TS-Evolutionary_Prototyping: A Python module for finding the prototype in large sets of time series

Time series analysis has become one of the basic building blocks for the technological fields of science and engineering. Therefore, there are a large number of software tools that encompass the preparation of the data, the performance of a large number of processing tasks with the data, the generation of datasets and finally the implementation of the necessary evaluation techniques. Of particular importance within the above tasks is the prototyping or summarisation of sets of time series as they have direct application in the resolution of clustering problems. In this work, we introduce a Python package that implements an evolutionary strategy to find prototypes. Given a set of time series, the implemented software finds prototypes using dynamic time warping (DTW) as the distance measure between series and does not restrict the search space for the prototype to the series of the input set. The software also includes use cases for clustering and classification. •A software for time series summarisation and prototyping in large datasets.•An evolutionary strategy mainly based in genetic algorithms to find the prototype.•Use of elastic distances as distance measures between series.•A Python library that obtains the centroid as prototype in large sets of time series.