State of the Art of Automatic Keyphrase Extraction Methods (État de l'art des méthodes d'extraction automatique de termes-clés) [in French]

State of the Art of Automatic Keyphrase Extraction Methods This article presents the state of the art of the automatic keyphrase extraction methods. The aim of the automatic keyphrase extraction task is to extract the most representative terms of a document. Automatic keyphrase extraction methods can be divided into two categories : supervised methods and unsupervised methods. For supervised methods, the task is reduced to a binary classification where terms are classified as keyphrases or non keyphrases. This classification requires a learning step which is not required by unsupervised methods. The unsupervised methods use features extracted from the analysed document (sometimes a document collection) to check properties which allow keyphrase identification. MOTS-CLÉS : extraction de termes-clés ; méthodes supervisées ; méthodes non-supervisées ; état de l’art .