Information Evaluation

Chapter 7

Quantitative Information Evaluation: Modeling and Experimental Evaluation

7.1. Introduction

As part of the process of information watch, information evaluation, such as it is defined in the military doctrinal corpus, aims to evaluate a piece of information using two ratings: the reliability of the source and the credibility of the informational content. As discussed in Chapter 4, the rules and good practices set out in the doctrine to quantify these ratings suffer from certain shortcomings. For instance, we can point to the independence between the reliability of a source and the credibility of the informational content, or the lack of consideration given to relations between sources reporting the same information.

The problem of evaluating the quality of information is by no means limited to the military domain. In the civilian domain, it is accorded great importance in sectors such as competitive intelligence or health, as demonstrated by the setting up of projects such as Net Scoring.¹ The approaches developed are usually based on manual, rather than automated, procedures to judge the quality of information available on the Web. As this type of approach is very costly to implement, projects drawing on the collaborative aspect of an increasing number of publication platforms have emerged. The study by [SET 10] or the Web of Trust² project are examples of this. The general idea in such a case is to exploit the relations between the users of a social network and the ratings ...