Data Mining - Concepts and Techniques.

Complex network approaches to nonlinear time series analysis

Revealing complicated behaviors from time series constitutes a fundamental problem of continuing interest and it has attracted a great deal of attention from a wide variety of fields on account of its significant importance. The past decade has witnessed a rapid development of complex network studies, which allow to characterize many types of systems in nature and technology that contain a large number of components interacting with each other in a complicated manner. Recently, the complex network theory has been incorporated into the analysis of time series and fruitful achievements have been obtained. Complex network analysis of time series opens up new venues to address interdisciplinary challenges in climate dynamics, multiphase flow, brain functions, ECG dynamics, economics and traffic systems.

https://iopscience.iop.org/article/10.1209/0295-5075/116/50001/pdf

Complex network analysis of time series

As an extension of probability theory, evidence theory is able to better handle unknown and imprecise information. Owing toits advantages, evidence theory has more flexibility and effectiveness for modeling and processing uncertain information.Uncertainty measure plays an essential role both in evidence theory and probability theory.In probability theory, Shannon entropy provides a novel perspective for measuring uncertainty. Various entropies exist for measuring the uncertainty of basic probability assignment (BPA) in evidence theory. However, from the standpoint of the requirements of uncertainty measurement and physics, these entropies are controversial. Therefore, the process for measuring BPA uncertainty currently remains an open issue in the literature.Firstly, this paper reviews the measures of uncertainty in evidence theory followed by an analysis of some related controversies. Secondly, we discuss the development of Deng entropy as an effective way to measure uncertainty, including introducing its definition, analyzing its properties, and comparing it to other measures. We also examine the concept of maximum Deng entropy, the pseudo-Pascal triangle of maximum Deng entropy, generalized belief entropy, and measures of divergence.In addition, we conduct an analysis of the application of Deng entropy and further examine the challenges for future studies on uncertainty measurement in evidence theory. Finally, a conclusion is provided to summarize this study.

Uncertainty measure in evidence theory

Visibility graph has established itself as a powerful tool for analyzing time series. We in this paper develop a novel multiscale limited penetrable horizontal visibility graph (MLPHVG). We use nonlinear time series from two typical complex systems, i.e., EEG signals and two-phase flow signals, to demonstrate the effectiveness of our method. Combining MLPHVG and support vector machine, we detect epileptic seizures from the EEG signals recorded from healthy subjects and epilepsy patients and the classification accuracy is 100%. In addition, we derive MLPHVGs from oil-water two-phase flow signals and find that the average clustering coefficient at different scales allows faithfully identifying and characterizing three typical oil-water flow patterns. These findings render our MLPHVG method particularly useful for analyzing nonlinear time series from the perspective of multiscale network analysis.

/pdf/multiscale-limited-penetrable-horizontal-visibility-graph-cfg625pd79.pdf

Multiscale limited penetrable horizontal visibility graph for analyzing nonlinear time series

Failure mode and effects analysis (FMEA) is a widely used reliability analysis tool for identifying and eliminating known or potential failures in system, design, and process. In traditional FMEA, failure modes are evaluated by FMEA team members with respect to three risk factors: severity (S), occurrence (O), and detectability (D), and ranked via their risk priority number (RPN), which is obtained by multiplying the crisp values of S, O, and D. However, traditional RPN has been considerably criticized due to the following shortcomings: not considering the different weights of risk factors; the identical value of RPN for different combinations of S, O, and D; the diversity and uncertainty of evaluation information given by FMEA team members and without considering the dependence among different failure modes. Although significant efforts have been made in FMEA literatures to overcome these shortcomings, there are still some deficiencies. In this paper, a new risk priority model is presented for FMEA by using the house of reliability (HoR)-based rough VIsekriterijumska optimizacija i KOmpromisno Resenje (VIKOR) approach. In the proposed model, the HoR is introduced to identify the dependence among different failure modes and the link between failure modes and the risk factors of O, D and the subcriteria of S. Rough number is introduced to manipulate the subjectivity and vagueness in decision making and VIKOR approach is used to determine the risk priority order of failure modes in a comprehensive way. Finally, an illustrative case in transmission system of a vertical machining center has demonstrated the effectiveness and practicality of the proposed model.

Failure Mode and Effects Analysis by Using the House of Reliability-Based Rough VIKOR Approach

Performance pattern identification is the key basis for fault detection and condition prediction, which plays a major role in ensuring safety and reliability in complex electromechanical systems (CESs). However, there are a few problems related to the automatic and adaptive updating of an identification model. Aiming to solve the problem of identification model updating, a novel framework for performance pattern identification of the CESs based on the artificial immune systems and incremental learning is proposed in this paper to classify real-time monitoring data into different performance patterns. First, an unsupervised clustering technique is used to construct an initial identification model. Second, the artificial immune and outlier detection algorithms are applied to identify abnormal data and determine the type of immune response. Third, incremental learning is employed to trace the dynamic changes of patterns, and operations such as pattern insertion, pattern removal, and pattern revision are designed to realize automatic and adaptive updates of an identification model. The effectiveness of the proposed framework is demonstrated through experiments with the benchmark and actual pattern identification applications. As an unsupervised and self-adapting approach, the proposed framework inherits the preponderances of the conventional methods but overcomes some of their drawbacks because the retraining process is not required in perceiving the pattern changes. Therefore, this method can be flexibly and efficiently used for performance pattern identification of the CESs. Moreover, the proposed method provides a foundation for fault detection and condition prediction, and can be used in other engineering applications.

An artificial immune and incremental learning inspired novel framework for performance pattern identification of complex electromechanical systems

Failure mode and effects analysis using Dempster-Shafer theory and TOPSIS method: Application to the gas insulated metal enclosed transmission line (GIL)

Fault recognition using an ensemble classifier based on Dempster–Shafer Theory

In order to recognize the different operating conditions of a distributed and complex electromechanical system in the process industry, this work proposed a novel method of condition recognition by combining complex network theory with phase space reconstruction. First, a condition-space with complete information was reconstructed based on phase space reconstruction, and each condition in the space was transformed into a node of a complex network. Second, the limited penetrable visibility graph method was applied to establish an undirected and un-weighted complex network for the reconstructed condition-space. Finally, the statistical properties of this network were calculated to recognize the different operating conditions. A case study of a real chemical plant was conducted to illustrate the analysis and application processes of the proposed method. The results showed that the method could effectively recognize the different conditions of electromechanical systems. A complex electromechanical system can be studied from the systematic and cyber perspectives, and the relationship between the network structure property and the system condition can also be analyzed by utilizing the proposed method.

Zhiyong Gao

Papers

Failure Mode and Effects Analysis by Using the House of Reliability-Based Rough VIKOR Approach

An artificial immune and incremental learning inspired novel framework for performance pattern identification of complex electromechanical systems

Failure mode and effects analysis using Dempster-Shafer theory and TOPSIS method: Application to the gas insulated metal enclosed transmission line (GIL)

Fault recognition using an ensemble classifier based on Dempster–Shafer Theory

Complex network theory-based condition recognition of electromechanical system in process industry