Named Entity Recognition in Turkish Bank Documents

Named Entity Recognition (NER) is the process of automatically recognizing entity names such as person, organization, and date in a document. In this study, we focus on bank documents written Turkish propose CRF model to extract named entities. The main contribution study twofold: (i) domain-specific features law, regulation, reference which frequently appear documents; (ii) contribute NER research document not mature other languahes English German. Experimental results based 10-fold cross validation conducted 551 real life, anonymized show proposed CRF-NER achieves 0.962 micro average F1 score. Morespecifically, score for identification law 0.979, regulation name 0.850, article no 0.850