The Method for Time Series Based on Symbolic Form and Area Difference

Abstract

Symbolic Aggregate Approximation (SAX) is a popular algorithm in the symbolic methods, but it doesn’t take the form characteristic of sequence into consideration and its description of time series information is incomplete. In this paper, a method for time series based on symbolic form and area difierence is introduced. This method applies the idea of layered in unvaried-time series similarity measure to combine the symbolic method with the area of sequence and coordinate axis, and the similarity can be searched from the rough to the subtle. Ultimately, not only can the overall trend of sequence be matched, but also the goal of fltting can be reached in detail. The experiments show that this method can be used efiectively for time series similarity matching.