Computer Engineering and Applications ›› 2007, Vol. 43 ›› Issue (34): 223-225.

• 工程与应用 • Previous Articles     Next Articles

Application of research on Chinese prosodic model base on multi-association rule

ZHAO Huan1,TAN Hua2   

  1. 1.School of Computer and Communication,Hunan University,Changsha 410082,China
    2.Management School,Hunan University,Changsha 410082,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-12-01 Published:2007-12-01
  • Contact: ZHAO Huan

多维关联规则在汉语韵律模型研究中的运用

赵 欢1,谭 华2   

  1. 1.湖南大学 计算机与通信学院,长沙 410082
    2.湖南大学 工商管理学院,长沙 410082
  • 通讯作者: 赵 欢

Abstract: Chinese prosody is characterized by its hierarchical structure when influenced by linguistic environments.Base on this,can apply association rules to deal with Chinese prosodic problem,obtain more interested rules that investigators want,then apply network model to analysis and output more efficient prosodic parameter.In the end,can construct the Chinese prosodic model better than ever.Put the prosodic model into software to emulate.The emulator output improves the model can meet the requirement of the speech system.

Key words: speech analysis, data mining, multi-association rule, neural network

摘要: 汉语韵律特征受语境参数影响时受前后音节的影响较大,利用数据挖掘中的多维关联规则方法处理汉语韵律问题,得出研究者感兴趣的有用规则,然后利用神经网络方法分析输出更为准确的韵律参数,构建语音合成系统的韵律模型。将构建的韵律模型在软件平台上进行模拟实验,实验证明,该模型能满足语音合成系统的要求。

关键词: 语音合成, 数据挖掘, 多维关联规则, 神经网络