Optimal Construction of Multi-Dimensional Indexes in Time-Series Databases: A Physical Database Design Approach

Similarity search in time-series databases is an operation that finds such data sequences whose changing patterns are similar to that of a query sequence. Typically, it hires the multi-dimensional index for its efficient processing. In order to alleviate the dimensionality curse, a problem in high-dimensional cases, the previous methods for similarity search apply the Discrete Fourier Transform(DFT) to data sequences, and take only the first two or three DFT coefficients for selecting organizing attributes of the multi-dimensional index. Other than this ad-hoc approach, there have been no research efforts on devising a systematic guideline for choosing the best organizing attributes among all the DFT coefficients. This paper first points out the problems occurred in the previous methods, and proposes a novel solution to construct the optimal multi-dimensional index. The proposed method analyzes the characteristics of a target database, and then identifies the organizing attributes having the best discrimination power. Finally, it determines the optimal number of organizing attributes by using a cost model for similarity search. We show the effectiveness of the proposed method through a series of experiments.