Database high-dimensional data query method based on three-input network and commanding height tree

The invention provides a database high-dimensional data query method based on a three-input network and a commanding height tree, which comprises the following steps of: constructing the three-input network, extracting a standard sample from a data set for training the three-input network, and const...

Full description

Saved in:
Bibliographic Details
Main Authors BAI YUNHAI, HE JIAYU, PEI YUXUAN
Format Patent
LanguageChinese
English
Published 28.05.2024
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:The invention provides a database high-dimensional data query method based on a three-input network and a commanding height tree, which comprises the following steps of: constructing the three-input network, extracting a standard sample from a data set for training the three-input network, and constructing a triple sample according to the standard sample; training a three-input network by using a triple sample, randomly selecting one piece of sample data as a root node, taking the rest of sample data as child nodes, and constructing a commanding height tree; calculating the distance between each child node and a root node in the data set through a three-input network, and storing the distance in a node set; and constructing a corresponding commanding height tree through the node set, and querying through the commanding height tree to obtain the node closest to the target node. The method has the advantages that the three-input network and the commanding height tree are split and combined again, the query accu
Bibliography:Application Number: CN202410488146