SUBSEQUENCE MATCHING METHOD USING DUALITY OF WINDOW CONFIGURATION IN TIME SERIES DATABASE

摘要

PURPOSE: A subsequence matching method using a duality of a window configuration in a time series database is provided to reduce an erroneous answer, and to improve a performance by using a duality of a window configuration which divides a data sequence into a disjoint window and a query sequence into a sliding window. CONSTITUTION: An f-dimensional index is generated and initialized(S301). A data sequence is read from a database to a memory(S302). The data sequence read is divided into a disjoint window(S303). Each disjoint window is converted into an f-dimensional point(S304). A record of a converted point, a data sequence identifier and a window starting position is formed(S305). A point into which the disjoint window is converted is made as a key. In addition, the record is inserted into the f-dimensional index(S306). It is judged whether a present sequence is the last sequence(S307). A next data sequence is read from the database(S308).