dear All,

I've the following problem:

I have a data.frame in which some of the columns are numerical values and some are categorical. Just for data exploration I would like to do sort of a clustering or a correlation matrix on the columns to see which of them are actually correlated. But how can I correlate between numeric and continuous values?

Thanks in advance for any hints and suggestions on that!

cheers, jo

Some thoughts:

• Consider transforming the categorical variables to numeric if that makes sense.
• Consider transforming the numeric variables to categorical and then using an appropriate distance metric (this may take some thought).

A really interesting idea is to use a non-linear measure such as random forest proximities for clustering.  Steve Horvath discusses the idea here:

http://www.slideshare.net/Pammy98/using-random-forest-proximity-for-unsupervised-learning-in-tissue