CN117349386A - Digital humane application method based on data strength association model - Google Patents

Digital humane application method based on data strength association model Download PDF

Info

Publication number
CN117349386A
CN117349386A CN202311318274.0A CN202311318274A CN117349386A CN 117349386 A CN117349386 A CN 117349386A CN 202311318274 A CN202311318274 A CN 202311318274A CN 117349386 A CN117349386 A CN 117349386A
Authority
CN
China
Prior art keywords
node
nodes
information
relation
weak
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
Application number
CN202311318274.0A
Other languages
Chinese (zh)
Other versions
CN117349386B (en
Inventor
佘雨芙
赵龙霄
潘生林
王新鑫
滕子仪
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jijiu Tianjin Technology Co ltd
Original Assignee
Jijiu Tianjin Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Jijiu Tianjin Technology Co ltd filed Critical Jijiu Tianjin Technology Co ltd
Priority to CN202311318274.0A priority Critical patent/CN117349386B/en
Publication of CN117349386A publication Critical patent/CN117349386A/en
Application granted granted Critical
Publication of CN117349386B publication Critical patent/CN117349386B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/28—Databases characterised by their database models, e.g. relational or object models
    • G06F16/284—Relational databases
    • G06F16/288—Entity relationship models
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
    • G06F16/45—Clustering; Classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
    • G06F16/48—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/483—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/02—Knowledge representation; Symbolic representation
    • G06N5/022—Knowledge engineering; Knowledge acquisition
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Multimedia (AREA)
  • Library & Information Science (AREA)
  • Artificial Intelligence (AREA)
  • Computational Linguistics (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention discloses a digital humane application method based on a data strength association model, which relates to the technical field of digital humane and comprises the following steps: s1, data collection; s2, preprocessing the original data acquired in the step S1; s3, digitizing and information library building; s4, constructing a data relationship; s5, learning reconstruction; s6, converting the entity relationship into a node information network; s7, calculating the strong and weak relation between the nodes; s8, constructing a data strength relation model; s9, constructing and managing node interaction relations; s10, revealing hidden relations. The invention adopts the digital humane application method based on the data strength association model, is based on big data analysis, data mining and machine learning, adopts the data strength association model, combines the technologies of strength relation, dynamic weight regulation and the like, and can more accurately predict the relation between nodes, optimize the information propagation path and improve the utilization efficiency of information.

Description

Digital humane application method based on data strength association model
Technical Field
The invention relates to the technical field of digital humanization, in particular to a digital humanization application method based on a data strength association model.
Background
With the continuous development of the Internet age, innovation and application of digital technology and artificial intelligence have been deep into the mind, not only changing the daily life style of people, but also remodelling the research ecological environment in the human research field. The widespread use of information communication technology and related infrastructure has enabled researchers to more conveniently obtain information, analyze content, and compose text. Digital technology and digital applications convert a large number of real things into information data, which is widely used for retrieval and analysis, providing a more accurate and reliable information source for human and text research.
However, this round of human reformulation is more represented as "human datamation", and, although involving information collection and quantitative studies based on data analysis, is still highly dependent on the manner in which the individual researchers experience reading, subjective analysis, and language interpretation, as well as the authenticity and validity of the data material itself. Therefore, how to objectively judge and accurately push data information is one of the challenges faced by the current humane research.
Disclosure of Invention
The invention aims to provide a digital humane application method based on a data strength association model, which is based on big data analysis, data mining and machine learning, and adopts the data strength association model, and combines the technologies of strength relation, dynamic weight regulation and the like.
In order to achieve the above purpose, the invention provides a digital humane application method based on a data strength association model, which comprises the following steps:
s1, data collection;
s2, preprocessing the original data acquired in the step S1;
s3, digitizing and information library building;
for different types of files, the system needs to conduct context identification to understand the content of the files, metadata information is extracted from each file, object detection and extraction are conducted on images and video files, the extracted metadata information and the content identified by the context are stored in a relational database, a gallery database is established for media files such as pictures and videos, and extracted object information, image features and the like are stored in the gallery database;
s4, constructing a data relationship;
forming a set of data of a database, establishing a knowledge organization model, a person association data set and a data field ontology according to the data set, and finding out data static association;
s5, learning reconstruction;
dynamic knowledge organization based on dimension expansion, semantic label extraction based on entity portraits, and fusion into data dynamic aggregation based on knowledge graphs;
s6, converting the entity relationship into a node information network;
S7, calculating the strong and weak relation between the nodes;
s8, constructing a data strength relation model;
s81, defining a model; s82, defining a total node set, an edge set, a limited information set, a strong-weak relation node set, neighbor nodes and an information network; s83, establishing a process model; combining a process model and a prediction algorithm, and carrying out dynamic regulation and control on the relationship by utilizing a cosine similarity principle and a dynamic weight of the strong-weak relationship to obtain an optimal path of resources necessary for individual development;
s9, constructing and managing node interaction relations;
s10, revealing hidden relations.
Preferably, in step S7, the strong-weak relationship calculation between nodes is performed by the correlation score, the strong relationship value is S, the weak relationship value is w, the correlation score is g, the common history period weight is m,
s＝g×m； (1)
w＝1-(g×m) (2)
the weak relation value is obtained by reducing the strong relation value, the correlation between two exhibits is lower or belongs to different historical periods, the weak relation value is correspondingly higher, and the strong relation threshold value and the weak relation threshold value are set to judge whether the node is in strong association or weak association.
Preferably, in step S81, it is included that there are N network nodes in the information network, where node i is one of them,
defining one, a total node set: i is the total node set, i= {1,2,3,..n }, I, j e I;
Defining two, edge sets: e is a set of relationships for nodes, e= {1,2,3,., N }, E;
defining three, a limited information set: the finite set of information owned by node i is denoted as R i Represented by R i ＝{R i,1 ,...,R i,H H.gtoreq.0, representing the total number of information for node i;
defining four strong and weak relation node sets: the strong relation node set NS (i) is a set of strong relation nodes of node i, and uses binary vector X i,j Representing the relationship between the strength of nodes, if X i,j =1, meaning j e NS (i), i.e. node j is a strong relational node of node i; the weak relational node set NW (i) is a set of weak relational nodes of node i, if X i,j =α, then represents j∈nw (i), where 0<α<1, node j is a weak relationship node of node i;
definition five, neighbor nodes: the nodes connected with the node i by the edges are all neighbor nodes, the number of the neighbor nodes is recorded as K, all the neighbor nodes of the node i form a neighbor node set N (i), wherein N (i) =NS (i) UW (i), and the neighbor nodes of the node j are the neighbor nodes of the i and satisfy the binary vector relation: x is 0 < X i,j Less than or equal to 1, wherein j is N (i);
definition six, information network: the information network is composed of individual nodes and relationships between the nodes, expressed as RG, i.e., rg= { (I, E) |0<X i,j <1,i,j∈I,e∈E}。
Preferably, in step S82, the following steps are included:
S821, initializing strong relation node set NS (i) of node i and limited information set R owned by node i i ；
S822, at time t, the information category of the node i is H, and the information set is And->Respectively representing the quantity of information r owned by a node i and a node j at a moment t, wherein the node i acquires information from an information network, and firstly, the nodes in a strong relation node set NS (i) perform information interaction;
s823, at time t+w, information owned by node i is changed, and the change is influenced by two reasons, namely, a formula
Wherein,representing the finite set of information R at time t+w i,r The variation of ρ is the volatilization coefficient, 0<ρ<1, which means that the amount of information owned by a node gradually decreases with the passage of time, γ means a trust parameter for information change, 0<γ<1，/>Representing the finite set of information at time tR i,r Is an information amount of (a);
as time goes by and its own growth and environment changes, new information is continually incorporated into the new collection,or/>when representing information R i,r Changes occur at time w;
when the node interacts with the strong relation network member, the information of the node i changes, and the formula is as follows
Formula (VI)
Determining whether information owned by different network nodes to be exchanged is the same, and whether the amount of the same information is consistent, wherein phi represents a trust parameter for interaction with a strong relationship node, j is epsilon NS (i), i.e., node j is a strong relationship node of node i;
s824, after t+w time, when node i interacts with the weak relation network node, the node i is added to the weak relation set NW (i) with probability p, information change of the node i is affected by the strong and weak relation, and the formula is that
Where k ε NW (i), i.e. node k is the weak relational node of node i, delta represents the trust parameter for interaction with the weak relational node, where Representing the finite set of information R at time t+w j,r Is used for the information amount of the (a),representing the finite set of information R at time t+w k,r Is an information amount of (a);
s825, judging the stability of the information network at time t+w through standard deviation sigma
Sigma (t) is variance, I is total node set number, I is node, H is limited information set number, r is limited information node, and N is network node number.
Preferably, in step S83, the following steps are included:
s831, constructing a network graph G= (I, E) based on the definition one and the definition two in the step S81, marking a central node and other nodes by using labels, wherein the node I corresponds to the label f (I);
s832, respectively denoted as d according to the radius of the node i relative to the node x and the node y (i,x) And d (i,y) The label of the center node is 1, (d) (i,x) ,d (i,y) ) Node label f (i) =2 corresponding to= (1, 1); (d) (i,x) ,d (i,y) ) Node label f (i) =3 corresponding to (1, 2) or (2, 1), and so on;
two central nodes, where the labels f (i) and the double radius (d (i, x), d (i, y)) satisfy:
d (i,x) +d (i,y) ≠d (j,x) +d (j,y) in the time-course of which the first and second contact surfaces,
d (i,x) +d (i,y) ＝d (j,x) +d (j,y) in the time-course of which the first and second contact surfaces,
generating an integer tag function f according to DRNL algorithm l (i)，
Wherein d x ＝d (i,x) ，d y ＝d (i,y) ，d＝d x -d y For d (i,x) = infinity or d (i,y) The node of = ≡is denoted as empty tag 0, d x ，d y Represents the radius of node i relative to node x and node y, respectively, and d represents d x ，d y Is a difference in (2);
s833, judging the relation between the node i and the nodes x and y according to the radius of the node i relative to the node x and the node y, and when d (i,x) ＝1＝d (i,y) When f (i) =2, the nodes i and x, y have the same relation degree and all show strong relation, otherwise, weak relation; d, d (i,x) ＜d (i,y) When the relation between the node i and the node x is stronger than the relation between the node i and the node y;
s834, reflecting the strength relation between nodes by using a cosine similarity calculation formula, wherein the cosine similarity measures the similarity between individuals according to the cosine value of the included angle of two vectors, the value is between 0 and 1, the closer 1 is to the higher the similarity, the closer 0 is to the lower the similarity, and for multidimensional vectors A (a 1, B1, c 1.) and B (a 2, B2, c2..), a1, B1, c 1.) and a2, B2, c2., the components in different dimensions are represented by specific calculation formulas
Where cos θ represents the cosine inside the trigonometric function, A and B are the two vectors to be compared, A and B represent the norms of A and B, typically Euclidean norms, expressed as itThe moulds A i And B i Representing the i-th element in vector a and vector B respectively,representing the sum of squares of the elements in vector a and vector B, respectively;
s835, because node i generates different strong and weak relations for different central nodes, different dynamic weights are given to each relation.
Preferably, in step S835, each relationship is given a different dynamic weight, which is specifically calculated as follows:
for similarity between node i and node x
For similarity between node i and node y
Wherein w is 1 And w 2 Respectively strong relation and weak relation weights of nodes; w (w) 3 And w 4 The weak relation degree weights between the nodes i and x and y are respectively; k is the number of nodes in the network, k, except f (i) =1 1 Number of nodes, k, of g (i) =2 in the network 2 Divide k by k 1 The specific calculation of the number of external nodes is shown in the formula
The prediction algorithm is combined with similarity and dynamic weight among nodes and judged relation calculation and dynamic weight regulation to determine the possibility of generating strong and weak relations among the nodes, and when the relations among some nodes are not clearly identified, the unidentified hidden relations are revealed by utilizing cosine similarity and dynamic weight, so that an optimal path of resources required by individual development is obtained.
Preferably, in step S9, the following steps are included:
s91, initializing node information; establishing a total node set according to a node information network, and initializing an edge set, a limited information set, a strong-weak relation node set and neighbor nodes by each node;
s92, information interaction and change processes;
in each time step, information interaction is carried out among nodes, the nodes acquire information from strong relation nodes and weak relation nodes, the information interaction among the strong relation nodes is more frequent and dense, the information interaction among the weak relation nodes is less, and the information set of the nodes is updated by using a formula (5)Calculating the variation of information according to the interaction of the node with other nodes>
S93, information change and relation regulation;
calculating the variation of the information in the information set by using the formulas (3) and (4)Over time, the interaction between some nodes is enhanced, the relationship between nodes is changed, and the nodes are added into the strong relationship node set of other nodes with a certain probability, and the nodes are usedThe formula (6) judges whether the information owned by different network nodes to be exchanged is the same or not and whether the quantity of the same information is consistent or not;
s94, adding weak relation nodes;
over time, the interaction between certain nodes is enhanced, the relationship between the nodes is changed, the nodes are added into the weak relationship node set of other nodes with a certain probability, and the formula (7) is used for calculating the change amount of information
S95, evaluating the stability of the information network;
after each time step is finished, the stability of the information network is calculated by using a standard deviation formula, the distribution condition of the information is evaluated by calculating the standard deviation of the information quantity of each node by using a formula (8), the standard deviation is stabilized or reaches the minimum value, and the information network is stabilized.
Preferably, in step S10, the following steps are included:
s101, constructing a network diagram and a node label;
building a relational network graph according to the data set, wherein nodes represent different elements, exhibits and subjects, marking the central node and other nodes by using labels, and calculating the labels of the nodes according to a formula (9);
s102, predicting the possibility of the relationship among the nodes, and revealing unrecognized hidden relationship;
calculating cosine similarity between the node i and the nodes x and y by using a formula (10), quantifying the similarity between the nodes, calculating the similarity between the node i and the nodes x and y according to formulas (11) and (12), and calculating strong relationship and weak relationship weights and weak relationship degree weights between the node i and the nodes x and y by using formulas (13), (14), (15) and (16);
s103, optimizing a propagation and recommendation mechanism of information according to a result of the prediction algorithm;
When the predictive strong relationship score between two nodes is high, the system recommends users to view the content of one of the nodes so that they can learn about the possible relationship, and by optimizing the information propagation, users obtain more relevant and valuable content, thereby enriching their knowledge and experience.
Therefore, the invention adopts the digital humane application method based on the data strength association model, and has the following technical effects:
(1) The reality things are converted into information data through digitization, information database construction, data relation construction and knowledge reconstruction, so that the retrieval and analysis are facilitated;
(2) The entity relationship is converted into a node information network, so that the inherent relationship between the data can be revealed, and researchers are helped to better understand the meaning and relationship of the data;
(3) The strength relation among the nodes is calculated, the association degree among the nodes can be judged, and the accurate pushing of related data information is facilitated;
(4) Establishing a data strong-weak relation model and management node interaction relation, revealing an implicit relation and improving the accuracy and reliability of information;
(5) According to the method, a user is not limited to passively acquiring information, and the requirement of the user on knowledge can be better met by means of strong and weak association of data, so that the utilization efficiency of resources and the accuracy of information are improved.
The technical scheme of the invention is further described in detail through the drawings and the embodiments.
Drawings
FIG. 1 is a flow chart of a digital humane application method based on a data strength association model of the invention;
fig. 2 is a diagram illustrating a strong-weak relationship node set according to the present invention.
Detailed Description
The technical scheme of the invention is further described below through the attached drawings and the embodiments.
Unless defined otherwise, technical or scientific terms used herein should be given the ordinary meaning as understood by one of ordinary skill in the art to which this invention belongs.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.
Furthermore, it should be understood that although the present disclosure describes embodiments, not every embodiment is provided with a separate embodiment, and that this description is provided for clarity only, and that the disclosure is not limited to the embodiments described in detail below, and that the embodiments described in the examples may be combined as appropriate to form other embodiments that will be apparent to those skilled in the art. Such other embodiments are also within the scope of the present invention.
It should also be understood that the above-mentioned embodiments are only for explaining the present invention, the protection scope of the present invention is not limited thereto, and any person skilled in the art should be able to cover the protection scope of the present invention by equally replacing or changing the technical scheme and the inventive concept thereof within the scope of the present invention.
Techniques, methods, and apparatus known to one of ordinary skill in the relevant art may not be discussed in detail, but are intended to be considered part of the specification where appropriate.
The disclosures of the prior art documents cited in the present specification are incorporated by reference in their entirety into the present invention and are therefore part of the present disclosure.
Example 1
As shown in fig. 1, the digital humane application method based on the data strength association model provided by the invention comprises the following steps:
s1, data collection;
the data of works of art, cultural relics, historical culture, science, exhibition and display data and other open data disclosed by the national museum are taken as data sets.
S2, preprocessing data;
errors, deletions and outliers in the dataset are cleared. The data is normalized and formatted for subsequent analysis. Inconsistent naming, date and other identifiers are handled.
S3, digitizing and information library building;
and carrying out context recognition, metadata extraction, object detection and extraction on different format files (pdf, jpg, avi, txt, gif and the like) in the original data set to form a relational database and a gallery database.
S31, identifying the context;
for different types of files, the system needs to perform context recognition to understand the content of the file. For example, for text files (txt), natural language processing techniques (e.g., text classification, named entity recognition, etc.) are used to identify information about the subject, author, etc. of the text; for audio video files (avi), audio-video processing techniques (e.g., audio-to-text, image analysis, etc.) are used to extract content information.
S32, extracting metadata;
for each file, metadata information is extracted from it, which may include file name, size, creation date, modification date, author, keywords, etc. For pdf files, information such as titles, authors, abstracts and the like of the documents can be extracted; for the picture files (jpg, gif), information such as resolution, shooting date, and the like can be extracted.
S33, object detection and extraction;
for image and video files, object detection and extraction may be required in order to identify objects, persons, etc. in the image. Using computer vision techniques such as object detection, face recognition, etc.
S304, establishing a relational database;
the extracted metadata information and context-identified content are stored in a relational database. Different data tables can be established according to file types, and various information can be sorted and stored for subsequent retrieval and management.
S35, establishing a drawing database;
for media files such as pictures and videos, a gallery database can be established, and extracted object information, image features and the like are stored in the gallery database to support image retrieval and similarity matching.
S4, constructing a data relationship;
S41, forming a set of data of a database;
the data in the database is sorted into different sets for better understanding and analysis. These sets include a set of context descriptions for classifying data by time, place, subject, etc.; a set of discrete entities comprising individual entities extracted from a database; and the entity relation set is used for recording the relation between different entities, such as family relation, participation relation and the like.
S42, establishing a knowledge organization model, a person associated data set and a data field body according to the data set;
an organized knowledge model is constructed from the data sets. The model helps to represent structures, such as graphs or knowledge maps, between entities, relationships, and contexts; the person association data set is used for describing association relations among the persons, such as social networks, cooperation relations and the like; and the data field ontology is used for establishing a model of concepts and relations in the field so as to better understand the semantics of the data.
S43, data static association;
by analyzing the entities and relationships in the data set, static associations between different entities can be discovered revealing information and patterns hidden in the data.
S5, learning reconstruction;
s51, dynamic knowledge organization based on dimension expansion;
in knowledge organization, by continuously expanding different dimensions, the organization of knowledge is more comprehensive and multi-angle. This may include collecting and sorting information from different time, place, subject, etc. dimensions, thereby forming a richer knowledge network.
S52, extracting semantic tags based on the entity portraits;
by analyzing the entity (e.g., person, place, event, etc.), a representation of the entity, including its attributes, relationships, etc., is created. Semantic tags are then extracted from these entity representations to describe information about the entity's characteristics, attributes, relationships, etc., to better organize and understand the data.
S53, fusing the data into data dynamic aggregation based on the knowledge graph;
the data of various dimensions are fused into a knowledge graph, which is a graph structure with entities and relations as nodes and is used for representing the association between different entities. This dynamic aggregation means that the knowledge graph can be updated and expanded continuously as new information is added, thereby maintaining an up-to-date presentation of data knowledge.
S6, converting the entity relationship into a node information network;
Knowledge graph is a graphic structure used to represent and organize knowledge, where entities (e.g., people, places, concepts) are connected together by relationships. Node organization networks are also a similar graphical structure that is commonly used to represent relationships between various elements, such as interpersonal relationships, information flows, and the like. And mapping the entities in the knowledge graph into nodes, and mapping the relation between the entities into the connection between the nodes. In this way, various elements, exhibits, topics, historic events, etc. of the digital museum can be represented as nodes and connected by relationship edges to form a node organization network.
S7, calculating the strong and weak relation between the nodes;
for example, the node strength relation calculation of the character relation considers a digital museum, and the exhibits show characters, events and cultures in different historic periods. Nodes may represent exhibits (characters, events, etc.), and the strength relationship may be expressed as a degree of association between the exhibits. We will assume that the calculation of the strong-weak relationship is based on the relationship of the correlation of the exhibit and the historical period.
The relevance score may be a score representing the relevance of an exhibit calculated based on information such as exhibit descriptions, keywords, and the like. For example, if the keywords in two presentation descriptions are similar, the relevance score may be higher. The common historical period weight represents a degree of historical period association between two exhibits. For example, if two shows belong to the same historical period, then the common historical period weight may be higher.
The strong relation value is s, the weak relation value is w, the relevance score is g, and the common historical period weight is m.
s＝g×m； (1)
w＝1-(g×m)； (2)
In this calculation, the weak relationship value is obtained by reducing the strong relationship value. If the correlation between two shows is low or they belong to different historic periods, the weak relationship value will be correspondingly higher.
And judging whether the nodes are strongly associated or weakly associated by setting a strong-weak relation threshold value.
S8, constructing a data strength relation model;
the construction of the data strength relation model focuses on comprehensively considering the strength relation among the nodes, and a dynamic model of information transmission and strength relation is established. The interaction, information change and evolution of strong and weak relations among the nodes are described in detail, and the possible relations among the nodes are predicted through dynamic weight regulation and control and the resource acquisition path is optimized.
S81, defining a model;
it is assumed that there are N network nodes in the information network, of which node i is one.
Defining one, a total node set: i is the total node set, i= {1,2,3,..n }, I, j e I.
Defining two, edge sets: e is a set of relationships for nodes, e= {1,2,3,..n }, E.
Defining three, a limited information set: the finite set of information owned by node i is denoted as R i Represented by R i ＝{R i,1 ,...,R i,H And, wherein H.gtoreq.0, represents the total number of information for node i.
Defining four strong and weak relation node sets: the strong relation node set NS (i) is a set of strong relation nodes of node i, and uses binary vector X i,j Representing the relationship between the strength of nodes, if X i,j =1, meaning j e NS (i), i.e. node j is a strong relational node of node i; the weak relational node set NW (i) is a set of weak relational nodes of node i, if X i,j =α, then represents j∈nw (i), where 0<α<1, node j is a weak relationship node of node i, and the specific relationship is shown in fig. 2.
Definition five, neighbor nodes: the nodes connected with the edges of the node i are all neighbor nodes, and the number of the neighbor nodes is recorded as K. All neighbor nodes of the node i constitute a neighbor node set N (i), where N (i) =ns (i)/(NW) i. If node j is a neighbor node of i, the binary vector relationship is satisfied: x is 0 < X i,j And less than or equal to 1, wherein j is E N (i).
Definition six, information network: the information network is composed of individual nodes and relationships between the nodes, expressed as RG, i.e., rg= { (I, E) |0<X i,j <1,i,j∈I,e∈E}。
S82, a process model;
s821, initializing strong relation node set NS (i) of node i and limited information set R owned by node i i ；
S822, at time t, the information category of the node i is H, and the information set is And->The number of information r owned by node i and node j at time t, respectively. The node i acquires information from an information network, and firstly, information interaction is carried out on the nodes in the strong relation node set NS (i);
s823, at time t+w, information owned by the node i is changed. This change is affected by two reasons, e.g. formulas
Wherein γ represents a trust parameter for information change, 0<γ<1。Representing the finite set of information R at time t i,r Information amount of->Representing the finite set of information R at time t+w i,r Is a variable amount of (a).
The importance and practicality of a certain information decreases over time and the growth and environment itself changes, and the volatility coefficient ρ indicates that the amount of information possessed by the node i gradually decreases over time, where 0< ρ <1.
New information is continually incorporated into new collections as time progresses and their own growth and environment changes. If it isor/>Representation information R i,r The change occurs at time w. Where γ represents a trust parameter for information changes, 0<γ<1. This parameter takes into account the difference in the degree of trust of the nodes when they acquire information from different strong and weak relationship networks.
When the node interacts with the strong relation network member, the information of the node i changes, and the formula is as follows
Formula (VI)
And judging whether the information owned by different network nodes to be exchanged is the same or not, and whether the quantity of the same information is consistent or not. Where j ε NS (i), i.e., node j is a strong relationship node to node i,representing trust parameters for interaction with strongly related nodes,/->
S824, after going through the time t+w, when the node i interacts with the weak relation network node, the node i joins the weak relation set NW (i) with the probability p. The information change of the node i is affected by the strong and weak relation, such as formula
k e NW (i), i.e. node k is a weak relational node of node i, delta represents a trust parameter for interaction with the weak relational node, wherein Representing the finite set of information R at time t+w j,r Information amount of->Representing the finite set of information R at time t+w k,r Is an information amount of (a);
s825, at time t+w, judging the stability of the information network through the standard deviation sigma.
Sigma (t) is variance, I is total node set number, I is node, H is limited information set number, r is limited information node, and N is network node number. Through this process model, interactions and variations of information between nodes can be controlled, thereby controlling the stability of the information network. The model considers the strong and weak relation between nodes, the change of the importance of information along with time and the trust degree between different nodes, thereby better simulating the propagation and change of information between nodes in an actual information network;
S83, a prediction algorithm;
and using a process algorithm to realize one-to-many and many-to-many interactive association of the nodes in the relation space and judging how to optimize the strong and weak relation combination of the nodes. The link prediction algorithm may determine the likelihood of generating a strong or weak relationship between two nodes and find existing but unidentified hidden relationships. And combining the process model and a prediction algorithm, and carrying out dynamic regulation and control on the relationship by utilizing the cosine similarity principle and the dynamic weight of the strong-weak relationship to obtain an optimal path of the necessary resource for individual development.
S831, define a first and a second based on the foregoing model, and construct a network graph g= (I, E). Marking the central node and other nodes with labels, such as node i corresponding to label f (i);
s832, respectively denoted as d according to the radius of the node i relative to the node x and the node y (i,x) And d (i,y) . The label of the center node is 1, if (d (i,x) ,d (i,y) ) The corresponding node label f (i) =2; if (d) (i,x) ,d (i,y) ) The corresponding node label f (i) =3 for either = (1, 2) or (2, 1), and so on.
Two central nodes, where the labels f (i) and the double radius (d (i, x), d (i, y)) satisfy:
if d (i,x) +d (i,y) ≠d (j,x) +d (j,y) Then
If d (i,x) +d (i,y) ＝d (j,x) +d (j,y) Then
Generating an integer label f according to DRNL algorithm l (i) Functions, e.g. formulas
Wherein d x ＝d (i,x) ，d y ＝d (i,y) ，d＝d x -d y . For d (i,x) = infinity or d (i,y) The node of = ≡is denoted as empty tag 0, d x ，d y Represents the radius of node i relative to node x and node y, respectively, and d represents d x ，d y Is a difference in (2);
s833, judging the relation between the node i and the nodes x and y according to the radius of the node i relative to the node x and the node y. When d (i,x) ＝1＝d (i,y) When f (i) =2, the nodes i and x, y have the same relation degree and all show strong relation, otherwise, weak relation; if d (i,x) ＜d (i,y) The relation between the node i and the node x is stronger than the relation between the node i and the node y;
s834, reflecting the strong and weak relation between the nodes by using a cosine similarity calculation formula. Cosine similarity measures the similarity between individuals according to the cosine value of the included angle of the two vectors, wherein the closer to 1 the similarity is, the higher the similarity is, and the closer to 0 the similarity is, the lower the similarity is. For the multidimensional vectors a (a 1, B1, c 1.) and B (a 2, B2, c2..) a1, B1, c 1.) and a2, B2, c2..) represent components in different dimensions, the specific calculations are as the formula
cos θ denotes the cosine inside the trigonometric function, A and B are the two vectors to be compared, A and B denote the norms of vector A and vector B, typically Euclidean norms, expressed as their modulo (length), A i And B i Representing the i-th element in vector a and vector B respectively, Representing the sum of the squares of the elements in vector a and vector B, respectively.
S835, because node i generates different strong and weak relations to different central nodes, different dynamic weights are given to each relation, and the specific calculation is as follows:
for similarity between node i and node x
For similarity between node i and node y
Wherein W is 1 And W is 2 Respectively strong relation and weak relation weights of nodes; w (W) 3 And W is 4 The weak relation degree weights between the nodes i and x and y are respectively; k is the number of nodes in the network, k, except f (i) =1 1 Number of nodes, k, of f (i) =2 in the network 2 Divide k by k 1 Number of nodes outside. The specific calculation is shown as formula
/>
Through the prediction algorithm, the probability of generating strong and weak relations between the nodes can be predicted by combining the similarity, the dynamic weight and the judged relations between the nodes. If the relationship between certain nodes has not been identified explicitly, a potential hidden relationship is revealed using cosine similarity and dynamic weights.
S9, constructing and managing node interaction relations;
s91, initializing node information;
and establishing a total node set according to the node information network, and initializing an edge set, a limited information set, a strong-weak relation node set and neighbor nodes by each node.
S92, information interaction and change processes;
and in each time step, information interaction is carried out among the nodes, and the nodes can acquire information from strong relation nodes and weak relation nodes. Information interaction between strong relationship nodes may be more frequent and dense, while interaction between weak relationship nodes may be less. Updating information sets of nodes using equation (5)Calculating the variation of information according to the interaction of the node with other nodes>ρ is the volatility coefficient, indicating that the amount of information that a node possesses gradually decreases over time.
S93, information change and relation regulation;
calculating the variation of the information in the information set using formulas (3) and (4)Over time, interactions between certain nodes may be enhanced and relationships between nodes may change. A node may join a strong-relationship set of nodes with some probability to other nodes. Using equation (6) to determine whether the information owned by different network nodes to be exchanged is the same and whether the amount of the same information is consistent.
S94, adding weak relation nodes;
over time, interactions between certain nodes may be enhanced and relationships between nodes may change. A node may join a weak set of relational nodes for other nodes with a certain probability. Calculating the amount of change in information using equation (7)
S95, evaluating the stability of the information network;
after each time step is finished, the stability of the information network can be calculated by using a standard deviation formula, and the distribution situation of the information can be evaluated by calculating the standard deviation of the information quantity of each node by using a formula (8). If the standard deviation is stable or reaches the minimum value, the information network is in a stable state
S10, revealing hidden relations;
s1001, constructing a network diagram and a node label;
a relational network graph is built from the data sets, wherein the nodes represent different elements, such as exhibits, themes, etc. For the central node and other nodes, labels are used for marking. The labels of the nodes are calculated according to equation (9).
S1002, predicting the possibility of the relationship among the nodes, and revealing unrecognized hidden relationship;
the cosine similarity between node i and nodes x, y is calculated using equation (10), and this similarity quantifies the degree of similarity between the nodes. The similarity between node i and nodes x, y is calculated according to formulas (11) and (12). Strong and weak relationship weights and weak relationship degree weights with nodes x, y are calculated using equations (13), (14), (15) and (16).
S1003, optimizing information propagation and recommendation;
based on the outcome of the predictive algorithm, the system may optimize the information propagation and recommendation mechanism. If the predicted strong relationship between two nodes is scored higher, the system may recommend that the user view the content of one of the nodes so that they get a thorough understanding of the possible associations. By optimizing the information dissemination, users can obtain more relevant, valuable content, enriching their knowledge and experience.
Therefore, the digital humane application method based on the data strength association model is based on big data analysis, data mining and machine learning, and adopts the data strength association model, and combines the technologies of strength relation, dynamic weight regulation and the like, so that the model can more accurately predict the relation between nodes, optimize the information propagation path and improve the information utilization efficiency.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention and not for limiting it, and although the present invention has been described in detail with reference to the preferred embodiments, it will be understood by those skilled in the art that: the technical scheme of the invention can be modified or replaced by the same, and the modified technical scheme cannot deviate from the spirit and scope of the technical scheme of the invention.

Claims (8)

1. The digital humane application method based on the data strength association model is characterized by comprising the following steps of:
s1, data collection;
s2, preprocessing the original data acquired in the step S1;
s3, digitizing and information library building;
for different types of files, the system needs to conduct context identification to understand the content of the files, metadata information is extracted from each file, object detection and extraction are conducted on images and video files, the extracted metadata information and the content identified by the context are stored in a relational database, a gallery database is established for media files such as pictures and videos, and extracted object information, image features and the like are stored in the gallery database;
S4, constructing a data relationship;
forming a set of data of a database, establishing a knowledge organization model, a person association data set and a data field ontology according to the data set, and finding out data static association;
s5, learning reconstruction;
dynamic knowledge organization based on dimension expansion, semantic label extraction based on entity portraits, and fusion into data dynamic aggregation based on knowledge graphs;
s6, converting the entity relationship into a node information network;
s7, calculating the strong and weak relation between the nodes;
s8, constructing a data strength relation model;
s81, defining a model; s82, defining a total node set, an edge set, a limited information set, a strong-weak relation node set, neighbor nodes and an information network; s83, establishing a process model; combining a process model and a prediction algorithm, and carrying out dynamic regulation and control on the relationship by utilizing a cosine similarity principle and a dynamic weight of the strong-weak relationship to obtain an optimal path of resources necessary for individual development;
s9, constructing and managing node interaction relations;
s10, revealing hidden relations.
2. The method for applying digital humanization based on a data strength association model according to claim 1, wherein in step S7, the strength relationship between nodes is calculated by the relevance score, the strength relationship value is S, the strength relationship value is w, the relevance score is g, the common history period weight is m,
s＝g×m； (1)
w＝1-(g×m) (2)
The weak relation value is obtained by reducing the strong relation value, the correlation between two exhibits is lower or belongs to different historical periods, the weak relation value is correspondingly higher, and the strong relation threshold value and the weak relation threshold value are set to judge whether the node is in strong association or weak association.
3. The method according to claim 1, wherein in step S81, it is assumed that there are N network nodes in the information network, wherein node i is one of them,
defining one, a total node set: i is the total node set, i= {1,2,3,..n }, I, j e I;
defining two, edge sets: e is a set of relationships for nodes, e= {1,2,3,., N }, E;
defining three, a limited information set: the finite set of information owned by node i is denoted as R i Represented by R i ＝R i,1 ,...,R i,H H.gtoreq.0, representing the total number of information for node i;
defining four strong and weak relation node sets: the strong relation node set NS (i) is a set of strong relation nodes of node i, and uses binary vector X i,j Representing the relationship between the strength of nodes, if X i,j =1, meaning j e NS (i), i.e. node j is a strong relational node of node i; the weak relational node set NW (i) is a set of weak relational nodes of node i, if X i,j =α, then represents j∈nw (i), where 0<α<1, node j is a weak relationship node of node i;
definition five, neighbor nodes: the nodes connected with the node i by the edges are all neighbor nodes, the number of the neighbor nodes is recorded as K, all the neighbor nodes of the node i form a neighbor node set N (i), wherein N (i) =NS (i) UW (i), and the neighbor nodes of the node j are the neighbor nodes of the i and satisfy the binary vector relation: x is 0 < X i,j Less than or equal to 1, wherein j is N (i);
definition six, information network: the information network is composed of individual nodes and relationships between the nodes, expressed as RG, i.e., rg= { (I, E) |0<X i,j <1,i,j∈I,e∈E}。
4. A method for applying digital humanization based on a data strength association model according to claim 3, comprising the steps of:
s821, initializing strong relation node set NS (i) of node i and limited information set R owned by node i i ；
S822, at time t, the information category of the node i is H, and the information set isAnd->Respectively representing the quantity of information r owned by a node i and a node j at a moment t, wherein the node i acquires information from an information network, and firstly, the nodes in a strong relation node set NS (i) perform information interaction;
s823, at time t+w, information owned by node i is changed, and the change is influenced by two reasons, namely, a formula
Wherein,representing the finite set of information R at time t+w i,r The variation of ρ is the volatilization coefficient, 0<ρ<1, which means that the amount of information owned by a node gradually decreases with the passage of time, γ means a trust parameter for information change, 0<γ<1，/>Representing the finite set of information R at time t i,r Is an information amount of (a);
new forms of development and environmental changes over time and on their ownContinuously into the new set,when representing information R i,r Changes occur at time w;
when the node interacts with the strong relation network member, the information of the node i changes, and the formula is as follows
Formula (VI)
Determining whether information owned by different network nodes to be exchanged is the same, and whether the amount of the same information is consistent, wherein phi represents a trust parameter for interaction with a strong relationship node,j is epsilon NS (i), i.e., node j is a strong relationship node of node i;
s824, after t+w time, when node i interacts with the weak relation network node, the node i is added to the weak relation set NW (i) with probability p, information change of the node i is affected by the strong and weak relation, and the formula is that
Where k ε NW (i), i.e. node k is the weak relational node of node i, delta represents the trust parameter for interaction with the weak relational node, where Representing the finite set of information R at time t+w j,r Information amount of->Representing the finite set of information R at time t+w k,r Is an information amount of (a);
s825, judging the stability of the information network at time t+w through standard deviation sigma
Sigma (t) is variance, I is total node set number, I is node, H is limited information set number, r is limited information node, and N is network node number.
5. The method for digital humane application based on the data strength association model according to claim 4, wherein in step S83, the method comprises the steps of:
s831, constructing a network graph G= (I, E) based on the definition one and the definition two in the step S81, marking the central node and other nodes by using labels, wherein the node I corresponds to the label G (I);
s832, respectively denoted as d according to the radius of the node i relative to the node x and the node y (i,x) And d (i,y) The label of the center node is 1, (d) (i,x) ,d (i,y) ) Node label f (i) =2 corresponding to= (1, 1); (d) (i,x) ,d (i,y) ) Node label f (i) =3 corresponding to (1, 2) or (2, 1), and so on;
two central nodes, where the labels f (i) and the double radius (d (i, x), d (i, y)) satisfy:
d (i,x) +d (i,y) ≠d (j,x) +d (j,y) in the time-course of which the first and second contact surfaces,
d (i,x) +d (i,y) ＝d (j,x) +d (j,y) in the time-course of which the first and second contact surfaces,
generating an integer tag function f according to DRNL algorithm l (i)，
Wherein d x ＝d (i,x) ，d y ＝d (i,y) ，d＝d x -d y For d (i,x) = infinity or d (i,y) The node of = ≡is denoted as empty tag 0, d x ，d y Represents the radius of node i relative to node x and node y, respectively, and d represents d x ，d y Is a difference in (2);
s833, judging the relation between the node i and the nodes x and y according to the radius of the node i relative to the node x and the node y, and when d (i,x) ＝1＝d (i,y) When f (i) =2, the nodes i and x, y have the same relation degree and all show strong relation, otherwise, weak relation; d, d (i,x) ＜d (i,y) When the relation between the node i and the node x is stronger than the relation between the node i and the node y;
s834, reflecting the strength relation between nodes by using a cosine similarity calculation formula, wherein the cosine similarity measures the similarity between individuals according to the cosine value of the included angle of two vectors, the value is between 0 and 1, the closer 1 is to the higher the similarity, the closer 0 is to the lower the similarity, and for multidimensional vectors A (a 1, B1, c 1.) and B (a 2, B2, c2..), a1, B1, c 1.) and a2, B2, c2., the components in different dimensions are represented by specific calculation formulas
Wherein cos θ represents the cosine inside the trigonometric function, A and B are the two vectors to be compared, A and B represent the norms of the vectors A and B,usually euclidean norms, expressed as their modes, a i And B i Representing the i-th element in vector a and vector B respectively,representing the sum of squares of the elements in vector a and vector B, respectively;
S835, because node i generates different strong and weak relations for different central nodes, different dynamic weights are given to each relation.
6. The method for digital humane application based on the data strength association model according to claim 5, wherein in step S835, each relationship is given a different dynamic weight, which is specifically calculated as follows:
for similarity between node i and node x
For similarity between node i and node y
Wherein w is 1 And w 2 Respectively strong relation and weak relation weights of nodes; w (w) 3 And w 4 The weak relation degree weights between the nodes i and x and y are respectively; k is the number of nodes in the network, k, except f (i) =1 1 Number of nodes, k, of g (i) =2 in the network 2 Divide k by k 1 The specific calculation of the number of external nodes is shown in the formula
The prediction algorithm is combined with similarity and dynamic weight among nodes and judged relation calculation and dynamic weight regulation to determine the possibility of generating strong and weak relations among the nodes, and when the relations among some nodes are not clearly identified, the unidentified hidden relations are revealed by utilizing cosine similarity and dynamic weight, so that an optimal path of resources required by individual development is obtained.
7. The method for applying digital humanization based on the data strength association model according to claim 4, wherein in step S9, the method comprises the steps of:
S91, initializing node information; establishing a total node set according to a node information network, and initializing an edge set, a limited information set, a strong-weak relation node set and neighbor nodes by each node;
s92, information interaction and change processes;
in each time step, information interaction is carried out among nodes, the nodes acquire information from strong relation nodes and weak relation nodes, the information interaction among the strong relation nodes is more frequent and dense, the information interaction among the weak relation nodes is less, and the information set of the nodes is updated by using a formula (5)Calculating the variation of information according to the interaction of the node and other nodes
S93, information change and relation regulation;
calculating the variation of the information in the information set by using the formulas (3) and (4)Over time, interaction between certain nodes is enhanced, the relation between the nodes is changed, the nodes are added into strong relation node sets of other nodes with a certain probability, and whether information owned by different network nodes to be exchanged is the same or not and whether the quantity of the same information is consistent or not is judged by using a formula (6);
s94, adding weak relation nodes;
over time, the interaction between certain nodes is enhanced, the relationship between the nodes is changed, the nodes are added into the weak relationship node set of other nodes with a certain probability, and the formula (7) is used for calculating the change amount of information
S95, evaluating the stability of the information network;
after each time step is finished, the stability of the information network is calculated by using a standard deviation formula, the distribution condition of the information is evaluated by calculating the standard deviation of the information quantity of each node by using a formula (8), the standard deviation is stabilized or reaches the minimum value, and the information network is stabilized.
8. The method for applying digital humanization based on the data strength association model according to claim 5, wherein in step S10, the method comprises the steps of:
s101, constructing a network diagram and a node label;
building a relational network graph according to the data set, wherein nodes represent different elements, exhibits and subjects, marking the central node and other nodes by using labels, and calculating the labels of the nodes according to a formula (9);
s102, predicting the possibility of the relationship among the nodes, and revealing unrecognized hidden relationship;
calculating cosine similarity between the node i and the nodes x and y by using a formula (10), quantifying the similarity between the nodes, calculating the similarity between the node i and the nodes x and y according to formulas (11) and (12), and calculating strong relationship and weak relationship weights and weak relationship degree weights between the node i and the nodes x and y by using formulas (13), (14), (15) and (16);
S103, optimizing a propagation and recommendation mechanism of information according to a result of the prediction algorithm;
when the predictive strong relationship score between two nodes is high, the system recommends users to view the content of one of the nodes so that they can learn about the possible relationship, and by optimizing the information propagation, users obtain more relevant and valuable content, thereby enriching their knowledge and experience.
CN202311318274.0A 2023-10-12 2023-10-12 Digital humane application method based on data strength association model Active CN117349386B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311318274.0A CN117349386B (en) 2023-10-12 2023-10-12 Digital humane application method based on data strength association model

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311318274.0A CN117349386B (en) 2023-10-12 2023-10-12 Digital humane application method based on data strength association model

Publications (2)

Publication Number Publication Date
CN117349386A true CN117349386A (en) 2024-01-05
CN117349386B CN117349386B (en) 2024-04-12

Family

ID=89370550

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311318274.0A Active CN117349386B (en) 2023-10-12 2023-10-12 Digital humane application method based on data strength association model

Country Status (1)

Country Link
CN (1) CN117349386B (en)

Citations (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105740366A (en) * 2016-01-26 2016-07-06 哈尔滨工业大学深圳研究生院 Inference method and device of MicroBlog user interests
CN106599133A (en) * 2016-12-02 2017-04-26 中国科学院软件研究所 Draft interaction based monitoring video visual analysis method
CN107957998A (en) * 2016-10-14 2018-04-24 富士通株式会社 The method of relation between the method and estimation entity pair of production Methods computation model
US20180336221A1 (en) * 2017-05-22 2018-11-22 Fujitsu Limited System, method, and program for reconciling input datasets with a model ontology
CN109194509A (en) * 2018-08-28 2019-01-11 浙江工业大学 A method of the prediction network based on path strong or weak relation and community information connects side
CN109857871A (en) * 2019-01-28 2019-06-07 重庆邮电大学 A kind of customer relationship discovery method based on social networks magnanimity context data
CN110188422A (en) * 2019-05-16 2019-08-30 深圳前海微众银行股份有限公司 A kind of method and device of feature vector that extracting node based on network data
CN111309944A (en) * 2020-01-20 2020-06-19 同方知网(北京)技术有限公司 Digital human search algorithm based on graph database
WO2021212682A1 (en) * 2020-04-21 2021-10-28 平安国际智慧城市科技股份有限公司 Knowledge extraction method, apparatus, electronic device, and storage medium
CN113947259A (en) * 2021-11-09 2022-01-18 吉林大学 Method for predicting vehicle speeds of drivers in different styles based on GRU neural network
CN114117220A (en) * 2021-11-26 2022-03-01 东北大学 Deep reinforcement learning interactive recommendation system and method based on knowledge enhancement
CN115063251A (en) * 2022-05-30 2022-09-16 华侨大学 Social communication propagation dynamic network representation method based on relationship strength and feedback mechanism
CN115239442A (en) * 2022-09-22 2022-10-25 湖南快乐通宝小额贷款有限公司 Method and system for popularizing internet financial products and storage medium
CN116150401A (en) * 2023-02-16 2023-05-23 大连海洋大学 Strong robustness knowledge graph triplet quality inspection network model training method and quality inspection method based on noisy data set
CN116841741A (en) * 2023-06-30 2023-10-03 哈尔滨理工大学 Self-evolution reasoning method for mobile crowd sensing node relation prediction

Patent Citations (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105740366A (en) * 2016-01-26 2016-07-06 哈尔滨工业大学深圳研究生院 Inference method and device of MicroBlog user interests
CN107957998A (en) * 2016-10-14 2018-04-24 富士通株式会社 The method of relation between the method and estimation entity pair of production Methods computation model
CN106599133A (en) * 2016-12-02 2017-04-26 中国科学院软件研究所 Draft interaction based monitoring video visual analysis method
US20180336221A1 (en) * 2017-05-22 2018-11-22 Fujitsu Limited System, method, and program for reconciling input datasets with a model ontology
CN109194509A (en) * 2018-08-28 2019-01-11 浙江工业大学 A method of the prediction network based on path strong or weak relation and community information connects side
CN109857871A (en) * 2019-01-28 2019-06-07 重庆邮电大学 A kind of customer relationship discovery method based on social networks magnanimity context data
CN110188422A (en) * 2019-05-16 2019-08-30 深圳前海微众银行股份有限公司 A kind of method and device of feature vector that extracting node based on network data
CN111309944A (en) * 2020-01-20 2020-06-19 同方知网(北京)技术有限公司 Digital human search algorithm based on graph database
WO2021212682A1 (en) * 2020-04-21 2021-10-28 平安国际智慧城市科技股份有限公司 Knowledge extraction method, apparatus, electronic device, and storage medium
CN113947259A (en) * 2021-11-09 2022-01-18 吉林大学 Method for predicting vehicle speeds of drivers in different styles based on GRU neural network
CN114117220A (en) * 2021-11-26 2022-03-01 东北大学 Deep reinforcement learning interactive recommendation system and method based on knowledge enhancement
CN115063251A (en) * 2022-05-30 2022-09-16 华侨大学 Social communication propagation dynamic network representation method based on relationship strength and feedback mechanism
CN115239442A (en) * 2022-09-22 2022-10-25 湖南快乐通宝小额贷款有限公司 Method and system for popularizing internet financial products and storage medium
CN116150401A (en) * 2023-02-16 2023-05-23 大连海洋大学 Strong robustness knowledge graph triplet quality inspection network model training method and quality inspection method based on noisy data set
CN116841741A (en) * 2023-06-30 2023-10-03 哈尔滨理工大学 Self-evolution reasoning method for mobile crowd sensing node relation prediction

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
王曙燕 等: "融合节点标签与强弱关系的链路预测算法", 《计算机工程与应用》, vol. 58, no. 18, 12 April 2022 (2022-04-12), pages 71 - 77 *

Also Published As

Publication number Publication date
CN117349386B (en) 2024-04-12

Similar Documents

Publication Publication Date Title
Shen et al. A web-scale system for scientific knowledge exploration
CN111680173B (en) CMR model for unified searching cross-media information
CN106776711B (en) Chinese medical knowledge map construction method based on deep learning
WO2016045153A1 (en) Information visualization method and intelligent visible analysis system based on textual resume information
US20110264649A1 (en) Adaptive Knowledge Platform
US20100205176A1 (en) Discovering City Landmarks from Online Journals
CN107818183B (en) Three-stage combined party building video recommendation method based on feature similarity measurement
Huang et al. Research on multi-label user classification of social media based on ML-KNN algorithm
Nie et al. Cross-domain semantic transfer from large-scale social media
CN117349386B (en) Digital humane application method based on data strength association model
Zheng et al. The best answer prediction by exploiting heterogeneous data on software development Q&A forum
CN116244497A (en) Cross-domain paper recommendation method based on heterogeneous data embedding
Yin et al. Geographic knowledge graph attribute normalization: improving the accuracy by fusing optimal granularity clustering and co-occurrence analysis
Mekthanavanh et al. Social web video clustering based on multi-modal and clustering ensemble
Guesmi et al. Community detection in multi-relational bibliographic networks
Shang et al. Intelligent optimization method of resource recommendation service of mobile library based on digital twin technology
Yu et al. Construction of hierarchical cognitive academic map
Pushpalatha et al. A tree based representation for effective pattern discovery from multimedia documents
Zheng et al. A method of dividing clinical data set for medical image AI training
Hammouda Distributed document clustering and cluster summarization in peer-to-peer environments
Liao Construction of Intelligent Recommendation Retrieval Model of FuJian Intangible Cultural Heritage Digital Archives Resources.
Sheng et al. Measuring semantic similarity in ontology and its application in information retrieval
Yang et al. [Retracted] Structural Analysis of PE Teaching Strategy and System Knowledge Management Based on Mobile Digital Information Technology
Li Music recommendation algorithm based on user portrait
Zhang et al. Representation Learning in Academic Network Based on Research Interest and Meta-path

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