Legal Documents Links Extraction

Abstract

The presence of links in a text always causes difficulties in reading. It happens because it is often not easy to follow a text link, what especially comes to legal references. And there are still no solutions published for automate links extraction regarding to Russian legal documents. Therefore, the purpose of this work was to build a method of automatic search, extraction and structure allocation of legal references using a formal grammar. Besides the method we also build a link describing manifest and dataset for evaluation. Moreover, during the work there has been made a research showing how strict are government recommendations to the links writing observed in real life. Evaluation of our method showed that it helps to extract 94% of links with 98% quality.