SyntacticSearchOnTimeSeries

The significant increase of fields that generate time series has converged into massive amounts of data with different behaviours and therefore different solutions to answer in terms of pattern and query search. With this rises a need for a more expressive way of exploring and retrieve information from the available data. In this work we addressed this problem by developing an interactive tool, called Syntactic Search in Time Series (SSTS), that relies in expressing the morphological reasoning based on the visual interpretation of patterns in time series. The proposed framework has three steps: (1) Pre-Processing: applies typical pre-processing methods to prepare the signal, (2) Symbolic Connotation: translation of the signal into a symbolic representation based on the properties that most matter and ease the search, and (3) Search: string pattern that parses the connotation string, which in this case is a regular expression. We demonstrate the utility of this approach by solving typical query and pattern search tasks on several time series and evaluate the readability of the symbolic strings generated to solve the examples by comparing it with a traditional coding approach. The SSTS tool is a step forward in building a mechanism capable of augmenting the cognitive expressiveness to design the search method and increase the abstraction over the entire solving time series query task, easing the retrieval of information from time series.