Analyse formelle de concepts incertains pour l’étude d’un corpus de textes

The analysis of a corpus by an expert takes a relatively long time. The development of digital tools made it possible to generate a summary of information contained in the corpus very quickly. In this research, we will focus on the contribution of formal concept analysis (FCA) to the analysis of a corpus. FCA makes it possible to build a model also called Hasse diagram which can be queried to find relevant concepts. It can happen that some words are misspelled or quantitative variables are associated to the corpus, which can generate uncertainties during the processing. Uncertainties may also arise in queries when human knowledge is imprecise. Possibility theory allows us to represent these imperfections. The combination of textual analysis tools and AFC allows us to present more relevant results.