Hypervector Analysis of the Voynich Manuscript

[dvallis]

Let's not distract from the topic of this thread.

The analysis is promising, but I have considerable difficulty in understand what exactly is being done.
I checked the link to the paper of Pentti Kanerva, but this is only concerned with the hypervectors, not the text analysis.

This piece of text:

Assign random hypervectors to the twenty six letters of the alphabet and a space character. Parse any large text sample with a sliding three letter window, building a series of trigrams. Combining the trigrams yields a hyperdimensional vector signature for the text sample.

is especially difficult to follow. It does not allow one to understand what the author has done exactly.

Yeah the math is hard to follow, so I wrote the first piece in the spirit of "Here's what I found". I thought most people would not want to slog through all the math. Except  you guys LoL.

Now I'll definitely write the next piece with all the nuts and bolts of hypervector language analysis. Its largely based on an example Kanerva proposed in one of his lectures. Took a lot of code and head scratching to implement. 

I knew the algorithm was right when language families began to show up with no input from me as I fed in the data.