会議発表論文

会議発表論文

text

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). 2019, 11707, p. 275-290

Correlation Set Discovery on Time-Series Data

Springer

英語

text

2019-08-06

https://doi.org/10.1007/978-3-030-27618-8_21

This version of the contribution has been accepted for publication, after peer review (when applicable) but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: https://doi.org/10.1007/978-3-030-27618-8_21.

Time-series data analysis is essential in many modern applications, such as financial markets, sensor networks, and data centers, and correlation discovery is a core technique for the analysis. In this paper, we address a novel problem that computes a k-sized time-series dataset where the minimum Pearson correlation of any two time-series in the set is maximized. This problem discovers a group of time-series, which are highly correlated with each other, from a given time-series dataset without any prior knowledge, thus helps many analytical applications. We show that this problem is NP-hard, and design an approximate heuristic solution that provides a high quality result with fast response time. Extensive experiments on real and synthetic datasets verify the efficiency, effectiveness, and scalability of our solution.

© 2019, Springer Nature Switzerland AG.

92852

Correlation Set Discovery on Time-Series Data

Springer

英語

2019-08-06

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

11707

275

290

https://doi.org/10.1007/978-3-030-27618-8_21

https://hdl.handle.net/11094/92852