UseR! 2013 Highlights
The conference was excellent this year. My highlights:
Bojan Mihaljevic gave a great presentation on machine learning models built from network models. Their package isn’t on CRAN yet, but I’m really looking forward to it.
Jim Harner’s presentation on K-NN models with feature selection was also very interesting, especially the computational aspects of training the model in parallel.
Tal Galili gave a pretty interesting summary of models on tRNA seqeunces. This was a great case study in cross-validation. A simple CV loop was compared with one that used the two Archaea families. The results were very different.
The first session on Computational Challenges in Molecular Biology showed a set of applications of modeling in biology. For me, the presentations of Thomas Poulsen and Insa Winzenborg really stood out (in a positive way).
Hadley Wickham’s talk on big data and R was excellent. His new dplyr package should solve a lot of problems for me. I find his work to be excellent despite one member of R Core remarking to me that
caretshould not depend onplyrsince it was considered by them to be “unreliable”. I feel that nothing is farther form the truth.Another RStudio-ista, Joe Cheng, gave yet another great presentation on shiny.
Finally, I really enjoy the lightning talks. The presentation was setup so that the speaker could have 15 slides and each slide will be shown for 20 seconds. It is sort of a combination of a talk and a game of boggle.
I’m hoping that the presentations will be posted on the website.
Albacete was very nice, as was the food and beer. Next year, the conference will be in Los Angeles. It should be a lot of fun.
(This article was originally posted at http://appliedpredictivemodeling.com)