Practical Named Entity Tagging using Co-training

3] and [1] opened the possibility of using an unlabeled corpus through co-training, a semi-supervised learning algorithm, to classify named entities. Our approach to solve the problem of Korean named entity classification also adopted a co-training method called DL-CoTrain. However, we use only a part-of-speech tagger and a simple noun phrase chunker instead of a full parser to extract the contextual features of a named entity for robustness and practicality. We will discuss the linguistic features in Korean which are valuable for named entity classification and experimentally show how large a labeled corpus and which unlabeled corpus is necessary for the superior performance and portability of a named entity classifier. With only a quarter of the labeled corpus, our method can compete with its supervised counterpart.