25.08.2015 Views

In the Beginning was Information

6KezkB

6KezkB

SHOW MORE
SHOW LESS
  • No tags were found...

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Figure 31: Model of a discrete source for generating sequences of symbols.The source has a supply of N different symbols (e. g. an alphabet with 26 letters),of which a long sequence of n symbols is transmitted one after <strong>the</strong> o<strong>the</strong>r at a certaintime. The source could be a symbol generator which releases randomsequences of symbols according to a given probability distribution, or it could bean unknown text stored on magnetic tape which is transmitted sequentially (i. e.one symbol at a time).The first requirement (i) I(p) = I(p 1 x p 2 ) = I(p 1 ) + I(p 2 ) is met ma<strong>the</strong>maticallywhen <strong>the</strong> logarithm of equation (1) is taken. The second requirement (ii) issatisfied when p 1 and p 2 are replaced by <strong>the</strong>ir reciprocals 1/p 1 and 1/p 2 :I (p 1 x p 2 ) = log(1/p 1 ) + log(1/p 2 ). (2)As yet <strong>the</strong> base b of <strong>the</strong> logarithms in equation (2) entails <strong>the</strong> question ofmeasure and is established by <strong>the</strong> third requirement (iii):I = log b (1/p) = log b (1/0.5) = log b 2 = 1 bit (3)It follows from log b 2 = 1 that <strong>the</strong> base b = 2 (so we may regard it as abinary logarithm, as notation we use log 2 = lb; giving lb x = (log x)/(log2); log x means <strong>the</strong> common logarithm that employ <strong>the</strong> base 10: log x =log 10 x). We can now deduce that <strong>the</strong> definition of <strong>the</strong> information contentI of one single symbol with probability p of appearing, is172I(p) = lb(1/p) = - lb p ≥ 0. (4)

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!