CAJ | 학술논문

利用遗传算法优越的全局搜索能力对传统DTW算法进行改进，重点研究遗传动态时间规划算法(GA_DTW)的实现机理、编码方式、适应度函数设计、种群初始化、选择机理、交叉运算、变异操作和终止策略。实验结果表明，在孤立词的说话人识别上，该算法具有识别率更高、耗时更少的优点。
이용유전산법우월적전국수색능력대전통DTW산법진행개진，중점연구유전동태시간규화산법(GA_DTW)적실현궤리、편마방식、괄응도함수설계、충군초시화、선택궤리、교차운산、변이조작화종지책략。실험결과표명，재고립사적설화인식별상，해산법구유식별솔경고、모시경소적우점。
Using genetic algorithms superior global search capability,it improved the traditional DTW algorithm. Focus on implementation mechanism, encoding, fitness function, population initialization, selection mechanism, crossover operator, mutation operator and termination policies of genetic algorithm dynamictimewarping (GA_DTW). Experimental results show that about the isolated word speech recognition, the algorithm has a higher recognition rate, consuming fewer advantages.