Relative entropy fuzzy c-means clustering

The Application of LCS and Information Entropy as a Novel Fusion Algorithm for Degradation Feature Extraction

“…In this situation, the correlation coefficient can hardly evaluate the correlations. Based on information entropy theory, the relative entropy can reveal the correlation between nonlinear data by the analysis of the complexity [30]. Therefore, the feature fusion algorithm is proposed to make full use of the extracted features' information and to improve features conciseness.…”

Method for Feature Extraction Based on Composite Spectrum and Relative Entropy Fusion

“…通过分析可知，DCS 奇异谱能够从细节上相对 具体地反映信号复杂性的变化情况。当性能退化程 度较轻时， 复合谱中不同模式分量的分布比较均匀， 信号复杂度较高，相应地熵值越大。 2 基于关联熵的特征融合方法 特征融合的主要目的是通过对特征信息的综合 利用，改善融合结果的完整性和简洁性。其方法主 要有两大类 [21] ： 基于 JBT(Joint-based technique)的融 合技术和基于 UBT(Union-based technique)的融合技 术，其中，JBT 侧重于保持特征信息的简洁性，而 UBT 侧重于保持特征信息的完整性。本文将采用基 于 JBT 技术的融合方法， 其典型代表就是关联熵 [22] 。 关联熵能够有效地反映不同变量间的信息冗余程 度，关联熵越大，两变量间的关联性越强，信息冗 余度越大 [23] …”

A hybrid fuzzy time series model based on granular computing for stock price forecasting

“…Finally, the cutting point selection policy is used repeatedly on each partition until stopping criterion is fulfilled. Entropy-based discretization has been successfully applied in many domains [32,36,64], but has never been used in forecasting with fuzzy time series models. For this reason, we adopt entropy-based discretization to measure whether the linguistic values approach a steady state belonging to the fuzzy set.…”