Interesting Words Key

Is there an explanation of what we are looking at in the Interesting words? It is definitely interesting, and there are lots of words. But I am really not sure what is what. If there was some control over it like the graph search results, it could start making sense.

Find more posts tagged with

Comments

Kevin Becker

Interesting words is just a pretty graphical representation of what an algorithm has determined is "interesting."

Terry Poperszky

https://community.logos.com/discussion/comment/79507#Comment_79507

Interesting words is just a pretty graphical representation of what an algorithm has determined is "interesting."

I find it more pretty, than interesting.

secateurs

https://community.logos.com/discussion/comment/79512#Comment_79512

Too bad; it has such potential to be a graphical way to show statistics.

MJ. Smith

https://community.logos.com/discussion/comment/384807#Comment_384807

it has such potential to be a graphical way to show statistics.

Most algorithms to build wordles (this kind of tag cloud) do use size to indicate frequency - sometimes a weighted frequency.

DMB

https://community.logos.com/discussion/comment/384822#Comment_384822

This thread is a good illustration of the value of 'documentation' for Logos. If I understand Bob correctly, the word sizes are relative to the 'unlikeliness' of the probability of the words. It's equivalent to someone saying something and your ear catching it, because it's different.

Now, I'd bet what I just wrote (which is my understanding) is completely wrong. That's unfortunate, since I do suspect Bob's approach is unusually good if indeed we knew what it was!

In my heavy-duty pattern matching algorithms, syntactical mix is a decent predictor of an author (i.e. a computer can reproduce the writing with a decently low error level). But the pattern of vocabulary use (not the statistical histogram which is what Logos4 displays) is far, far, more accurate. If the computer has a list of 50 common words used in NT theological arguments, and you give it a couple of random verses (string of 25 words), it can nail the book almost every time. It's amazing.

MJ. Smith

https://community.logos.com/discussion/comment/384866#Comment_384866

the word sizes are relative to the 'unlikeliness' of the probability of the words

For the benefit of non-techies, that would be equivalent to the "weighting" that I referred to.