A maximum entropy approach to natural language processing
Computational Linguistics
Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data
ICML '01 Proceedings of the Eighteenth International Conference on Machine Learning
Hi-index | 0.00 |
Automatic recognition of prepositional usage is of great significance in parsing and syntax analysis. Many researches have been focused on preposition usage. In this paper, we introduce the triune knowledge base (usage dictionary, usage rule and usage corpus) of Contemporary Chinese preposition that we have finished .On this basis, we firstly adopt rule-based method to automatically annotate the prepositions in the corpus of People's Daily, in which the precision rate achieves 68.68%. Then to the prepositions whose precision rate is less than 80%, we use statistics-based method to annotate them with different models, features and context windows. The best precision rate achieves 90.86%. Experiments show that the statistics-based method can efficiently meet the need of the automatic recognition of prepositions' usage.