摘要
农业命名实体的识别对于农业领域信息抽取起着重要的作用,本文采用条件随机场(CRF)模型,利用2种标注集,选取词汇、词形、语法等一系列特征进行训练,通过不断调整模板和特征选择改进训练模型,实验取得较好的识别效果。
Agricultural named entity recognition has played a very important role in information extraction field of agriculture . In this paper, We use Conditional Random Fields (CRF) model and use two sets of labeling strategies including a series of feature training by selecting vocabu- lary, morphology, syntax features. By continuously adjusting the templates and feature selection improving training models, We has obtained better recognition results.
出处
《河北农业大学学报》
CAS
CSCD
北大核心
2014年第1期132-135,共4页
Journal of Hebei Agricultural University
关键词
农业命名实体识别
CRF模型
特征选择
agricultural named entity recognition~ CRF model lfeature selection.