Estimating \iSAX Parameters for Efficiency

TitleEstimating \iSAX Parameters for Efficiency
Publication TypeConference Paper
Year of Publication2023
AuthorsTsoukalos M, Platis N, Vassilakis C
Conference NameNew Trends in Database and Information Systems (Proceedings of the 2023 European Conference on Advances in Databases and Information Systems / ADBIS 2023)
PublisherSpringer Nature Switzerland
Conference LocationCham
ISBN Number978-3-031-42941-5
KeywordsIndexing, Parameter tuning, SAX, Time series
Abstract

\iSAX is considered one of the most efficient indexes for time series. Several parameters affect the construction of an \iSAX index: the sliding window size, the threshold value, the number of segments and the maximum cardinality, the last two being related to the SAX representation. In this paper (i) we consider the effect of each parameter on the efficiency of the \iSAX index, (ii) we evaluate the importance of each parameter, and, (iii) suggest how to optimize these parameters.

DOI10.1007/978-3-031-42941-5_1