引入集成学习的最大熵短语调序模型  

An Ensemble Learning Method for Maximum Entropy Based Phrase Reordering Model

作　　者：何钟豪 苏劲松[2] 史晓东[1] 陈毅东[1] 黄研洲 

机构地区：[1]厦门大学智能科学与技术系,福建厦门361005 [2]厦门大学软件学院,福建厦门361005

出　　处：《中文信息学报》2014年第1期87-93,共7页Journal of Chinese Information Processing

基　　金：国家自然科学基金(61303082;61005052);国家科技支撑计划(2012BAH14F03);高等学校博士学科点专项科研基金(20120121120046)

摘　　要：基于最大熵的括号转录语法模型具有翻译能力强、模型训练简单的优点,成为近些年统计机器翻译研究的热点。然而,该模型存在短语调序实例样本分布不平衡的缺点。针对该问题,该文提出了一种引入集成学习的短语调序模型训练方法。在大规模数据集上的实验结果表明,我们的方法能有效改善调序模型的训练效果,显著提高翻译系统性能。The Maximum Entropy Based BTG model becomes a hot topic in statistical machine translation in recent years due to its strong translation and easy to-train abilities. However, the distribution of reordering examples in this model is imbalanced. To solve this problem, we introduce an ensemble learning method for training phrase reor- dering model. Experimental results show that,the reordering model can reach a better training effect via our method and the performance of the translation system is improved significantly in a large-scale dataset.

关 键 词：最大熵 短语调序 不平衡分类 集成学习 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]