Access

You are not currently logged in.

Access your personal account or get JSTOR access through your library or other institution:

login

Log in to your personal account or through your institution.

If you need an accessible version of this item please contact JSTOR User Support

Universal Prediction of Random Binary Sequences in a Noisy Environment

Tsachy Weissman and Neri Merhav
The Annals of Applied Probability
Vol. 14, No. 1 (Feb., 2004), pp. 54-89
Stable URL: http://www.jstor.org/stable/4140490
Page Count: 36
  • Read Online (Free)
  • Download ($19.00)
  • Cite this Item
If you need an accessible version of this item please contact JSTOR User Support
Universal Prediction of Random Binary Sequences in a Noisy Environment
Preview not available

Abstract

Let $X = {(X_t, Y_t)}_{t\in Z}$ be a stationary time series where Xt is binary valued and Yt, the noisy observation of Xt, is real valued. Letting P denote the probability measure governing the joint process $\lbrace(X_t, Y_t)\rbrace$, we characterize U(l, P), the optimal asymptotic average performance of a predictor allowed to base its prediction for Xt on Y1, ..., $Y_{t-1}$, where performance is evaluated using the loss function l. It is shown that the stationarity and ergodicity of P, combined with an additional "conditional mixing" condition, suffice to establish U(l, P) as the fundamental limit for the almost sure asymptotic performance. U(l, P) can thus be thought of as a generalized notion of the Shannon entropy, which can capture the sensitivity of the underlying clean sequence to noise. For the case where $X = {X_t}$ is governed by P and Yt given by $Y_t = g(X_t, N_t)$ where g is any deterministic function and $N = {N_t}$, the noise, is any i.i.d. process independent of X (namely, the case where the "clean" process X is passed through a fixed memoryless channel), it is shown that, analogously to the noiseless case, there exist universal predictors which do not depend on P yet attain U(l, P). Furthermore, it is shown that in some special cases of interest [e.g., the binary symmetric channel (BSC) and the absolute loss function], there exist twofold universal predictors which do not depend on the noise distribution either. The existence of such universal predictors is established by means of an explicit construction which builds on recent advances in the theory of prediction of individual sequences in the presence of noise.

Page Thumbnails

  • Thumbnail: Page 
54
    54
  • Thumbnail: Page 
55
    55
  • Thumbnail: Page 
56
    56
  • Thumbnail: Page 
57
    57
  • Thumbnail: Page 
58
    58
  • Thumbnail: Page 
59
    59
  • Thumbnail: Page 
60
    60
  • Thumbnail: Page 
61
    61
  • Thumbnail: Page 
62
    62
  • Thumbnail: Page 
63
    63
  • Thumbnail: Page 
64
    64
  • Thumbnail: Page 
65
    65
  • Thumbnail: Page 
66
    66
  • Thumbnail: Page 
67
    67
  • Thumbnail: Page 
68
    68
  • Thumbnail: Page 
69
    69
  • Thumbnail: Page 
70
    70
  • Thumbnail: Page 
71
    71
  • Thumbnail: Page 
72
    72
  • Thumbnail: Page 
73
    73
  • Thumbnail: Page 
74
    74
  • Thumbnail: Page 
75
    75
  • Thumbnail: Page 
76
    76
  • Thumbnail: Page 
77
    77
  • Thumbnail: Page 
78
    78
  • Thumbnail: Page 
79
    79
  • Thumbnail: Page 
80
    80
  • Thumbnail: Page 
81
    81
  • Thumbnail: Page 
82
    82
  • Thumbnail: Page 
83
    83
  • Thumbnail: Page 
84
    84
  • Thumbnail: Page 
85
    85
  • Thumbnail: Page 
86
    86
  • Thumbnail: Page 
87
    87
  • Thumbnail: Page 
88
    88
  • Thumbnail: Page 
89
    89