Original scientific paper
https://doi.org/10.20532/cit.2020.1004574
Fast Online Similarity Search for Uncertain Time Series
Ruizhe Ma
; Georgia State University, Atlanta, USA
Diwei Zheng
; Nanjing University of Aeronautics and Astronautics, Nanjing, China
Li Yan
; Nanjing University of Aeronautics and Astronautics, Nanjing, China
Abstract
To achieve fast retrieval of online data, it is needed for the retrieval algorithm to increase throughput while reducing latency. Based on the traditional online processing algorithm for time series data, we propose a spatial index structure that can be updated and searched quickly in a real-time environment. At the same time, we introduce an adaptive segmentation method to divide the space corresponding to nodes. Unlike traditional retrieval algorithms, for uncertain time series, the distance threshold used for screening will dynamically change due to noise during the search process. Extensive experiments are conducted to compare the accuracy of the query results and the timeliness of the algorithm. The results show that the index structure proposed in this paper has better efficiency while maintaining a similar true positive ratio.
Keywords
time series, uncertainty, online similarity search, index
Hrčak ID:
240968
URI
Publication date:
10.7.2020.
Visits: 1.175 *