Standardized Log-Likelihood Ratio Statistics for Mixtures of Discrete and Continuous Observations

Statistics of the Spectral Ratio and Log-likelihood Ratio Seismic Discriminants

The effect of noise variations on two of the short period seismic discriminants, spectral ratio and log-likelihood ratio is studied. The mean square errors of two different estimators of these two discriminants are evaluated and some theoretical and experimental results for two seismic signals in seismic noise are presented. Accepted for the Air Force Joseph R. Waterman, Lt. Col., USAF Chief, L...

Generalized Likelihood Ratio Statistics

A New Method for Performance Evaluation of Turbo-Codes Using Statistics of the Log Likelihood Ratio

In this paper1, we present a new method for the performance evaluation of Turbo-Codes. The method is based on estimating the Probability Density Function (pdf ) of the bit Log Likelihood Ratio (LLR) by using an exponential model. It is widely known that the pdf of the bit LLR is close to the normal density, and the proposed approach takes advantage of this property to simplify the calculations....

Likelihood Ratio Determination for Stochastic Processes Recognition Problem with Respect to the Set of Continuous and Discrete Memory Observations

This paper considers the problem of likelihood ratio determination for recognition of the stochastic processes with continuous time on the set continuous and discrete time memory observations. The research of memory influence on the detection quality of anomalous noises in the discrete channel observation with applying the general obtained results is realized for the one particular problem.

Adjusting Composite Likelihood Ratio Statistics

Composite likelihood may be useful for approximating likelihood based inference when the full likelihood is too complex to deal with. Stemming from a misspecified model, inference based on composite likelihood requires suitable corrections. Here we focus on the composite likelihood ratio statistic for a multidimensional parameter of interest, and we propose a parameterization invariant adjustme...