Partially ordered template-based matching algorithm for financial time series

  • Authors:
  • Yin Tang

  • Affiliations:
  • Lab of E-Commerce, Management College, Jinan Univ., GZ, P.R. China

  • Venue:
  • IEA/AIE'06 Proceedings of the 19th international conference on Advances in Applied Artificial Intelligence: industrial, Engineering and Other Applications of Applied Intelligent Systems
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

Based on definitions of 1st and 2nd order atomic pattern of time series, this paper deduces n-th order atomic pattern, where partially-ordered relationship within these patterns is discussed. The framework enables more refined comparison between sequences, based on which we propose Template-Based Matching Algorithm. The experimental result has verified its distinct advantages over some similar and classical approaches both in accuracy and performance. Index Terms: time series, pattern recognition, case-based reasoning, partially order, lattice.