Comparative Review of Likelihood Ratio (LR) and Probability of Random Correspondence (PRC) as Statistical Tools for Fingerprint Evidence Evaluation

Comparative Review of Likelihood Ratio (LR) and Probability of Random Correspondence (PRC) as Statistical Tools for Fingerprint Evidence Evaluation

Length: 1397 words (4 double-spaced pages)

Rating: Term Papers

Open Document

Essay Preview

The first time fingerprint comparison evidence was used in court against a defendant was in 1892 in Argentina [1]
For purposes of forensic identification in cases of law enforcement and other areas where human identification is needed, fingerprints have been widely acclaimed to be of an invaluable importance and has therefore seen a close to unanimous acceptance as the gold standard of forensic evidence where biometric identity is concerned. Recently however, as was rarely done in times past, the scientific foundations of fingerprint expert testimonies in court are beginning to be challenged [2]. There are some commentators who now query the scientific validity of forensic fingerprint identification. Reference has been made to the Daubert’s standard which stipulates the conditions for acceptability of expert testimony [1], where scientific methods and techniques employed must themselves be based on testable and falsifiable theories that have gone through the peer review process. These methods also must have known and/or predictable error rates and must comply with recognized standards relevant to their application.
These Daubert’s principles, as it has come to be known in the US since 1993, has gone beyond the American jurisdiction to influence the admissibility of expert testimony internationally. The UK Law commission recently prescribed similar standards in its expert evidence consultation paper in which it laid emphasis on the scientific method much more than falsifiability [3].
The ACE-V method that has been in use by fingerprint expert examiners (which has been the subject of controversy) has had several objections raised against them, some of which according to [2] are:

- the contextual bias that arises from...

... middle of paper ...

...= {GPy,Ry,Nty}, and the continuous feature vectors xc and yc contain the remaining features in x and y respectively [5].
By this model, a value of 1 is assigned to the numerator of the discrete likelihood, while frequencies of general pattern multiplied region and minutia-type probabilities were used for calculating the denominator.
The between-finger and within-finger LRs were also evaluated in two experiments. 216 fingerprints from 4 different fingers were used to evaluate the within-finger LR while 818 fingerprints dataset was used to evaluate the between-finger LR. Normally, likelihood ratios for within-finger tests are expected to be greater than between-finger LRs[2].
4.1.2 Model of Neumann et al (2012)
Neumann and co. also in 2012 [8], proposed another LR model based on Feature vectors but this time around, they proposed the use of radial triangulation. The

Need Writing Help?

Get feedback on grammar, clarity, concision and logic instantly.

Check your paper »