| 语音识别中ARMA模型的研究 |
| |
| 引用本文： | 龚国勇.语音识别中ARMA模型的研究[J].玉林师范学院学报,2009,30(5). |
| |
| 作者姓名： | 龚国勇 |
| |
| 作者单位： | 玉林师范学院,数学与计算机科学系,广西,玉林,537000 |
| |
| 摘 要： | 在用线性时间序列模型拟合语音数据时(例如用AR模型拟合),模型阶数越高,参数个数越多,统计计算时误差越大.本文在理论上对ARMA(1,q)模型的参数估计进行了推导,在数值计算上,得到了ARMA(1,0)模型的参数估计并对模型进行拟合,解决了用AR模型拟合语音数据时,统计计算时误差较大的问题.
|
| 关 键 词： | ARMA(1 q)模型 参数估计 模拟 |
Research on ARMA Model in Speech Recognition |
| |
| GONG Guo-yong.Research on ARMA Model in Speech Recognition[J].Journal of Yulin Teachers College,2009,30(5). |
| |
| Authors: | GONG Guo-yong |
| |
| Institution: | GONG Guo-yong (Associate Professor; Department of Mathematics and Computer Science; Yulin Normal University; Yulin; Guangxi 537000); |
| |
| Abstract: | When fitting speech data by using a linear time series model, the higher order of the model and the more parameters are involved, the greater error will be in statistics calculation. This thesis deduces parameter estimation for ARMA(1,q)model theoretically, and in numerical calculation, parameter estimation for ARMA(1,0)model is obtained and the model is fit. It solves the error problems in statistics calculation when using AR model fitting speech data. |
| |
| Keywords: | ARMA（1 q）model parameter estimation simulation |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
