The French national Library is involved in one of the most interesting current European projects, Europeana14-18 which aims to give in open access digital European corpora. Among the corpora provided,
The French national Library is involved in one of the most interesting current European projects, Europeana14-18 which aims to give in open access digital European corpora. Among the corpora provided, the French national Library has decided to provide to the project press issues related to the murder of Franz Ferdinand in Sarajevo in June 1914 which led, through the mechanism of military alliances, to the outbreak of the First World war.This studies at the crossroad of technical aspects (what we could expect from optical recognition of characters and named entity recognition processes regarding the physical constraints of such a corpus) and intellectual ones both in terms of valorisation and from an epistemological point of view (how the use of new technologies allows new way of understanding the past). We intend first to present the latest technological developments concerning named entity recognition process applied to a corpus characterized by poor quality both of the paper and the ink. Regarding named entity recognition, we have used the results of the very first OCR outputs in order to evaluate existing named entities resources thanks to dictionaries specially dedicated to this purpose.We have applied this process to the study of the French public opinion during the days following the murder of Franz Ferdinand in Sarajevo. Using title such as Le Temps, Le Figaro and l’Humanité, we have mainly focus on the awareness of the possible consequences of this murder during these turning point days which led to the war. Our study intends to share with the community the very first results of this original and ambitious study which not belong only to French History but to European memory.
Lire la suite
