Recurrent neural network based loanwords identification in Uyghur

摘要: Comparable corpus is the most important resource in several NLP tasks. However, it is very expensive to collect manually. Lexical borrowing happened in almost all languages. We can use the loanwords to detect useful bilingual knowledge and expand the size of donor-recipient/recipient-donor comparable corpora. In this paper, we propose a recurrent neural network (RNN) based framework to identify loanwords in Uyghur. Additionally, we suggest two features: inverse language model feature and collocation feature to improve the performance of our model. Experimental results show that our approach outperforms several sequence labeling baselines.