Unequal-step multiscale integrated mapping dispersion Lempel-Ziv complexity: A novel complexity metric for signal analysis

Dispersion Lempel-Ziv complexity (DLZC), as a more effective complexity metric proposed in recent years, can effectively quantify the dynamic change of signal. However, DLZC still has shortcomings of insufficient applicability and lack of scale information. To surmount these drawbacks, we fully utilize the advantages of multiple mapping methods and integrate these mapping methods into DLZC, thus the integrated mapping DLZC (IMDLZC) is proposed, which has higher separability and applicability. Moreover, we developed a coarse-grained method called unequal step multi scale analysis, which solves the problem that the defects of subsequences decrease sharply with the increase of scale in existing coarse-grained methods. Then a new complexity metric for signal analysis called unequal-step multiscale IMDLZC (UM-IMDLZC) is proposed by coupling unequal-step multiscale analysis and IMDLZC. The experimental results of three kinds of synthetic signals show that UM-IMDLZC is more sensitive to the dynamic changes of signals and has the more excellent robustness; moreover, compared with other four complexity metrics, the proposed UM-IMDLZC has the best classification effect for different types of realistic signals. •The integrated mapping dispersion Lempel-Ziv complexity (IMDLZC) is proposed.•The unequal-step multiscale analysis is developed.•The unequal-step multiscale IMDLZC (UM-IMDLZC) is proposed.•The proposed UM-IMDLZC has advanced performance in signal analysis.