DiscussMormonism.com

MCB said: Your assumption of open set, sets up your means of analysis. True, but I don’t see how ‘open-set’ is an assumption. ‘Closed set’ is an assumption, and a very strong one at that. But ‘open-set’ is simply doing away with that assumption. And from work with the Hamilton texts, it appears that...

I'm not yet persuaded of that. Cross-genre tests still seem to be an intractable problem for the method.

I don't know about intractable, but I agree that it's a difficult problem. With this paper published, I hope to look at that problem.

Scholarly debate does not work like that. There will be a response, addressing those issues. I hope there will be a response. As Ben mentioned, the Jockers/Criddle paper did a lot of good. For me, itreawakened my dormant interest in statistical authorship attribution and generated some really inter...

Bruce wrote: Quote: The principal component plot convinces me that the true author is not in the training set. Chris, do you agree with that? Now that I see what I said, I’m not sure I agree with myself. What I meant is that the PCA plot convinces me that based on Criddle’s assumptions the true aut...

First, I think you did a very nice job correcting for some, but not all, of the flaws of the Jockers study. In particular, I don't think you have gone far enough to filter out the potentially biased word vectors. As a very simple example, not all the "came to passes" in the Book of Mormon...

When you tested the Extended method (page 19) on the later (post-LDS) Rigdon texts, 'naïve' NSC attributed most texts to Rigdon, and a sizable minority to Smith (based on your Smith corpus) - you don't give the figures, but I count about 25% attributed to Smith. Your Extended NSC attributes less th...

Chris, Does "centroid" in this context refer to the "center" or "mean" of each candidate author's 110-dimensional cloud? (If we could visualize such a thing?) Yes - it's the 'shrunken mean' of each candidates author's 110-dimensional cloud. I think you've got it right. ...

Danna, My understanding is that the latent author is based on the test text not the training (candidate) texts, and texts must fall outside of the bounds of the Latent Author to be assigned to a candidate author. So a data point falling within the distribution of both the Latent Author and a Candida...

Roger, You raised some good points. if the various "pc" charts represent "tests" with differing variables and in some of those charts the results are less distinct (or even overlap) why would you say that the "distinctness" of one chart applies to the rest? Or am I not ...

Just for the sake of clarification, I wasn't endorsing the idea that you had backed out of the discussion. Chris, I knew that you had not. Thanks for just bringing this dicussion to my attention. You have a very good idea of what I was saying and have represented the paper well. I doubt that we two...