And to define redundancy of the transmitted data we will use Shannon's theorem. Under a condition if Shannon's theorem will be carried out, redundancy κ will equal 0, information means is transferred without loss. If is not present, then κ there will be more than zero (κ>. I.e. the less size κ, the less will be probability of an error of decoding.
According to property of mutual information 2 it is possible to write down: Ssim=max (H (B) - H(B|A)). Let's paint H(B|A). Proceeding from statements of the problem probability of the correct transfer of a symbol on the channel - 1-p, and probability of wrong transfer of one p / (1-m) symbol, where m - number of various symbols which are transferred on the channel. Total of right transfers - m; total of wrong transitions - m * (m-. From this it follows that:
We draw a conclusion that the sense of the theorem of Shannon is that at H’(A)> With faultless transmission of messages on this channel if H’(A)