La descrizione dei reati di criminalità organizzata e di tipo mafioso nel testo delle sentenze

This paper proposes the use of data mining techniques for the analysis of criminal sentences. In particular, we are interested in the criminal sentences emitted for mafia-related crimes (ex art. 51, co 3 bis c.p.p.) by the four districts of the Appeal Courts of Sicily from 2000 up to 2006, and deemed irrevocable for at least one defendant.This research is particularly important for three reasons. (1) There is currently a lack of comparative analysis of the four districts of the Appeal Courts of Sicily (2) there is currently no collection of data describing the institutional answer to the mafia phenomenon; (3) Sicilian jurisprudence leads the Italian jurisprudence for this kind of phenomenon. The main objective is to extract from the analysis of the text of the sentences a description of the development of the sentences, as well as the socio-economic scope in which the events relative to the sentences happened.The research has shown that sentences can provide a lot of information; in particular if one takes into account the juridical statistics, their insufficient reliability and their lack of informative character in relation to the recipients of the data, and the difficulty in collecting and comparing the data, that is caused by the speed of the legislative changes and by the classification criteria.Thanks to the close collaboration between social scientists and computer scientists, we have developed an ad hoc reading algorithm, that analyzes the internal structure of the sentences, and that can be generalized to other areas of the jurisprudence, as well as to other temporal and geographical scopes