本论文针对词库外的词汇和错误识别进行拒识，以期达到减少识别错误，提高系统的识别率，降低虚警率的目的。

In this paper, the word out of vocabulary and misrecognition are rejected to cut down the recognition faults, improve the system recognition ratio and reduce the false acceptance ratio.

youdao

拒识的平均正识率为86.33%，拒识后平均候选个数降为3.46(未进行拒识前是10名候选)，总的拒识错误率为0.27%。

The average recognition accuracy is 86.33% with 3.46 candidate number (the number was 10 before rejection) on an average, and the total error rejection rate is 0.27%.

youdao

在系统的后端处理中，提出了一种基于置信测度的拒识方法改善系统的稳健性，最终使610个孤立词的识别任务，系统的等错误率小于5%；

In order to obtain an available equipment, a confidence measure method is suggested, which cost a little additive computation. For 610 words of the corpus, an equal-error-rate is less than 5%.

youdao

在系统的后端处理中，提出了一种基于置信测度的拒识方法改善系统的稳健性，最终使610个孤立词的识别任务，系统的等错误率小于5%；

In order to obtain an available equipment, a confidence measure method is suggested, which cost a little additive computation. For 610 words of the corpus, an equal-error-rate is less than 5%.

youdao