<< Chapter < Page | Chapter >> Page > |
Letting denote the worst-case density, out minimax procedure results in the following densities for each model in thelikelihood ratio test. The constants and determine the rate of decay of the exponential tails of these worst-case distributions. Their specific values havenot yet been determined, but since they are not needed to compute the likelihood ratio, we don't need them. The constants and are required so that a unit-area density results. The likelihood ratio for each observation in the robust model evaluationproblem becomes
The evaluation of the likelihood ratio depends entirely on determining values for and . The ratios and are easily found; in the tails, the value of thelikelihood ration equals that at the edges of the central region for continuous densities. At the left boundary, for example, the distribution functions must satisfy . In terms of the nominal densities, we have This equation also applies the value right edge . Thus, for a given value of , the integral of the difference between the nominal densities should equal the ratio for two values. illustrates this effect for a Gaussian example. The bi-valued nature of thisintegral may not be valid for some values of ; the value chosen for can be too large, making it impossible to distinguish the models! This unfortunatecircumstance means that the uncertainties, as described by the value of , swamp the characteristics that distinguish the models. Thus, the modelsmust be made more precise (more must be known about the data) so that smaller deviations from the nominal models can describe theobservations.
Returning to the likelihood ratio, the "robust" decision rule consists of computing a clipped function of each observed value, multiplying them together, and comparingthe product computed over the observations with a threshold value. We assume that the nominal distributions of each of the observations are equal; the values of the boundaries and then do not depend on the observation index in this case. More simply, evaluating the logarithm of the quantities involved results inthe decision rule where the function is the clipping function given by If the observations were not identically distributed, then the clipping function would depend on the observation index.
Determining the threshold
that meets a specific performance criterion is difficult in the
context of robust model evaluation. By the very nature of theproblem formulation, some degree of uncertainty in the
Let's consider the Gaussian model evaluation problem we have been using so extensively. The individual observations arestatistically independent and identically distributed with variance five: . For model , the mean is zero; for , the mean is one. These nominal densities describe our best models for the observations, but we seek to allowslight deviations (10%) from them. The equation to be solved for the boundaries is the implicit equation The quantity on the left side of the equation is shown in . If the uncertainty in the Gaussian model, as expressed by the parameter , is larger than 0.15 (for the example values of and ), no solution exists. Assuming that equals 0.1, the quantity and the clipping thresholds are and . Between these values, the clipping function is given by the logarithm of the likelihood ratio, which is givenby .
We can decompose the clipping operation into a cascade of two operations: a linear scaling and shifting (as described by theprevious expression) followed by a clipper having unit slope (see ). Let denote the result of the scaling and shifting operation. This quantity has mean and variance under and the opposite signed mean and the same variance under . The threshold values of the unit-clipping function are thus given by the solution of the equation By substituting for in this equation, we find that the two solutions are negatives of each other. We have now placed the unit-clipper'sthreshold values symmetrically about the origin; however, they do depend on the value of the mean . In this example, the threshold is numerically given by . The expected value of the result of the clipping function with respect to the worst-case density is given by thecomplicated expression The variance is found in a similar fashion and can be used to find the threshold on the sum of clipped observation values.
Notification Switch
Would you like to follow the 'Statistical signal processing' conversation and receive update notifications?