Minimally Supervised Japanese Named Entity Recognition: Resources and Evaluation

Abstract

Approaches to named entity recognition that rely on hand-crafted rules and/or supervised learning techniques have limitations in terms of their portability into new domains as well as in the robustness over time. For the purpose of overcoming those limitations, this paper evaluates named entity chunking and classification techniques in Japanese named entity recognition in the context of minimally supervised learning. This experimental evaluation demonstrates that the minimally supervised learning method proposed here improved the performance of the seed knowledge on named entity chunking and classification. We also investigated the correlation between performance of the minimally supervised learning and the sizes of the training resources such as the seed set as well as the unlabeled training data.

Resources

Details