基于集成学习的英文语义识别方法研究与实现  

A study of English semantic intelligence analysis method based on ensemble learning

作　　者：张慧[1] 许大炜[1] ZHANG Hui;XU Dawei(City College,Xi’an Jiaotong University,Xi’an 710018,China)

机构地区：[1]西安交通大学城市学院,陕西西安710018

出　　处：《电子设计工程》2020年第24期43-47,共5页Electronic Design Engineering

基　　金：陕西省社科基金(2018M30)。

摘　　要：针对英文语义自动化识别的需求,文中对英文的语义角色标注方法进行了研究,通过引入集成学习算法,实现了英文语句中施事、受事、来源、目的等论元标签的识别。文中的集成学习算法使用Adaboost作为集成策略,集成学习中使用的学习器为回归树,其内部结构为二叉树,训练收敛速度快,具有良好的分类效率。为了验证算法的性能,在开放数据集ENTBv-1上进行了算法的性能测试。结果表明,算法对文中定义的不同英文组块具有不同的识别效率,其中对DNP、DP、PP、QP等的识别效果较为理想,准确率、召回率与F值均在95%以上。模型的总体准确率可达到93.24%,召回率90.83%,F值92.02%,能够实现较为准确的英文语义角色识别与标注。In order to meet the needs of automatic recognition of English semantics,this paper studies the semantic role tagging method in English,and realizes the recognition of agent,patient,source and purpose in English sentences by introducing the integrated learning algorithm.The ensemble learning algorithm in this paper uses Adaboost as the integration strategy,the learner used in the ensemble learning is CART,its internal structure is binary tree,the training convergence speed is fast,and it has good classification efficiency.In order to verify the performance of the algorithm,the performance test of the algorithm is carried out on the open data set ENTBv-1.The results show that the algorithm has different recognition efficiency for different English chunks defined in the paper.Among them,the recognition effect for DNP,DP,PP,QP,etc.is ideal,and the accuracy,recall rate and F value are all above 95%.The overall accuracy of the model can reach 93.24%,recall rate 90.83%,F value 92.02%,which can achieve more accurate English semantic role recognition and annotation.

关 键 词：语义分析 角色标注 集成学习 回归树 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]