用融和丰富字特征的不等式平滑最大熵模型消解组合型歧义  

Using Conciliatory Luxuriant Word Character to Resolve Ambiquity of Knowledge to Improve the Performance of Classification

作　　者：范荣 

机构地区：[1]辽宁省锦州市机电工程学校,锦州121013

出　　处：《电脑编程技巧与维护》2011年第2期11-13,22,共4页Computer Programming Skills & Maintenance

摘　　要：重点在于融和丰富字特征作为消歧知识以提高分类性能和引入不等式平滑技术来克服数据稀疏问题,同时不等式平滑技术还使特征选择嵌入到参数估计过程中,显著压缩模型规模。This paper＇s keystone rests with using conciliatory luxuriant word character to resolve ambiquity of knowledge to improve the performance of classification and introducing inequality smoothing technique to overcome data sparsity problem.Meanwhile,inequality smoothing technique can also embed character selection into the parameter estimation process,which can significantly reduce model size

关 键 词：消歧 不等式平滑技术 熵模型 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]