If you know a beer you like and want some recommendations for a style of beer to try, check out the yhat Beer Recommender:
This neat little app is the product of a recommendation system built using the R language by the folks behind the yhat blog. It's based on about 1.5 million beer reviews from the Beer Advocate. The ratings were read into a Postgres database and then imported to R using the RPostgreSQL package. From there, you can easily make a start at collaborative filtering in R by using the intersect function to identify all reviewers that have rated a particular pair of beers, and write a distance function in R to measure how close the two beers are according to the reviewers' ratings. Do this for all pairs of beers (using a neat application of the expand.grid function), put the results behind a JavaScript UI, and you've got yourself a beer recommender. (This is similar to the approach used for this recommendation engine for movies.) Check out the full details including R code at the link below.
yhat blog: Recommendation System in R
Hey I just saw you're post. I really liked it. I saw you put it on a heroku server? Can you say some world how did you implement the R server ?
Posted by: Laszlo Zsurzsa | June 29, 2013 at 13:03