[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: LSC sums for monkey texts
Rene, thanks for the message. Sorry, I have to leave in a few minutes,
but tomorrow the first thing in the morning I'll email to you some texts
you requested. At first glance, LSC test you conducted very clearly
shows that your monkey texts have a quite different distribution of
letter frequency variability as compared to meaningful texts, but are
similar in some respects to the gibberish I explored. VMs on the other
hand has LSC curves precisely like those for meaningful texts. It looks
like LSC can serve as a sharp tool, do you agree? Best, Mark
> Dear all,
> I have indeed done a few initial tests with the LSC technique on
> texts generated by Jacques' monkey program.
> Tentatively, I would say that the technique still sees a difference
> between real meaningful text and a 3rd or 4th order character monkey.
> Not very conclusive but quite promising.
> I used text length of only 20,000 characters, which is not enough
> to be absolutely sure of the conclusion.
> See a quick summary of what I did at:
> There are some plots in which the X-scale has numbers from 1 to 19.
> These are 'codes' and represent the following actual values:
> 1, 2, 3, 5, 7, 10, 15, 20, 30, 50, 70, 100, 150, 200, 300, etc.
> I should rewrite the code in C so that I can run it at home.
> And I would also appreciate if Mark could send me one of his sample
> texts so that I can validate the results of my program.
> One more comment: spaces were removed from all source texts. With the
> spaces included as an additional 'character', the sums change
> Comments are welcome.
> Cheers, Rene