C-CRF模型在农作物名识别中的应用  

Research on application of C-CRF model on Named Entity Recognition of the agricultural product

作　　者：方莹[1] 

机构地区：[1]商丘师范学院计算机科学系,河南商丘476000

出　　处：《广东农业科学》2011年第6期197-199,共3页Guangdong Agricultural Sciences

基　　金：河南省科技厅基础与前沿技术研究项目（102300410208）; 商丘师范学院青年基金研究项目（2010by11）

摘　　要：提出了一种基于层叠条件随机场模型（C-CRF）的农业领域命名实体识别方法。该方法在底层以字作为切分粒度,充分利用农作物特征词典,识别简单命名实体。识别结果作为高层条件随机场的输入,在高层以词作为切分粒度,结合复杂特征,识别嵌套命名实体。实证分析表明该方法获得了较好的结果。This paper proposed a method for the Named Entity Recognition of the agricultural domain based on the Cascading Conditional Random Fields model.On the first step,this paper adopted character as the grain processing unit;used the context information and especial cue information of agricultural domain names as the features of recognition,and adopted the artificially constructed dictionary to build the model.Then the results of the first step were sent to the second step,in which adopted the word as the unit with complex features.A good result had been obtained by the experiment.

关 键 词：农作物 层叠条件随机场 特征模板 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]