Abstract
A recurrence plot is a visualization tool for analyzing experimental data. These plots often reveal correlations in the data that are not easily detected in the original time series. Existing recurrence plot analysis techniques, which are primarily application-oriented and completely quantitative, require that the time-series data first be embedded in a high-dimensional space, where the embedding dimension d_E is dictated by the dimension d of the data set, with d_E >= 2d+1. One such set of recurrence plot analysis tools, Recurrence Quantification Analysis, is particularly useful in finding locations in the data where the underlying dynamics change. We have found that the same results can be obtained with no embedding. The work presented in this paper represents the beginning of an attempt to improve upon recurrence plot analysis in a way that incorporates and exploits their rich structural characteristics.
Full paper in gzipped postscript (3.7MB) and pdf (76.8MB). To unzip the postscript file, use the unix command gzip or the MacOS/Windows/DOS commands StuffIt or winzip.