I want to classify text by different topics.
However, one of the current problems is that there are several topics/categories that are quite intuitively independent and statistically standalone, but there are also several other topics/categories that are quite intertwined with each other.
In order to make a good decision on the final categories, I realize I'd better carry out some statistical analyses on my raw text collection and explore the statistical nature of the collection.
I was thinking using clustering to have a clearer view of my text collection. But I want to know, what are options that I get in this step to make a better decision on my categories?
Any help and suggestion would be appreciated.
