Implementação de um método para extração de citações bibliográficas usando uma base de conhecimento

Abstract

Information Retrieval is an area of computing that works with automatic retrieval of information contained in documents. The main objective of this study was to develop and implement a method to extract citation from scanned text documents, identifying each citation contained therein. The method uses a knowledge base of bibliographic citations used to classify each token of the document read and a context-free grammar that determines the patterns of references to identify the citations contained in the texts. After the development and implementation, was made an analysis of the results obtained with the execution of the method developed. It was made two experiments and calculated values of Precision, Recall and F-measure on the data of each experiment, getting satisfactory results.