基于基频的朝鲜语方言辨识方法的研究

发布时间：2018-07-16 18:35

【摘要】：该文提出了一种基于基音频率特征的中国朝鲜族语言、韩国朝鲜语和朝鲜朝鲜语方言的自动辨识方法。首先,选择具有良好区分度的基频移位差分系数作为三个方言的特征参数;其次,设计和采用了分层支持向量机分类器,并进一步引入投票法确定最佳的分类结果。实验结果表明该文提取的特征参数具有良好的区分性和较强的稳定性,该文提出的方言辨识方法比传统的移位差分倒谱系数特征方法识别率高,可以有效解决朝鲜朝鲜语、韩国朝鲜语和中国朝鲜族语言的方言辨识问题。
[Abstract]:This paper presents an automatic recognition method for Korean and Korean dialects in China based on pitch frequency features. Firstly, the fundamental shift difference coefficient with good differentiation is chosen as the characteristic parameter of the three dialects. Secondly, the hierarchical support vector machine classifier is designed and adopted, and the voting method is introduced to determine the best classification result. The experimental results show that the feature parameters extracted in this paper have good distinctiveness and strong stability. The dialect identification method proposed in this paper has a higher recognition rate than the traditional shift difference cepstrum coefficient feature method, which can effectively solve the Korean language. Identification of Korean and Chinese Korean dialects.
【作者单位】：延边大学计算机科学与技术学科智能信息处理研究室;
【基金】：吉林省科技厅自然科学基金(20140101225JC)
【分类号】：TN912.3

【相似文献】