Problems of Forensic Sciences 2007 Vol. 69 (LXIX) 62-70
INFORMATION-THEORETICAL EVALUATION OF LIKELIHOOD RATIOS
Daniel RAMOS, Joaquin GONZALEZ-RODRIGUEZ
ATVS – Biometric Recognition Group, Escuela Politecnica Superior, Universidad Autonoma de Madrid, Madrid, Spain
Streszczenie
In this paper, we present an evaluation methodology for any likelihood-ratio LR-based forensic discipline. The proposed method is based on information theory, showing the performance of the system in the form of cross-entropy as a measure of uncertainty and deviation between probability distributions. This assessment methodology takes into account not only the discrimination of the technique in use, but also the calibration of LR values. In order to illustrate the proposed method, an example is shown using an LR-based speaker recognition system, simulating a blind forensic test following the NIST Speaker Recognition Evaluation 2006 protocol.
Słowa kluczowe
Cross-entropy; Information theory; Likelihood ratio; Assessment.