Multiple types of keywords ?

I thought that I could have 3 different keyword generators (from 3 different training sets).

So I created 3 different sets and generated the appropriate files, the lexicons vary in size so I figure I am correct. Being lazy I put them all in the DCT to see how they differ, one training set was hotel amenities, one restuartants and one things to do. My shock came when I started tagging various DCRs and all the keywords were the same.

Steps: created foo_summ_idx.cfg and fooKeywords.smf (3 times) used 3 different training sets, generated and changed the DCT. I am digging through the debug logs, but no idea why.

Any tips/pointers/RTFMs appreciated

Andy

Find more posts tagged with

Comments

Migrateduser

Some followup questions:

1) Can you post your smf files?

2) Not sure if you really want to extract summary keywords (words that distinguish content from a collection) or want to extract recognized keywords (words in a taxonomy that occur in the content). My guess is the latter, in which case, you'd probably want to use iwgenclusters or iwgenvocab to build the taxonomy.

nipper

1. Certainly, no problem, it is attached.

2. Yea probably, I was being lazy, assuming that if you provide different training sets I would get different keywords. I thought this way would make the start quicker.

Migrateduser

You'll be better of with three keyword taxonomies that the summary keywords.