Scholarlink

<p>Multi-scale entropy can effectively characterize the complexity of signals at different scales as a signal complexity metric. However, a specific multi-scale entropy utilizes only limited amplitude information and cannot fully characterize the complexity of the signal. To address this limitation, multi-scale fusion simplified entropy (MFSE) has been proposed, which fused multi-source amplitude information by exploiting the complementarities of the amplitude information and amplitude relationships for overall and local, and eliminated redundant information and captured non-linear relationships in the signal by integrating t-Distributed Stochastic Neighbor Embedding (t-SNE) algorithm. Simulated experiments show that MFSE has better robustness and differentiation ability for chaotic signals under different signal-to-noise ratios (SNRs). Four sets of real world signal experiments show that compared with four types of multi-scale entropy dimensionality reduction before and after, MFSE has significant advantages in distinguishing different types of signals under different feature numbers.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multi-scale fusion simplified entropy: a novel measure for signal analysis

  • 期刊论文

  • 发表日期：

    2026年1月21日

查看全文
  • Yuxing Li,
  • Yang Liu

摘要

Multi-scale entropy can effectively characterize the complexity of signals at different scales as a signal complexity metric. However, a specific multi-scale entropy utilizes only limited amplitude information and cannot fully characterize the complexity of the signal. To address this limitation, multi-scale fusion simplified entropy (MFSE) has been proposed, which fused multi-source amplitude information by exploiting the complementarities of the amplitude information and amplitude relationships for overall and local, and eliminated redundant information and captured non-linear relationships in the signal by integrating t-Distributed Stochastic Neighbor Embedding (t-SNE) algorithm. Simulated experiments show that MFSE has better robustness and differentiation ability for chaotic signals under different signal-to-noise ratios (SNRs). Four sets of real world signal experiments show that compared with four types of multi-scale entropy dimensionality reduction before and after, MFSE has significant advantages in distinguishing different types of signals under different feature numbers.