Maturity of scientific theories has facilitated creation of advanced technology of human-engineered complex systems. A major challenge in these systems is online detection of behavioral uncertainties due to gradual evolution of anomalies (i.e., deviations from the nominal condition). These anomalies may alter the quasi-static behavior that causes performance degradation and can eventually lead to catastrophic failures. Therefore, for safe and reliable operation, it is essential to develop robust analytical tools for online degradation monitoring and for generating advanced warnings of emerging anomalies. Since it is often infeasible to achieve the required modeling accuracy due to the presence of i) high dimensionality, ii) non-stationarity (possibly chaotic behavior), iii) nonlinearity, and iv) exogenous disturbances, time series analysis of appropriate sensor data provides one of the most powerful tools for degradation monitoring of complex systems. This book presents a data-driven pattern identification methodology, built upon multidisciplinary concepts of Symbolic Dynamics, Automata Theory and Information Theory, with diverse applications to complex electromechanical systems.