This is an image and post i found that describes how the Advario contextual targeting engine reads the density of the page and then targets the red areas for the most relevance.
Anjo Anjewierden posts the image to the left which described the topics found in his blog. The topics are mined automatically and the image essentially represents a graph of the topics where proximity describes relatedness between topics. The colours indicate how strongly related neighbouring terms. Thus, one could interpret this map as a presentation of the topical coherence in the blog. As red indicates strong relatedness, the fewer and larger the red areas are, the more focused the content of the blog.
To really understand the potential value of this type of visualization, it would be great to see a grid of, say, 100 blogs so that one could see if there were any clear patterns. One might see, for example, highly focused blogs as a single, large red blob with a fringe of green. Other blogs which are less topically focused might look like the example above.
Digg this • Add to del.icio.us • Email this
September 16, 2006 in blogosphere, text mining Permalink Comments (0) TrackBack (0)
To really understand the potential value of this type of visualization, it would be great to see a grid of, say, 100 blogs so that one could see if there were any clear patterns. One might see, for example, highly focused blogs as a single, large red blob with a fringe of green. Other blogs which are less topically focused might look like the example above.
Digg this • Add to del.icio.us • Email this
September 16, 2006 in blogosphere, text mining Permalink Comments (0) TrackBack (0)
No comments:
Post a Comment