利用上下文关联的思想,提出了三音素模型的大型句子文本库中选取最佳句子文本的算法,充分考虑了每个句子涵盖的不同三音素模型,利用贪婪算法去除了众多句子之间的冗余度,从而达到了选择文本容量最小化的目标。通过C#语言实现了本算法,给出了算法流程和算法性能分析,结果表明此算法的有效性和实用性。 Using the idea of ​​contextualization, an algorithm for selecting the best sentence text in a large sentence text database of three-phone model is proposed. Different triphone models covered by each sentence are fully considered, and greedy algorithm is used to remove the redundancy between many sentences Degree, so as to achieve the goal of minimizing the choice of text size. The algorithm is implemented in C # language, and the algorithm flow and the algorithm performance analysis are given. The results show that the algorithm is effective and practical.