Item

Abstract

Recurrence analysis techniques based on recurrence plots (RPs) have become a standard tool for the study of dynamical systems, offering means to quantify nonlinear characteristics from time series and to apply machine-learning methods in their analysis. Among recent developments, recurrence microstates analysis (RMA) is emerging as a complementary approach, enabling the estimation of typical quantifiers from recurrence quantification analysis (RQA) using the microstate distributions, as well as the use of these distributions as input features for machine-learning models. In this work, we introduce a Python library designed to provide the computational efficiency required for the optimized application of RMA techniques. The proposed methodology and performance benchmarks demonstrate a substantial reduction in computation time, making the library sustainable and energy efficient, especially for large-scale data analysis, and aligned with green computing principles.