A Survey on the Methods to Determine the Sensitivity of Textual Documents: Solutions and Problems to Solve

Abstract

The identification of sensitive information, whether personal or institutional, is a fundamental step when dealing with the problem of information leakage. This problem is one of the most pressing to which companies and research centers dedicate a considerable amount of material and intellectual resources, as a particular case, to the development of methods or the application of some already known ones to the identification of sensitive information. This increased the proposals with promising results, but without yet offering a totally satisfactory solution to the problem. Under these conditions, it is considered necessary to make a critical analysis of the existing methods and techniques and their future projections. In this paper, a review of the proposals for the determination of sensitivity in textual documents is presented and a taxonomy is introduced to better understand the approaches with which this problem has been approached in the context of information leakage. Starting from the critical analysis and the practical needs raised by experts in the areas of possible application, lines of research on this subject are outlined that include the development of methods for the automation of the classification of sensitive textual documents. Possible extensions that these studies may have in similar application areas are proposed based on other information carriers, such as the cases of images, recordings and other forms of information object, each of which entails levels of complexity that merit studies analogous to the one carried out in this work.