Les apports du TAL à la lisibilité du français langue étrangère

This paper presents a set of experiments aiming to (1) assess the contribution of NLP to the specific issue of the readability of texts for French as a foreign language (FFL) readers and (2) to propose a new readability formula for FFL. This new model relies on 46 textual features representative of the lexical, syntactic, and semantic levels as well as some of the specificities of the FFL context. We report comparisons between several techniques for feature selection and for various learning algorithms. Our best model, based on support vector machines (SVM), significantly outperforms previous FFL formulas. Regarding the contribution of NLP to readability, our findings suggest that NLP-based models are not significantly better than the classic ones, although combining both type of information leads to significant improvment. MOTS-CLÉS : lisibilité du FLE, traitement automatique du langage, difficulté des textes.