Leveraging Integrated Learning for Open-Domain Chinese Named Entity Recognition

Named entity recognition (NER) is a fundamental technique in natural language processing that provides preconditions for tasks, such as question reasoning, text matching, and semantic similarity. Compared to English, the challenge of Chinese NER lies noise impact caused by complex meanings, diverse structures, ambiguous boundaries itself. At same time, compared with specific domains, open-domain types are more changeable, number entities considerably larger. Thus, task difficult. However, existing methods have low rates. Therefore, this paper proposes method based on bidirectional long short-term memory conditional random field (BiLSTM-CRF) model, which leverages integrated learning improve efficiency NER. single models, including CRF, BiLSTM-CRF, gated recurrent unit-CRF, proposed can significantly accuracy