Abstract
Recurrence is a fundamental property of dynamical systems and studying it helps in deepen our understanding of the systems and predict their evolution. Recurrence has been studied in many real systems, such as financial exchange rates, damage detection and neuroscience data, but, as far as we know have never been applied with the aim of detecting demand patterns needed to develop a reliable forecast model. The analysis is performed using the Recurrence Plot (RP) together with the tool called Recurrence Quantification Analysis (RQA). RP is based on the computation of a distance matrix between reconstructed points in phase space. RQA provides a set of measures for RP quantification. In this work we analyse the RPs corresponding to several simulated demand time series and the relationship between Determinism (DET, a RQA measure) and parameters traditionally used to assess demand patterns. The visualization of the RP and the calculation of DET of the simulated demand time series provide positive results when used to detect demand patterns.