I’m currently working on a problem where I need to sample from a very specific custom pdf. It has taken me a lot of time to get this pdf written down but now that I finally have it, I’m just missing one key ingredient… the selection function! This is when you have some underlying data…
I learned early on in my PhD that binning is not a good idea. The loss of information from binning will lead to biased results. Here I demonstrate this with a simple linear regression toy model. #generate toy data x = runif(nc, 13,15) y = 22+1.5*x+rnorm(nc, 0,0.5) #bin in y variable nbin = 10 ybins…
31.10.2017 The first ever dark matter day… Why the particle physicists decided to it should coincide with halloween I have no clue. Clearly the trending tags will be dominated by the later occasion right? Anyway I made a video about dark matter. I hope you enjoy it!
Okay so a few weeks ago at work a colleague asked me where I was from in the UK, and I replied “Oh, it’s a small little town, you’ve probably never heard of it, Coventry. It’s close to Birmingham”. This is a natural response for me given that I lived in the US for a…
I attended this conference in Aix-en-Provence this week. I gave a talk on my hierarchical analysis of galaxy clusters and there was a great turn out from the cluster community but lots of cluster environmentalists! Anyway some of the most interesting talks I found were: Michael Gregg from UC Davis, gave a really nice talk…
Relative entropy is a really cool tool that which haven’t heard of before you are in for a treat! Also known as Kullback-Leibler divergence, it is the measure of the distance between 2 probability distributions [e.g. p(x) and q(x) ] on a random variable X. It is defined as: Where H(P,Q) is the cross entropy…
Michael Betancourt recently wrote a nice case study describing the problems often encountered with gaussian mixture models, specifically the estimation of parameters of a mixture model and identifiability i.e. the problem with labelling mixtures (http://mc-stan.org/documentation/case-studies/identifying_mixture_models.html). Also there has been suggestions that GMM’s can’t be easily done in Stan. I’ve found various examples online of simple…
