Dimensionality Reduction for Fast Similarity Search in Large Time Series

Speaker: Sindy SHOU

Abstract

The problem of similarity search in large time series databases has attracted much attention recently. It is a non-trivial problem because of inherent high dimensionality of the data. The most promising solutions involve performing dimensionality reduction on the data, then indexing the reduced data with a spatial access method. The paper introduces a new dimensionality reduction technique called PAA (piecewise Aggregate Approximation), which shows its superiority to other existing dimensionality reduction approaches.