CN113922823A - Social media information propagation graph data compression method based on constraint sparse representation - Google Patents
Social media information propagation graph data compression method based on constraint sparse representation Download PDFInfo
- Publication number
- CN113922823A CN113922823A CN202111270061.6A CN202111270061A CN113922823A CN 113922823 A CN113922823 A CN 113922823A CN 202111270061 A CN202111270061 A CN 202111270061A CN 113922823 A CN113922823 A CN 113922823A
- Authority
- CN
- China
- Prior art keywords
- network
- matrix
- atoms
- atom
- node
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
- H03M7/30—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a social media information propagation graph data compression method based on constraint sparse representation, which comprises the following steps of: s1, data sampling: sampling each node of a propagation network to obtain a self-centering network of each node in the network; s2, preprocessing data: expressing the self-centering network into a form of an adjacent matrix, and obtaining a dictionary matrix and a sparse code matrix; s3, extracting atoms representing the network from the dictionary matrix; s4, performing high-frequency atom analysis according to the sparse code matrix; s5, selecting high-frequency atoms to construct a preparation set; s6, matching in the network by using the self-centering network in the alternative set; and S7, reconstructing the network and analyzing the result. The method takes the sparse representation result of the complex network as guidance, and performs constraint representation by combining the characteristics of the social network propagation network, thereby realizing the compression of the social media information propagation graph data and achieving the purpose of recovering the original social network graph data structure by using a small number of atoms.
Description
Technical Field
The invention relates to a social media information propagation graph data compression method based on constraint sparse representation.
Background
With the development of the internet, social network platforms are rapidly developing. No matter the people obtain information, share life interests or communicate with people, the social network cannot be separated, and becomes an indispensable important part of the life of the people.
Social media with a large amount of users are many today, such as Twitter, Facebook, microblog, WeChat, etc. Twitter is one of global social media with the widest popularity and the largest number of users at present, supports 33 language versions, has more than 5 hundred million users, and is widely gathered in users of various countries, grades and professions. The social media, which is popular and broiled, is the main channel for public to freely publish opinions. The Twitter limits the number of words to be pushed which can be published by each user to 140 words, supports other mobile devices such as mobile phones, pc and tablets, and enables the user to access the Twitter anytime and anywhere, acquire the latest information at the first time and quickly and conveniently release the information. Twitter combines information with social network services, changes the propagation mode, associates users together, and forms a huge social relationship network in which information flows.
The social network data preprocessing is to process the data acquired by data acquisition so as to facilitate storage. In reality, most data are incomplete, inconsistent and redundant dirty data before processing, and if the data are directly stored and subsequently processed, the storage burden is increased, and meanwhile, the experimental result is seriously influenced. Currently, data preprocessing mainly focuses on filling up missing data and processing redundant data.
Social network propagation networks are generally large in scale, and structural characteristics of the social network propagation networks cannot be found through manual observation. Although the propagation network can be described to a certain extent by network parameters such as the average degree, the average path length, the average clustering coefficient, the number of nodes, the number of edges and the like, the obvious structural features frequently appearing in the network cannot be found.
This problem can be solved using sparse representation techniques of graph data structures. The technique may decompose a large network having a graph data structure, extracting atoms therein as significant local structures, thereby representing the original network as a linear combination of atoms. In the sparse representation of the complex network, the concept of the self-centering network is used, and when the overall network is not focused, but the characteristic of a single node is researched, the self-centering network is used, the network node consists of only one central node and the neighbor of the node, and the edge only comprises the edge between the central node and the neighbor and the edge between the neighbor and the neighbor. The egocentric network can represent structural information of a node.
The sparse representation of the graph data structure is used for sampling and decomposing an original graph structure, and simultaneously, the sparse representation is used for extracting the atomic structure of the network, so that the network structure characteristics can be described, and meanwhile, the atoms extracted by the sparse representation can be used for analyzing the network structure, such as network similarity measurement, network classification and identification. The technology can be used for characterizing a complex network, but the generated atoms are too many, and the actual times of many atoms in the original network are less, namely the atoms are not high-frequency atoms and do not contribute much to the whole network. For these atoms, certain storage cost is required, and the representation accuracy of these atoms is not enough, which is not favorable for the analysis of the subsequent network structure. In addition, the characterization technique cannot clearly indicate which atoms a certain local structure in the original propagation network is composed of, which is not favorable for structural analysis of the network.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a data compression technology for a social media information propagation graph, wherein the data compression technology is used for guiding the sparse representation result of a graph data structure and combining a social network propagation network to carry out constraint representation so as to reduce the network storage space and achieve the purpose of recovering the original social network graph data structure by using a small number of atoms.
The purpose of the invention is realized by the following technical scheme: the social media information propagation graph data compression method based on constraint sparse representation comprises the following steps:
s1, data sampling: sampling each node of a propagation network to obtain a self-centering network of each node in the network;
s2, preprocessing data: expressing the self-centering network into a form of an adjacent matrix, and obtaining a dictionary matrix and a sparse code matrix;
s3, extracting atoms representing the network from the dictionary matrix;
s4, performing high-frequency atom analysis according to the sparse code matrix;
s5, selecting high-frequency atoms to construct a preparation set;
s6, matching in the network by using the self-centering network in the alternative set;
and S7, reconstructing the network and analyzing the result.
Further, the specific implementation method of step S2 is as follows:
s21, representing the self-centering network into a form of an adjacent matrix;
s22, processing each adjacent matrix, and connecting all column vectors of the matrix end to end in sequence to form a column vector;
s23, splicing all the column vectors together to form a sampling matrix, wherein the column number of the sampling matrix is the number of nodes in the propagation network;
and S24, decomposing the sampling matrix by using a KSVD algorithm to obtain a dictionary matrix and a sparse code matrix.
Further, the specific implementation method of step S3 is as follows: processing each column vector in the dictionary matrix, restoring the column vector into an adjacency matrix, restoring the adjacency matrix into a network after obtaining a plurality of adjacency matrices, and removing redundant isomers in the adjacency matrix, wherein finally the obtained atoms are the atoms for representing the network.
Further, the specific implementation method of step S4 is as follows: using atoms and a dictionary column matrix and a sparse code matrix generated in the process of matrix decomposition by a KSVD algorithm to obtain the following mapping relation:
(1) mapping relation between atoms and column vectors of a dictionary matrix;
(2) mapping relation between the column vector of the dictionary matrix and the row vector of the sparse code matrix;
fusing the two to obtain a mapping relation between atoms and sparse code matrix row vectors;
the sparse code matrix represents the use condition of the column vector of the dictionary matrix in the original matrix, and the use times of atoms can be obtained by counting the numerical sum of the row vector of the sparse code matrix mapped by each atom; calculating the use times of atoms, sequencing and filtering the atoms according to the use times, selecting some atoms which have the highest contribution to the original network, and forming a main structure of the network by the selected atoms;
for propagation networks, where the atomic proportion of the star shape is fairly high, star atoms are a major feature of the propagation network.
Further, the specific implementation method of step S5 is as follows: and constructing 18 star-shaped self-center networks with the node number of 3-20 as an alternative set for constraint atom generation.
Further, the specific implementation method of step S6 is as follows:
s61, constructing an atom use set and an atom node sequence set as storage of matching characterization results;
s62, taking the self-centering network with the maximum number of nodes from the alternative set as a matching network;
s63, in the transmission network, obtaining the degrees of all nodes, sorting the nodes with the degrees more than or equal to the matching network node number-1 according to the sequence from big to small, and taking out the nodes to wait for matching;
s64, sequentially matching the nodes to be matched, overlapping the central node of the matching network with the nodes to be matched, and aligning the neighbor nodes in a sequence from small to large; sequentially recording the aligned first 19 node serial numbers in an atom node sequence set, and simultaneously writing the size of the current matching network into an atom use set; deleting the edge between the point to be matched and the recorded neighbor;
s65, repeating the operation of S64 until no node to be matched with the conformity more than or equal to the matching network node number-1 exists;
and S66, returning to the step S62 until the alternative set is empty.
Further, the specific implementation method of step S7 is as follows: taking the atom use set and the atom node sequence set as the input of the reconstruction network; traversing the atom use set, and drawing an atom network according to the corresponding atom node sequence; meanwhile, atoms with different node numbers are marked in different colors, so that the composition condition of taking the atoms as units is shown in the original network; selecting nodes with high repeated occurrence frequency from the high-frequency atoms, and observing the same attributes of the nodes; and calculating the ratio of the space size of the atom set, the atom use set and the atom node sequence set to the space size of the original network adjacency matrix, namely the compression ratio of the compression method.
The invention has the beneficial effects that: the invention provides propagation map representation under constraint, and develops a propagation map compression method based on the guidance of a complex network sparse representation result. Because the number of finally generated atoms is less, the compression rate is high, and the atomic composition condition of the original network can be accurately represented, the following beneficial effects can be brought:
(1) the compression of social information transmission network data storage is facilitated;
(2) compared with the sparse representation of a complex network, the number of generated atoms is less for a specific social information propagation network, the method has more characteristic significance, and the method is beneficial to downstream application to carry out network structure analysis;
(3) the composition structure of the original propagation network with the nodes as units is converted into the composition structure with the atoms as units, so that the composition of the social information propagation network can be further researched.
Drawings
FIG. 1 is a flowchart of a social media information propagation graph data compression method based on constraint sparse representation according to the present invention.
Detailed Description
The technical scheme of the invention is further explained by combining the attached drawings.
The invention takes the representation result of the complex network sparse representation technology to the propagation network as guidance, constructs a corresponding structure matching algorithm and researches the structure composition of the original propagation network. As shown in FIG. 1, the social media information propagation graph data compression method based on constraint sparse representation of the invention includes the following steps:
s1, data sampling: sampling each node of the propagation network to obtain a self-centering network of each node in the network, so as to obtain a self-centering network set of the propagation network;
s2, preprocessing data: expressing the self-centering network into a form of an adjacent matrix, and obtaining a dictionary matrix and a sparse code matrix; the specific implementation method comprises the following steps:
s21, representing the self-centering network into an adjacent matrix form, thereby converting the original abstract network into an actual matrix form and facilitating the subsequent mathematical operation; the adjacency matrix is a representation of the social networking graph data structure. For a network of N nodes, the representation is performed using an N-size adjacency matrix, where aijThe value of (b) may be 0 or 1. A when the node i and the node j have connecting edgesijIs 1, otherwise is 0. Representing the network as a contiguous matrix facilitates subsequent mathematical operations.
S22, processing each adjacent matrix, and connecting all column vectors of the matrix end to end in sequence to form a column vector; e.g., a matrix of dimension (m, n), which is processed to convert to column vectors of (m x n, 1);
s23, splicing all the column vectors together to form a sampling matrix, wherein the column number of the sampling matrix is the number of nodes in the propagation network;
and S24, decomposing the sampling matrix by using a KSVD algorithm to obtain a dictionary matrix and a sparse code matrix.
S3, extracting atoms representing the network from the dictionary matrix; the specific implementation method comprises the following steps: processing each column vector in the dictionary matrix, restoring the column vector into an adjacency matrix, restoring the adjacency matrix into a network after obtaining a plurality of adjacency matrices, and removing redundant isomers in the adjacency matrix, wherein finally the obtained atoms are the atoms for representing the network.
Through the steps, the social information propagation network is characterized by using a sparse characterization technology of a complex network, and atoms of the propagation network are obtained. The method can find the remarkable local characteristics of the network and acquire the atoms of the original propagation network so as to guide the follow-up steps to be carried out.
S4, performing high-frequency atom analysis according to the sparse code matrix; the specific implementation method comprises the following steps: using atoms and a dictionary column matrix and a sparse code matrix generated in the process of matrix decomposition by a KSVD algorithm to obtain the following mapping relation:
(1) mapping relation between atoms and column vectors of a dictionary matrix;
(2) mapping relation between the column vector of the dictionary matrix and the row vector of the sparse code matrix;
fusing the two to obtain a mapping relation between atoms and sparse code matrix row vectors;
the sparse code matrix represents the use condition of the column vector of the dictionary matrix in the original matrix, and the use times of atoms can be obtained by counting the numerical sum of the row vector of the sparse code matrix mapped by each atom; calculating the use times of atoms, sequencing and filtering the atoms according to the use times, selecting some atoms which have the highest contribution to the original network, and forming a main structure of the network by the selected atoms; the general structure of the network can be composed only by these atoms, which are collectively referred to as the main feature of the network in the present invention.
By performing this step of analysis on the propagation network, for the propagation network in which the atom proportion of the star shape is considerably high, it can be said that the star atom is the main feature of the propagation network. A star atom is a self-centric network, except that all neighbors of a central node have no connecting edges to each other.
S5, selecting high-frequency atoms to construct a preparation set;
for a propagation network, under the guidance of a complex network sparse representation result, the composition condition of the network can be described by adopting star-shaped atoms without using other structures.
The method aims to perform characterization under constraint on an original propagation network, uses star-shaped atoms under the guidance of a complex network sparse characterization result as a candidate set, and constrains atom generation in the characterization process to make the star-shaped atoms. In the sparse representation technology of the complex network, the generation of atoms is not constrained, atoms are generated in iteration according to the actual situation of a sampling matrix, however, the actual use frequency of the generated large number of atoms is not much, and the contribution to the original network is not large.
In the characterization process, the generation of atoms is restricted, so that the atoms can be generated only from a pre-constructed alternative set. In order to control the size of an atom without causing the atom to expand indefinitely, the network size of the atom is specified to be 20 or less in total number of nodes. Thus, 18 star-type self-centering networks with nodes of 3-20 are constructed as an alternative set for constraint atom generation.
S6, matching in the network by using the self-centering network in the alternative set;
in the constraint characterization step, the invention does not adopt a way of representing each sampled self-centered network by using a adjacency matrix, although the method is quite effective in the sparse characterization of the complex network. The reason is that the number of adjacent matrixes of different isomers of the same network is huge in scale, and especially when the size upper limit of a specified atom is 20, the different adjacent matrixes are more in representation form. This can introduce significant time complexity that impacts the practical usefulness of the present invention.
The invention does not adopt the adjacency matrix to represent the network, but directly uses the network itself, and uses the star type self-center network in the alternative set to match in the network, thereby gradually decomposing the network into a plurality of different combinations of the star type self-center network and simultaneously generating corresponding star type atoms. The specific implementation method comprises the following steps:
s61, constructing an atom use set and an atom node sequence set as storage of matching characterization results;
s62, taking the self-centering network with the maximum number of nodes from the alternative set as a matching network;
s63, in the transmission network, obtaining the degrees of all nodes, sorting the nodes with the degrees more than or equal to the matching network node number-1 according to the sequence from big to small, and taking out the nodes to wait for matching;
s64, sequentially matching the nodes to be matched, overlapping the central node of the matching network with the nodes to be matched, and aligning the neighbor nodes in a sequence from small to large; sequentially recording the aligned first 19 node serial numbers in an atom node sequence set, and simultaneously writing the size of the current matching network into an atom use set; deleting the edge between the point to be matched and the recorded neighbor;
s65, repeating the operation of S64 until no node to be matched with the conformity more than or equal to the matching network node number-1 exists;
and S66, returning to the step S62 until the alternative set is empty.
S7, reconstructing the network and analyzing the result; the specific implementation method comprises the following steps: taking the atom use set and the atom node sequence set as the input of the reconstruction network; traversing the atom use set, and drawing an atom network according to the corresponding atom node sequence; meanwhile, atoms with different node numbers are marked in different colors, so that the composition condition of taking the atoms as units is shown in the original network; selecting nodes with high repeated occurrence frequency from the high-frequency atoms, wherein the nodes have high influence on the network, and observing the same attributes of the nodes can provide more advanced understanding for the network condition; and calculating the ratio of the space size of the atom set, the atom use set and the atom node sequence set to the space size of the original network adjacency matrix, namely the compression ratio of the compression method. Through the steps, the original propagation network composition structure taking the nodes as the units is converted into the composition structure taking the atoms as the units. It will be appreciated by those of ordinary skill in the art that the embodiments described herein are intended to assist the reader in understanding the principles of the invention and are to be construed as being without limitation to such specifically recited embodiments and examples. Those skilled in the art can make various other specific changes and combinations based on the teachings of the present invention without departing from the spirit of the invention, and these changes and combinations are within the scope of the invention.
Claims (7)
1. The social media information propagation graph data compression method based on constraint sparse representation is characterized by comprising the following steps of:
s1, data sampling: sampling each node of a propagation network to obtain a self-centering network of each node in the network;
s2, preprocessing data: expressing the self-centering network into a form of an adjacent matrix, and obtaining a dictionary matrix and a sparse code matrix;
s3, extracting atoms representing the network from the dictionary matrix;
s4, performing high-frequency atom analysis according to the sparse code matrix;
s5, selecting high-frequency atoms to construct a preparation set;
s6, matching in the network by using the self-centering network in the alternative set;
and S7, reconstructing the network and analyzing the result.
2. The social media information propagation graph data compression method based on constraint sparse representation as claimed in claim 1, wherein the step S2 is specifically implemented by:
s21, representing the self-centering network into a form of an adjacent matrix;
s22, processing each adjacent matrix, and connecting all column vectors of the matrix end to end in sequence to form a column vector;
s23, splicing all the column vectors together to form a sampling matrix, wherein the column number of the sampling matrix is the number of nodes in the propagation network;
and S24, decomposing the sampling matrix by using a KSVD algorithm to obtain a dictionary matrix and a sparse code matrix.
3. The social media information propagation graph data compression method based on constraint sparse representation as claimed in claim 1, wherein the step S3 is specifically implemented by: processing each column vector in the dictionary matrix, restoring the column vector into an adjacency matrix, restoring the adjacency matrix into a network after obtaining a plurality of adjacency matrices, and removing redundant isomers in the adjacency matrix, wherein finally the obtained atoms are the atoms for representing the network.
4. The social media information propagation graph data compression method based on constraint sparse representation as claimed in claim 1, wherein the step S4 is specifically implemented by: using atoms and a dictionary column matrix and a sparse code matrix generated in the process of matrix decomposition by a KSVD algorithm to obtain the following mapping relation:
(1) mapping relation between atoms and column vectors of a dictionary matrix;
(2) mapping relation between the column vector of the dictionary matrix and the row vector of the sparse code matrix;
fusing the two to obtain a mapping relation between atoms and sparse code matrix row vectors;
the sparse code matrix represents the use condition of the column vector of the dictionary matrix in the original matrix, and the use times of atoms can be obtained by counting the numerical sum of the row vector of the sparse code matrix mapped by each atom; calculating the use times of atoms, sequencing and filtering the atoms according to the use times, selecting some atoms which have the highest contribution to the original network, and forming a main structure of the network by the selected atoms;
for the propagation network, star-shaped atoms are adopted as the main characteristics of the propagation network.
5. The social media information propagation graph data compression method based on constraint sparse representation as claimed in claim 1, wherein the step S5 is specifically implemented by: and constructing 18 star-shaped self-center networks with the node number of 3-20 as an alternative set for constraint atom generation.
6. The social media information propagation graph data compression method based on constraint sparse representation as claimed in claim 1, wherein the step S6 is specifically implemented by:
s61, constructing an atom use set and an atom node sequence set as storage of matching characterization results;
s62, taking the self-centering network with the maximum number of nodes from the alternative set as a matching network;
s63, in the transmission network, obtaining the degrees of all nodes, sorting the nodes with the degrees more than or equal to the matching network node number-1 according to the sequence from big to small, and taking out the nodes to wait for matching;
s64, sequentially matching the nodes to be matched, overlapping the central node of the matching network with the nodes to be matched, and aligning the neighbor nodes in a sequence from small to large; sequentially recording the aligned first 19 node serial numbers in an atom node sequence set, and simultaneously writing the size of the current matching network into an atom use set; deleting the edge between the point to be matched and the recorded neighbor;
s65, repeating the operation of S64 until no node to be matched with the conformity more than or equal to the matching network node number-1 exists;
and S66, returning to the step S62 until the alternative set is empty.
7. The social media information propagation graph data compression method based on constraint sparse representation as claimed in claim 6, wherein the step S7 is specifically implemented by: taking the atom use set and the atom node sequence set as the input of the reconstruction network; traversing the atom use set, and drawing an atom network according to the corresponding atom node sequence; meanwhile, atoms with different node numbers are marked in different colors, so that the composition condition of taking the atoms as units is shown in the original network;
selecting nodes with high repeated occurrence frequency from the high-frequency atoms, and observing the same attributes of the nodes;
and calculating the ratio of the space size of the atom set, the atom use set and the atom node sequence set to the space size of the original network adjacency matrix to obtain the compression rate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111270061.6A CN113922823B (en) | 2021-10-29 | 2021-10-29 | Social media information propagation graph data compression method based on constraint sparse representation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111270061.6A CN113922823B (en) | 2021-10-29 | 2021-10-29 | Social media information propagation graph data compression method based on constraint sparse representation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113922823A true CN113922823A (en) | 2022-01-11 |
| CN113922823B CN113922823B (en) | 2023-04-21 |
Family
ID=79243557
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111270061.6A Active CN113922823B (en) | 2021-10-29 | 2021-10-29 | Social media information propagation graph data compression method based on constraint sparse representation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113922823B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114826278A (en) * | 2022-04-25 | 2022-07-29 | 电子科技大学 | Image data compression method based on Boolean matrix decomposition |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080170623A1 (en) * | 2005-04-04 | 2008-07-17 | Technion Resaerch And Development Foundation Ltd. | System and Method For Designing of Dictionaries For Sparse Representation |
| US20130191425A1 (en) * | 2012-01-20 | 2013-07-25 | Fatih Porikli | Method for Recovering Low-Rank Matrices and Subspaces from Data in High-Dimensional Matrices |
| CN104331913A (en) * | 2014-11-19 | 2015-02-04 | 西安电子科技大学 | Polarized SAR polarization method based on sparse K-SVD (Singular Value Decomposition) |
| CN105430416A (en) * | 2015-12-04 | 2016-03-23 | 四川大学 | Fingerprint image compression method based on adaptive sparse domain coding |
| CN105811993A (en) * | 2016-03-11 | 2016-07-27 | 北京航空航天大学 | Data collection method based on compression dictionary learning in wireless sensor network |
| US20160259826A1 (en) * | 2015-03-02 | 2016-09-08 | International Business Machines Corporation | Parallelized Hybrid Sparse Matrix Representations for Performing Personalized Content Ranking |
| CN107170020A (en) * | 2017-06-06 | 2017-09-15 | 西北工业大学 | Dictionary learning still image compression method based on minimum quantization error criterion |
| WO2019006835A1 (en) * | 2017-07-06 | 2019-01-10 | 清华大学深圳研究生院 | Target recognition method based on compressed sensing |
| US20190095787A1 (en) * | 2017-09-27 | 2019-03-28 | Hsiang Tsung Kung | Sparse coding based classification |
| CN110135488A (en) * | 2019-05-10 | 2019-08-16 | 南京邮电大学 | Merge the high-quality compression method of data of dictionary training and observing matrix optimization |
-
2021
- 2021-10-29 CN CN202111270061.6A patent/CN113922823B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080170623A1 (en) * | 2005-04-04 | 2008-07-17 | Technion Resaerch And Development Foundation Ltd. | System and Method For Designing of Dictionaries For Sparse Representation |
| US20130191425A1 (en) * | 2012-01-20 | 2013-07-25 | Fatih Porikli | Method for Recovering Low-Rank Matrices and Subspaces from Data in High-Dimensional Matrices |
| CN104331913A (en) * | 2014-11-19 | 2015-02-04 | 西安电子科技大学 | Polarized SAR polarization method based on sparse K-SVD (Singular Value Decomposition) |
| US20160259826A1 (en) * | 2015-03-02 | 2016-09-08 | International Business Machines Corporation | Parallelized Hybrid Sparse Matrix Representations for Performing Personalized Content Ranking |
| CN105430416A (en) * | 2015-12-04 | 2016-03-23 | 四川大学 | Fingerprint image compression method based on adaptive sparse domain coding |
| CN105811993A (en) * | 2016-03-11 | 2016-07-27 | 北京航空航天大学 | Data collection method based on compression dictionary learning in wireless sensor network |
| CN107170020A (en) * | 2017-06-06 | 2017-09-15 | 西北工业大学 | Dictionary learning still image compression method based on minimum quantization error criterion |
| WO2019006835A1 (en) * | 2017-07-06 | 2019-01-10 | 清华大学深圳研究生院 | Target recognition method based on compressed sensing |
| US20190095787A1 (en) * | 2017-09-27 | 2019-03-28 | Hsiang Tsung Kung | Sparse coding based classification |
| CN110135488A (en) * | 2019-05-10 | 2019-08-16 | 南京邮电大学 | Merge the high-quality compression method of data of dictionary training and observing matrix optimization |
Non-Patent Citations (3)
| Title |
|---|
| DINGDE JIANG等: "A Compressive Sensing-Based Approach to End-to-End Network Traffic Reconstruction" * |
| 潘梦阳: "基于布尔矩阵分解的知识图谱稀疏表征方法研究" * |
| 翟学萌: "多类型复杂网络的随机模型构建与分析方法研究" * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114826278A (en) * | 2022-04-25 | 2022-07-29 | 电子科技大学 | Image data compression method based on Boolean matrix decomposition |
| CN114826278B (en) * | 2022-04-25 | 2023-04-28 | 电子科技大学 | Graph data compression method based on Boolean matrix decomposition |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113922823B (en) | 2023-04-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | User profile preserving social network embedding | |
| CN110569353B (en) | Attention mechanism-based Bi-LSTM label recommendation method | |
| CN109960810B (en) | Entity alignment method and device | |
| CN104317801B (en) | A kind of Data clean system and method towards big data | |
| US8171029B2 (en) | Automatic generation of ontologies using word affinities | |
| CN111950594A (en) | Unsupervised graph representation learning method and unsupervised graph representation learning device on large-scale attribute graph based on sub-graph sampling | |
| Li et al. | Multi-level network embedding with boosted low-rank matrix approximation | |
| CN103793501B (en) | Based on the theme Combo discovering method of social networks | |
| CN112966091B (en) | Knowledge map recommendation system fusing entity information and heat | |
| CN108073576A (en) | Intelligent search method, searcher and search engine system | |
| CN113626723B (en) | Attribute graph community searching method and system based on representation learning | |
| CN111078876A (en) | Short text classification method and system based on multi-model integration | |
| CN107368534A (en) | A kind of method for predicting social network user attribute | |
| US20170034111A1 (en) | Method and Apparatus for Determining Key Social Information | |
| CN110929022A (en) | Text abstract generation method and system | |
| CN111460277B (en) | Personalized recommendation method based on mobile social network tree-shaped transmission path | |
| CN112860685A (en) | Automatic recommendation of analysis of data sets | |
| CN113922823A (en) | Social media information propagation graph data compression method based on constraint sparse representation | |
| CN117251779A (en) | Node classification method based on global perception neural network | |
| CN114461943A (en) | Deep learning-based multi-source POI semantic matching method and device and storage medium thereof | |
| CN113987126A (en) | Retrieval method and device based on knowledge graph | |
| CN110175289B (en) | Mixed recommendation method based on cosine similarity collaborative filtering | |
| CN116956289A (en) | Method for dynamically adjusting potential blacklist and blacklist | |
| CN109582953B (en) | Data support scoring method and equipment for information and storage medium | |
| CN110956199A (en) | Node classification method based on sampling subgraph network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |