Abstract

Due to wide use of Information Technology, substantial amount of information is being gathered for exploratory analyses, business operations and online networking in the big data era. Due to the gathering of information for several events at distinctive time periods, huge datasets are formed. A time series can be defined as a series of numeric values obtained at various points, occurring after regular intervals. An interesting research problem in time series is motif discovery. Motif discovery subroutine can be utilized as a part of algorithms for classification and summarization. In this paper, comparative analysis of various motif discovery methods in time series datasets has been done and a method for improving the same has been suggested.