摘要
对于关键词检测系统来讲,垃圾模型的结构和类型对整个系统的性能有很大的影响。文中针对中文关键词检测系统提出了一种与上下文相关的垃圾模型。通过实验验证,关键词的检测率有了很大的提高,实验结果表明了该模型的有效性。
For the keyword spotting system, the structure and type of garbage model have an important effect on the overall system performance. The garbage model which is context-dependent is presented in this paper. The experiment validates that the garbage model has a great improvement in the probabihty detection for the keyword. Experimental results show that the model is effective.
出处
《信息技术》
2009年第6期142-144,共3页
Information Technology