CAJ | 학술논문

变性，提出一种基于隐马尔科夫模型的参数估计方法。通过将非平稳信源的输出对应为隐马尔科夫模型的观察值序列，构造合适的隐马尔科夫模型，改进前向算法和后向算法，最后再利用改进后的算法对非平稳信源的统计特性进行估计。实验结果表明利用提出的方法非常逼近非平稳信源的统计特性，为更有效地进行非平稳信源的压缩提供了实现基础。在数据压缩领域，如何获取非平稳信源的统计特性一直是一个被广泛关注的热点问题。针对非平稳信源统计特性的时
변성，제출일충기우은마이과부모형적삼수고계방법。통과장비평은신원적수출대응위은마이과부모형적관찰치서렬，구조합괄적은마이과부모형，개진전향산법화후향산법，최후재이용개진후적산법대비평은신원적통계특성진행고계。실험결과표명이용제출적방법비상핍근비평은신원적통계특성，위경유효지진행비평은신원적압축제공료실현기출。재수거압축영역，여하획취비평은신원적통계특성일직시일개피엄범관주적열점문제。침대비평은신원통계특성적시
In the field of data compression , how to obtain the statistical property of non-stationary source has always been a widely concerned hot topic.In this paper we propose an HMM-based parameter estimation method in light of the time variability of statistical property of non-stationary source .By corresponding the output of non-stationary source to the observation value sequence of an HMM , the author firstly constructs an appropriate HMM , then revises the classical forward algorithm and backward algorithm , and finally estimates the statistical property of non-stationary source using the improved algorithm . Experimental results show that the proposed method can be used to highly approach the statistical property of non-stationary source , which offers the achievable basis for more efficient non-stationary source compression .