Comments on: Provocative Corollary to Andrew Gelman’s Folk Theorem http://hea-www.harvard.edu/AstroStat/slog/2007/provocative-corollary-to-october-3-2007-quote-of-the-week/ Weaving together Astronomy+Statistics+Computer Science+Engineering+Intrumentation, far beyond the growing borders Fri, 01 Jun 2012 18:47:52 +0000 hourly 1 http://wordpress.org/?v=3.4 By: hlee http://hea-www.harvard.edu/AstroStat/slog/2007/provocative-corollary-to-october-3-2007-quote-of-the-week/comment-page-1/#comment-110 hlee Thu, 04 Oct 2007 01:07:23 +0000 http://hea-www.harvard.edu/AstroStat/slog/2007/provocative-corollary-to-october-3-2007-quote-of-the-week/#comment-110 <p>The first sentence, <em>Model, for statisticians, has two meanings.</em> made me reflect challenges that I had at the beginning of my graduate studies in statistics, eventually not study (generalized) linear models (GLM), and dislike adopting kernel density estimation (KDE) in astrophysics, although statistically/mathematically GLM and KDE are beautiful themselves. After getting used to statistics, it seems I forgot the time of being a cocoon. Convincing the other what "model" means and its suitability has been the most long lasting obstacle during collaboration. </p> <p>Regarding MaxEnt, although not knowing what it is, the recent <a href="http://arxiv.org/abs/0709.4079" rel="nofollow">[arxiv/stat.MT:0709.4079]</a> discussed it. I was unable to comment it. Someday, would you introduce what MaxEnt is? (I guess the acronym came from Maximum Entropy).</p> The first sentence, Model, for statisticians, has two meanings. made me reflect challenges that I had at the beginning of my graduate studies in statistics, eventually not study (generalized) linear models (GLM), and dislike adopting kernel density estimation (KDE) in astrophysics, although statistically/mathematically GLM and KDE are beautiful themselves. After getting used to statistics, it seems I forgot the time of being a cocoon. Convincing the other what “model” means and its suitability has been the most long lasting obstacle during collaboration.

Regarding MaxEnt, although not knowing what it is, the recent [arxiv/stat.MT:0709.4079] discussed it. I was unable to comment it. Someday, would you introduce what MaxEnt is? (I guess the acronym came from Maximum Entropy).

]]>
By: vlk http://hea-www.harvard.edu/AstroStat/slog/2007/provocative-corollary-to-october-3-2007-quote-of-the-week/comment-page-1/#comment-108 vlk Wed, 03 Oct 2007 23:56:43 +0000 http://hea-www.harvard.edu/AstroStat/slog/2007/provocative-corollary-to-october-3-2007-quote-of-the-week/#comment-108 But how about MaxEnt? The models there (in both senses) are usually very well set up, but getting error bars on, say, an image inverted from fourier components is well nigh impossible. Wouldn't you say that that has more to do with what it means to say "error bar" for an image rather than a problem with the model? But how about MaxEnt? The models there (in both senses) are usually very well set up, but getting error bars on, say, an image inverted from fourier components is well nigh impossible. Wouldn’t you say that that has more to do with what it means to say “error bar” for an image rather than a problem with the model?

]]>