Statistica testuale e text mining: alcuni paradigmi applicativi

In this paper, after reconstructing some essential phases in the evolution of automatic analysis of texts, the steps of an ideal strategy for the statistical analysis of textual data are defined. The characteristics of lexical and textual analysis are described, as well as some techniques of information extraction, that employ resources which are endogenous and exogenous with respect to the texts to be examined. In order to show the potential of textual statistics and of the most recent Text Mining applications, some relevant case studies concerning statistical survey and document analysis are illustrated.