时间表达式识别技术主要分为两类:基于机器学习的序列标注方法和基于规则的方法。
There are two main technologies in time expression recognition:sequence labeling method based on machine learning theory and rule-based method.
我们的方法是将简称生成问题转化为等价的序列标注问题,并利用一阶条件随机场建立自动生成模型。
In our method, we transformed the problem into an equivalent sequence tagging problem, and built up the automatic generation model through the first order conditional random field.
这一方法将汉语语义角色标注从一个节点的分类问题转化为序列标注问题,我们使用了条件随机域这一模型,取得了较好的结果。
Based on the semantic chunking result, the Chinese SRL can be changed into a sequence labeling problem instead of the classification problem.
应用推荐