A nearest neighbor query method for searching objects with time and location informations based on spatiotemporal similarity

It is crucial for users to find lots of information with time and geographical tags on the Internet. Therefore, a nearest neighbor query method called STR-kNN is proposed. Using the method of calculating the spatiotemporal similarity between two objects, the spatiotemporal variables of the data obje...

Full description

Saved in:
Bibliographic Details
Published in:Evolutionary intelligence 2024, Vol.17 (4), p.3031-3041
Main Authors: Qian, Shenyi, Tian, Ziqiao
Format: Article
Language:English
Subjects:
Algorithms
Applications of Mathematics
Artificial Intelligence
Bioinformatics
Control
Engineering
Mathematical and Computational Engineering
Mechatronics
Queries
Query processing
Research Paper
Robotics
Search algorithms
Similarity
Spatiotemporal data
Statistical Physics and Dynamical Systems
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:It is crucial for users to find lots of information with time and geographical tags on the Internet. Therefore, a nearest neighbor query method called STR-kNN is proposed. Using the method of calculating the spatiotemporal similarity between two objects, the spatiotemporal variables of the data object are normalized and mapped to the three-dimensional space. The distance similarity between two data objects in the three-dimensional space is used to approximate their actual spatiotemporal similarity. In this way, a three-dimensional STR-tree index is built for the data object in the three-dimensional space. This index can effectively combine the spatial–temporal variables of the data object, ensuring that each data object is traversed no more than once during query processing. Finally, an accurate searching algorithm of STR-kNN is designed to find up to the top-k query results through a single calculation. In the experiment, when the data volume is 6 million, the query times of STR-kNN, 3DR-k NN, and RT-k NN algorithms are 12 ms, 43 ms, and 55 ms respectively. When k is taken as 50, the query times of the three algorithms are 12 ms, 43 ms, and 55 ms, respectively, indicating that the new algorithm can greatly improve the query efficiency.
ISSN:1864-5909
1864-5917
DOI:10.1007/s12065-024-00926-7