Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. It only takes a minute to sign up.
Sign up to join this community
I have to build a recommendation engine that will cluster users by their preferences.
For example: user that looks for yellow sport GM car should get recommendations for other yellow sports cars. But in my real life data I have about 20 preferences most of them are strings and some are numeric. What algorithm / package in R would be best to use in such a task?
You will need to combine expertise and software for two separate fields: text mining and recommendation systems. Both fields are challenging enough by themselves. I'd recommend getting familiar with them separately, rather than immediately jumping to soliciting software recommendations.
Probably best to run a couple of web searches to find the software - and read tutorials to get the underlying ideas of the two application areas. Unfortunately, there are no relevant task views at CRAN.
For text mining in R, I'd recommend the tm package.
There does not seem to be a dedicated R package to implement recommendation systems. However, a web search turns up a number of good-looking tutorials, like this one and this one and this one.
After reading about the subject I think that using Jaccard method from the recommendlab package will be a good fit.The reason is that Jaccard method uses binary data so it will be a good fit for factors as well as numeric values.
