Journal
of the American Statistical Association&lt;/cite&gt; &lt;strong&gt;105&lt;/strong&gt; (2010):
761--774&lt;/a&gt;
		&lt;/ul&gt;
	&lt;li&gt;Bradley Efron and Robert Tibshirani, &quot;Using Specially Designed Exponential Families for Density Estimation&quot;, &lt;a href=&quot;http://projecteuclid.org/euclid.aos/1032181161&quot;&gt;&lt;cite&gt;Annals of Statistics&lt;/cite&gt;
&lt;strong&gt;24&lt;/strong&gt; (1996): 2431--2461&lt;/a&gt;
	&lt;li&gt;Evarist Gin&amp;eacute; and Hailin Sang, &quot;Uniform asymptotics for kernel density estimators with variable bandwidths&quot;, &lt;a href=&quot;http://arxiv.org/abs/1007.4350&quot;&gt;arxiv:1007.4350&lt;/a&gt;
	&lt;li&gt;Evarist Gin&amp;eacute; and Richard Nickl
		&lt;ul&gt;
		&lt;li&gt;&quot;Adaptive estimation of a distribution function and its density in sup-norm loss by wavelet and spline projections&quot;, &lt;a href=&quot;http://projecteuclid.org/euclid.bj/1290092899&quot;&gt;&lt;citE&gt;Bernoulli&lt;/cite&gt; &lt;strong&gt;16&lt;/strong&gt; (2010): 1137--1163&lt;/a&gt;
		&lt;li&gt;&quot;Uniform limit theorems for
wavelet density estimators&quot;, &lt;a href=&quot;http://arxiv.org/abs/0805.1406&quot;&gt;arxiv:0805.1406&lt;/a&gt; = &lt;cite&gt;Annals of Probability&lt;/cite&gt; &lt;strong&gt;37&lt;/strong&gt; (2009):
1605--1646
		&lt;li&gt;&quot;Confidence bands in density estimation&quot;,
&lt;a href=&quot;http://projecteuclid.org/euclid.aos/1266586625&quot;&gt;&lt;cite&gt;Annals of Statistics&lt;/cite&gt; &lt;strong&gt;38&lt;/strong&gt; (2010):
1122--1170&lt;/a&gt;
		&lt;/ul&gt;
	&lt;li&gt;Han Liu, John Lafferty and Larry Wasserman, &quot;Tree
Density Estimation&quot;, &lt;a href=&quot;http://arxiv.org/abs/1001.1557&quot;&gt;arxiv:1001.1557&lt;/a&gt;
	&lt;li&gt;Han Liu, Min Xu, Haijie Gu, Anupam Gupta, John Lafferty, Larry Wasserman, &quot;Forest Density Estimation&quot;, &lt;a href=&quot;http://jmlr.csail.mit.edu/papers/v12/liu11a.html&quot;&gt;&lt;cite&gt;Journal of Machine
Learning Research&lt;/cite&gt; &lt;strong&gt;12&lt;/strong&gt; (2011): 907--951&lt;/a&gt;
	&lt;li&gt;Brendan P. M. McCabe, Gael M. Martin, David Harris, &quot;Efficient probabilistic forecasts for counts&quot;, &lt;a href=&quot;http://dx.doi.org/10.1111/j.1467-9868.2010.00762.x&quot;&gt;&lt;cite&gt;Journal
of the Royal Statistical Society&lt;/cite&gt; B &lt;strong&gt;73&lt;/strong&gt; (2011): 253--272&lt;/a&gt;
	&lt;li&gt;Abdelkader Mokkadem, Mariane Pelletier, Yousri Slaoui, &quot;The stochastic approximation method for the estimation of a multivariate probability density&quot;, &lt;a href=&quot;http://arxiv.org/abs/0807.2960&quot;&gt;arxiv:0807.2960&lt;/a&gt;
	&lt;li&gt;Andrew B. Nobel, Gusztav Morvai, Sanjeev R. Kulkarni, &quot;Density estimation from an individual numerical sequence&quot;, &lt;citE&gt;IEEE Transactions on
Information Theory&lt;/cite&gt; &lt;strong&gt;44&lt;/strong&gt; (1998): 537--541, &lt;a href=&quot;http://arxiv.org/abs/0710.2500&quot;&gt;arxiv:0710.2500&lt;/a&gt;
	&lt;li&gt;Andriy Norets, &quot;Approximation of conditional densities by smooth mixtures of regressions&quot;, &lt;cite&gt;Annals of Statistics&lt;/citE&gt; &lt;strong&gt;38&lt;/strong&gt;
(2010): 1733--1766&lt;/a&gt;, &lt;a href=&quot;http://arxiv.org/abs/1010.0581&quot;&gt;arxiv:1010.0581&lt;/a&gt;
	&lt;li&gt;Michael Nussbaum, &quot;Asymptotic Equivalence of Density Estimation and
Gaussian White Noise&quot;, &lt;a href=&quot;http://projecteuclid.org/euclid.aos/1032181160&quot;&gt;&lt;cite&gt;Annals of Statistics&lt;/cite&gt; &lt;strong&gt;24&lt;/strong&gt;
(1996): 2399--2430&lt;/a&gt;
	&lt;li&gt;Bruno Pelletier, &quot;Kernel density estimation on Riemannian
manifolds&quot;, &lt;a
href=&quot;http://dx.doi.org/10.1016/j.spl.2005.04.004&quot;&gt;&lt;cite&gt;Statistics and
Probability Letters&lt;/cite&gt; &lt;strong&gt;73&lt;/strong&gt; (2005): 297--304&lt;/a&gt;
	&lt;li&gt;Alessandro Rinaldo, Aarti Singh, Rebecca Nugent, Larry Wasserman, &quot;Stability of Density-Based Clustering&quot;, &lt;a href=&quot;http://arxiv.org/abs/1011.2771&quot;&gt;arxiv:1011.2771&lt;/a&gt;
	&lt;li&gt;Alessandro Rinaldo and Larry Wasserman, &quot;Generalized Density
Clustering&quot;, &lt;citE&gt;Annals of Statistics&lt;/cite&gt; &lt;strong&gt;38&lt;/strong&gt;
(2010): 2678--2722, &lt;a href=&quot;http://arxiv.org/abs/0907.3454&quot;&gt;arxiv:0907.3454&lt;/a&gt;
	&lt;li&gt;Olga Y. Savchuk, Jeffrey D. Hart, and Simon J. Sheather, &quot;Indirect
Cross-Validation for Density
Estimation&quot;, &lt;a href=&quot;http://dx.doi.org/10.1198/jasa.2010.tm08532&quot;&gt;&lt;cite&gt;Journal
of the American Statistical Association&lt;/cite&gt; &lt;strong&gt;105&lt;/strong&gt; (2010):
415--423&lt;/a&gt;
	&lt;li&gt;Yuefeng Wu, Subhashis Ghosal, &quot;Kullback Leibler property of kernel mixture priors in Bayesian density estimation&quot;, &lt;a href=&quot;http://dx.doi.org/10.1214/07-EJS130&quot;&gt;&lt;cite&gt;Electronic Journal
of Statistics&lt;/cite&gt; &lt;strong&gt;2&lt;/strong&gt; (2008): 298--331&lt;/a&gt;, &lt;a href=&quot;http://arxiv.org/abs/0710.2746&quot;&gt;arxiv:0710.2746&lt;/a&gt;
	&lt;/ul&gt;
</description>
  </item>
  </channel>
</rss>
