hmmm nice code guys,the idea of grouping based on the most popular 2/3 word phrases is definately a break through here. i've been doing a lot of work recently with the google keyword tool api and my initial reaction to running a few lists of words through these groupers is that for a production model (erect) i'd add the ability to re scrape keywords for any 2 word phrases that have less than 'n' number of keywords grouped. it would be like running a spider and could work well if the ability to re filter returned lists based on the keyphrase was taken in to account and a recursion factor of 1 was used. as the G api is currently priced at 200 pulls for a $1 (!!) you could make some monster targetted lists this way. great for the content network.
i'll have a go at writing a grouper myself later today as i have some ideas about varying the percentage variable on recurring filter passes and a few other ideas but i want to test a theory first.
i'll have a go at writing a grouper myself later today as i have some ideas about varying the percentage variable on recurring filter passes and a few other ideas but i want to test a theory first.