Linguistic comparison of time series using the End-Point Fit algorithm

In this paper, we present an approach to linguistically describe two time series based on their domain values and trends. The Iterative End-Point Fit algorithm has been selected for obtaining a suitable segmentation of the series to be used in the description process. The comparison between series is performed by calculating the difference between values of two time series, which are defined on the same time domain. The resulting time series is described by several kind of summaries according to the way the difference is calculated and the permissible error. We have illustrated this proposal with several time series that represent the CO2 Emission in different countries.