Comments on: Dance of the Errors http://hea-www.harvard.edu/AstroStat/slog/2008/errordance/ 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: vlk http://hea-www.harvard.edu/AstroStat/slog/2008/errordance/comment-page-1/#comment-158 vlk Sun, 27 Jan 2008 02:46:25 +0000 http://hea-www.harvard.edu/AstroStat/slog/2008/errordance/#comment-158 Ah, it appears that the figure captions don't show up when you mouse over, as I had expected they would. I have now added the captions explicitly in the text. The short answer is, there is no time dependence -- these are fluxes calculated from the MCMC draws of the DEMs. MCMC is used to draw the parameters that determine the DEM as a function of temperature, which in turn is used to calculate the intensity in each temperature bin. i.e., these <em>are</em> the model uncertainties. And the point is that the uncertainty structure is far richer than is usually represented, as e.g., in the third set of equations above, and even the graphical representation of Figure 1 is insufficient to capture the correlations and magnitudes across both temperatures and lines. Ah, it appears that the figure captions don’t show up when you mouse over, as I had expected they would. I have now added the captions explicitly in the text.

The short answer is, there is no time dependence — these are fluxes calculated from the MCMC draws of the DEMs. MCMC is used to draw the parameters that determine the DEM as a function of temperature, which in turn is used to calculate the intensity in each temperature bin.

i.e., these are the model uncertainties. And the point is that the uncertainty structure is far richer than is usually represented, as e.g., in the third set of equations above, and even the graphical representation of Figure 1 is insufficient to capture the correlations and magnitudes across both temperatures and lines.

]]>
By: hlee http://hea-www.harvard.edu/AstroStat/slog/2008/errordance/comment-page-1/#comment-157 hlee Sat, 26 Jan 2008 22:57:01 +0000 http://hea-www.harvard.edu/AstroStat/slog/2008/errordance/#comment-157 A naive question, here since I don't understand the process. What is the time related variable in the animation? Or is it realization of multiple simulations? I wonder if the point is that conditioning (temperature) affects and want to find a way of including the model uncertainty in the final estimates. A naive question, here since I don’t understand the process. What is the time related variable in the animation? Or is it realization of multiple simulations? I wonder if the point is that conditioning (temperature) affects and want to find a way of including the model uncertainty in the final estimates.

]]>