Pattern Recognition Through PerceptuallyImportant Points In Financial Time Series

Abstract

Technical Analysis is a financial risk management practice that has been in use since the advent of the stock market and Pattern Recognition is an indivisible part of it. There has been a lot of research into pattern recognition in time series. Existing pattern recognition techniques lack dynamic extensibility. They do not provide any interfaces in order to include new patterns for recognition dynamically. This limits the operability of these techniques to a particular domain. This research devises a new technique for domain independent pattern recognition while giving sufficient speed and accuracy. This enables it to be used by critical Decision Support Systems for time series of different domains. The system emulates the human visual cognition process by implementing the concept of Perceptually Important Points Identification (PIPI). Perceptually Important Points (PIP) represents the minimal set of data points which are necessary to form a pattern. For dynamic inclusion of patterns a Pattern Definition Language (PDL) has been conceptualized for defining patterns in time series by using a declarative programming paradigm. This also results in domain