Unlock instant, AI-driven research and patent intelligence for your innovation.

A Deep Hashing-Based Embedding Method for Symbolic Social Networks

A social network and hash technology, applied in the field of social media in computer natural language processing technology, can solve the problem of loss of negative side information and other issues

Inactive Publication Date: 2021-04-27
BEIJING INSTITUTE OF TECHNOLOGYGY +1
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to solve the problem of negative side information loss when the hash-based network embedding method processes the marked social network. In order to improve the effect of the marked social network analysis task, a symbolic social network based on deep hash is proposed. Embedding method, a network embedding that can preserve both positive and negative edge information

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A Deep Hashing-Based Embedding Method for Symbolic Social Networks
  • A Deep Hashing-Based Embedding Method for Symbolic Social Networks
  • A Deep Hashing-Based Embedding Method for Symbolic Social Networks

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0045] This embodiment describes the process adopted by the present invention when processing public data sets, the structure and detailed parameters of the involved deep hash model, and the experimental results.

[0046] A symbolic social network embedding method based on deep hashing. From the public data set, the dense network Epinions and the sparse network Slashdot are respectively selected as the experimental data set, and the following processes are performed in the order of data collection phase, training phase and testing phase. :

[0047] (1) Data collection stage.

[0048] First, take the dense network Epinions experimental data set as the object, and perform the following processing:

[0049] Step A: Extract each edge and edge weight from the experimental data set to form an edge set ε;

[0050] Among them, edge set ε={(v i ,v j ,e ij )}, v i , Represents the nodes in the network, the edge weight e ij ∈{1,-1} means positive and negative edges, the number o...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention proposes a symbolic social network embedding method based on deep hashing, which belongs to the technical field of network social media. This method uses triplets to simultaneously store positive and negative edge information in the network, and designs an end-to-end deep hash model to train triplet data, and finally obtains the Hamming space of each node in the labeled social network. The hash code representation in . The quality of the network embedding method is evaluated by the AUC value of the corresponding hash code in the link prediction task. Compared with the existing technology, the method of the present invention applies deep hashing to the network embedding task for the first time, and can simultaneously consider the positive edge and negative edge information in the tagged network, and perform the connection prediction task in two real tagged social networks, which significantly improves the tagged social network. Effects on network analysis tasks.

Description

technical field [0001] The invention relates to a symbolic social network embedding method, in particular to a symbolic social network embedding method based on deep hashing, and belongs to the field of social media in computer natural language processing technology. Background technique [0002] The network is an important form of expressing the relationship between objects and objects. A key issue in the analysis and research of the network is to study how to reasonably represent the characteristic information in the network. Network representation is a bridge connecting network raw data and network application tasks. Network representation learning is responsible for learning the vector representation of each node in the network from network data, and then these node representations can be used as node features for subsequent network application tasks. [0003] With the continuous development of social media represented by WeChat, Weibo, and Facebook, network representat...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06Q50/00
CPCG06Q50/01
Inventor 毛先领郭佳楠姜晓健孙英翔黄河燕牟其林邹佳
Owner BEIJING INSTITUTE OF TECHNOLOGYGY