Profile cover photo
Profile photo
Michael Kuhn
221 followers
221 followers
About
Michael's posts

Post has attachment
I'm still not sure coding all of this is faster than going back to Illustrator after each data update!

Post has attachment
(in case anyone is following on G+...)

Post has attachment
turns out that the standard R randomForest package is rather inefficient, so I could squeeze out lots of RAM and speed savings for predicting on genotype matrices

Post has shared content
PeerJ Pricing Plans have just been accounced #openaccess  

Post has shared content

Post has shared content

Some time ago there was a paper (in Science?) showing that when fitting data, very different parameter combinations can give very similar fitting error. Therefore, the "best fit" might have some extreme parameters, while a slightly worse fit might have more believable parameters. Does anyone remember the reference?

Post has attachment

Post has shared content
Google is really on a mission to increase the amount of clutter and reduce the amount of content. First Google Reader, now Google+. Actual posts get perhaps 40% of my browser window. Argh!
Since a picture is worth 1,000 words:

#TheNewGoogle+ #newgoogleplus
Photo

Post has shared content
took me 2 min... but I have a 1st grader and a pre-school kid at home :)
I'll post the answer in the comments within a few days. Don't spoil it for everyone else!
Photo
Wait while more posts are being loaded