Time Series Representation for Identification of Extremes

Extracting information from the huge time series is a challenging task. Databases are prepared by keeping in mind the type of information required. Indexing a time series is a difficult task, where shape may not be exact. Finding the minima and maxima of the time series is another difficult job dependent on the expert’s subjectivity. In this information age algorithmic trading is the buzz word. For the identification of various patterns, a machine can be made intelligent by embedding some good algorithm in the trading module to identify the various patterns. In this paper an attempt has been made to identify the extremes, where profit probability is maximized.