Abstract:Sparse measures have attracted lots of interests from many fundamental research domains to be as objective functions of signal processing algorithms, health indices of degradation modeling and input features to machine learning algorithms. Among them, kurtosis and negative entropy are the most two popular sparse measures to characterize the sparsity of signals. For example, kurtosis and negative entropy are used in machine condition monitoring to quantify the sparsity of repetitive transients caused by localized rotating machine faults and to indicate an onset of early rotating faults. When kurtosis and negative entropy are decomposed into the sum of weighted normalized square envelope, the main difference between kurtosis and negative entropy is whether the logarithm transformation is applied to normalized square envelope to form a weight. In this paper, Box-Cox transformation as generalized power transformation is introduced to generalize the weights used in kurtosis and negative entropy and subsequently a new family of sparse measures, coined as Box-Cox sparse measures (BCSM), are proposed. The only parameter in the proposed BCSM is a transformation parameter λ ⩾ 0 . The contributions of this paper are summarized as follows. Firstly, this paper provides new propositions for intuitive sparse attributes of the proposed BCSM, which theoretically prove that the proposed BCSM satisfies all six intuitive sparse attributes. Secondly, in numerical and experimental studies, it is shown that (1) the proposed BCSM converges when the length of a signal increases; (2) only when a distribution is quite sparse, the proposed BCSM with λ > 1 can indicate the sparsity of the distribution. Being different form the performance of the proposed BCSM with λ > 1 , the proposed BCSM with 0 ⩽ λ ⩽ 1 can steadily indicate that a distribution is getting sparser, which indicates that negative entropy ( λ = 0 ) is the best choice among all the proposed sparse measures and it is better than kurtosis ( λ = 1 ) to quantify the sparsity of repetitive transients caused by rotating faults for machine condition monitoring; (3) the proposed BCSM with 0 ⩽ λ ⩽ 1 is more effective in monitoring bearing and gear health conditions than the proposed BCSM with λ > 1 . Thirdly, the proposed BCSM of a complex Gaussian signal is investigated to provide a theoretical baseline for machine condition monitoring. Finally, the proposed BCSM can be applied to any situations, where sparse measures are needed.

Box-Cox sparse measures: A new family of sparse measures constructed from kurtosis and negative entropy

Generalized Gini indices: Complementary sparsity measures to Box-Cox sparsity measures for machine condition monitoring

Gini Indices II and III: Two New Sparsity Measures and Their Applications to Machine Condition Monitoring

Probability Distributions and Typical Sparsity Measures of Hilbert Transform-Based Generalized Envelopes and Their Application to Machine Condition Monitoring

Box-cox-sparse-measures-based blind filtering: Understanding the difference between the maximum kurtosis deconvolution and the minimum entropy deconvolution

Correlation dimension and approximate entropy for machine condition monitoring: Revisited

Research on sparsity measures for rotating machinery health monitoring

Understanding importance of positive and negative signs of optimized weights used in the sum of weighted normalized Fourier spectrum/envelope spectrum for machine condition monitoring

Broad learning system based on the quantized minimum error entropy criterion

Assessing an Alternative for `Negative Variance Components': A Gentle Introduction to Bayesian Covariance Structure Modelling for Negative Associations Among Patients with Personalized Treatments

The Sum of Weighted Normalized Square Envelope: A Unified Framework for Kurtosis, Negative Entropy, Gini Index and Smoothness Index for Machine Health Monitoring

Theoretical Investigations on Kurtosis and Entropy and Their Improvements for System Health Monitoring

Investigations on the Sensitivity of Sparsity Measures to the Sparsity of Impulsive Signals

Reconstruction Guarantee Analysis of Basis Pursuit for Binary Measurement Matrices in Compressed Sensing

Measuring Sparsity of Wireless Channels

New and Improved Conditions for Uniqueness of Sparsest Solutions of Underdetermined Linear Systems

A Novel Nonlinear Dynamic Measure for Early Detection of Bearing Fault Using Weighted Squared Envelope-Based Symbolic Lempel-Ziv Complexity

Bayesian Compressive Sensing Using Reweighted Laplace Priors

Mathematical Foundation of a New Complexity Measure

Cumulative Diversity Pattern Entropy (cden): A High-Performance, Almost-Parameter-Free Complexity Estimator for Nonstationary Time Series

Detecting and displaying size bimodality: Kurtosis, skewness and bimodalizable distributions