tag:blogger.com,1999:blog-3240271627873788873.post2567532668410609986..comments2024-03-26T06:46:11.752-04:00Comments on Doing Bayesian Data Analysis: It's getting warmer in Wisconsin!John K. Kruschkehttp://www.blogger.com/profile/17323153789716653784noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-3240271627873788873.post-48640546977768287132014-12-22T19:17:16.659-05:002014-12-22T19:17:16.659-05:00If I recall correctly, it's easy to estimate t...If I recall correctly, it's easy to estimate the wl parameter by putting a uniform prior on it over a reasonable range. Perhaps something like<br /><br />wl ~ dunif( (365.24219-10)/(2*pi) , (365.24219+10)/(2*pi) )<br /><br />You don't need to put an infinite-support prior on wl (or frequency), just something that's broad on a reasonable range for the data.John K. Kruschkehttps://www.blogger.com/profile/17323153789716653784noreply@blogger.comtag:blogger.com,1999:blog-3240271627873788873.post-17524543298120828272014-12-18T12:01:55.137-05:002014-12-18T12:01:55.137-05:00How would you have modeled the WL parameter if you...How would you have modeled the WL parameter if you wanted to?<br />I have a similar problem with 8 data points, which looks like a M shape, so symmetric around the center. I am trying to fit a "BASELINE+ AMP*cos(2*pi*x*FREQ)" model to the data with BASELINE and AMP parameters dnorm distributed and FREQ as dlnorm as negative frequencies doesn't exist. I have difficulties to make it converge, especially the FREQ.MU and FREQ.TAU parameter. Any help would be appreciated...dronathttps://www.blogger.com/profile/10626412503597280767noreply@blogger.comtag:blogger.com,1999:blog-3240271627873788873.post-17814520530190845342013-09-20T23:03:40.880-04:002013-09-20T23:03:40.880-04:00As both a Bayesian, however empirical my priors mi...As both a Bayesian, however empirical my priors might be, and a scientist with the implemented mission of educating public on climate change, you need to know (1) how refreshing I find your blog and book, and how useful, and (2) how good it is to have a series of problems and puzzlers and questions posed which exercise the Bayesian regimen. My own focus of late has been the predictive underpinnings of Bayesian analysis, namely those identified by Geisser in his work on <i>Predictive Inference</i> and seconded by Christensen, Johnson, Branscum, and Hanson in their <i>Bayesian Ideas and Data Analysis</i>. I think this stuff totally <i>awesome</i> and look forward to seeing your work and that of your students in the ISBA publications and elsewhere. With respect, <br /><br /> -- Jan Galkowski, Akamai Technologies.<br /><br />Anonymousnoreply@blogger.com