Mathematical approach in forensic science and use of probability in evalution of evidence

In the last years, the forensic scientific community has proposed the use of scientific methods for the evalution of evidences based on probability. The Bayesian approach subsequently clarifies and restrains the role of the forensic scientist in legal proceedings and forces him to present the results of his analyses in a such appropriate way that no doubt of interpretation should arise. Hence, a "beyond a reasonable doubt" truth should be overcome by a quantification of an error rate, a confidence level, a reliability level, a likelihood ratio etc. Questions about the admissibility of the scientific evidence in Court are much-discussed. In particular, objections to the use of probability in legal proceedings are based on the presumed breach of the casuqality principle, according to which the proof of the personal responsability requires that the occured behaviour be a necessary condition for the crime. This thesis proposes a general panorama of the use of mathematics in forensic science, trying to give a systematic approach to the general principles of identity, and to the processes of identification and individualization. Moreover, the probability approach in evalution of evidences is introduced, and its applications in particular non-standard forensic branches are proposed. It is author's belief that the language of science (and then of forensic science) is only mathematics. Although its rigor, mathematics has pratical use in daily forensic scientist job, and this is the reason for which a formal mathematical approach is here suggested.