WO2021135290A1 - 基于知识图谱的信息可视化方法、装置、设备及存储介质 - Google Patents

基于知识图谱的信息可视化方法、装置、设备及存储介质 Download PDF

Info

Publication number
WO2021135290A1
WO2021135290A1 PCT/CN2020/111114 CN2020111114W WO2021135290A1 WO 2021135290 A1 WO2021135290 A1 WO 2021135290A1 CN 2020111114 W CN2020111114 W CN 2020111114W WO 2021135290 A1 WO2021135290 A1 WO 2021135290A1
Authority
WO
WIPO (PCT)
Prior art keywords
knowledge element
knowledge
element entity
entity
preset
Prior art date
Application number
PCT/CN2020/111114
Other languages
English (en)
French (fr)
Inventor
徐永泽
赖长明
薛凯文
Original Assignee
深圳Tcl新技术有限公司
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 深圳Tcl新技术有限公司 filed Critical 深圳Tcl新技术有限公司
Publication of WO2021135290A1 publication Critical patent/WO2021135290A1/zh

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
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
    • G06F16/367—Ontology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/34—Browsing; Visualisation therefor

Definitions

  • This application relates to the technical field of knowledge graphs, and in particular to a method, device, equipment, and storage medium for information visualization based on knowledge graphs.
  • the knowledge graph technology was formally proposed by Google in May 2012. Its original intention is to improve the capabilities of search engines and enhance the search quality and experience of users.
  • knowledge graph technology has been widely used in fields such as intelligent search, intelligent question answering, and personalized recommendation.
  • the data storage and usage forms involved in the knowledge graph technology have played an important role in promoting the understanding of the knowledge of the human world by machines.
  • Knowledge graph technology displays complex knowledge fields through data mining, information processing, knowledge measurement and graph drawing, reveals the dynamic development of knowledge fields, and provides practical and valuable references for subject research.
  • the knowledge graph technology is generated from search engines, and it is also the largest development and application in this field. Now, more application products will consider using the information brought by knowledge graph technology to optimize recommendation, intelligent question answering and other information systems.
  • knowledge graph technology to optimize recommendations and intelligent question and answer, users need to provide clearly targeted information. Therefore, it is very inconvenient for users to retrieve and view information when they do not have specific goals, resulting in a sense of experience for users in resource retrieval. Poor.
  • the main purpose of this application is to provide an information visualization method, device, equipment and storage medium based on a knowledge graph, which aims to solve the inconvenience and poor experience of information retrieval and viewing for users of the prior art when there is no specific goal. problem.
  • this application provides an information visualization method based on a knowledge graph.
  • the method includes the following steps:
  • the target knowledge element entity and the initial knowledge element entity are displayed through the current interface.
  • this application also proposes an electronic device, the device including: a memory, a processor, and an information visualization program based on a knowledge graph that is stored on the memory and can run on the processor,
  • the information visualization program based on the knowledge graph is configured to implement the steps of the information visualization method based on the knowledge graph as described above.
  • this application also proposes a storage medium that stores an information visualization program based on a knowledge graph, and the information visualization program based on a knowledge graph is executed by a processor as described above The steps of the information visualization method based on the knowledge graph.
  • the initial knowledge element entity is determined according to the knowledge element entity information input by the user; the associated knowledge element entity corresponding to the initial knowledge element entity is searched in the preset knowledge graph, and the associated knowledge element entity and the initial knowledge are determined
  • the map distance of the meta-entity determine that the map distance belongs to the preset target distance range, and use the associated knowledge meta-entity corresponding to the map distance as the target knowledge meta-entity; compare the target knowledge meta-entity and the initial knowledge through the current interface Meta entities are displayed.
  • FIG. 1 is a schematic structural diagram of an electronic device in a hardware operating environment involved in a solution of an embodiment of the present application
  • FIG. 2 is a schematic flowchart of the first embodiment of the information visualization method based on the knowledge graph of this application;
  • FIG. 3 is a schematic diagram of general dimensional information visualization according to an embodiment of the application.
  • FIG. 4 is a schematic flowchart of a second embodiment of the information visualization method based on the knowledge graph of this application;
  • FIG. 5 is a schematic flowchart of a third embodiment of the information visualization method based on the knowledge graph of this application.
  • FIG. 6 is a schematic diagram of the visualization of jump dimension information according to an embodiment of the application.
  • FIG. 7 is a schematic flowchart of a fourth embodiment of an information visualization method based on a knowledge graph of this application.
  • FIG. 8 is a schematic diagram of visualization of mobile center information according to an embodiment of this application.
  • FIG. 9 is a structural block diagram of the first embodiment of the information visualization device based on the knowledge graph of this application.
  • FIG. 1 is a schematic structural diagram of an electronic device in a hardware operating environment involved in an embodiment of the application.
  • the electronic device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005.
  • the communication bus 1002 is used to implement connection and communication between these components.
  • the user interface 1003 may include a display screen (Display) and an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface.
  • the network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (WI-FIdelity, WI-FI) interface).
  • WI-FIdelity wireless fidelity
  • the memory 1005 may be a high-speed random access memory (Random Access Memory, RAM) memory, or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk memory.
  • RAM Random Access Memory
  • NVM Non-Volatile Memory
  • the memory 1005 may also be a storage device independent of the aforementioned processor 1001.
  • FIG. 1 does not constitute a limitation on the electronic device, and may include more or less components than those shown in the figure, or combine certain components, or arrange different components.
  • the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and an information visualization program based on a knowledge graph.
  • the network interface 1004 is mainly used for data communication with a network server; the user interface 1003 is mainly used for data interaction with users; the processor 1001 and the memory 1005 in the electronic device of this application can be set in In the electronic device, the electronic device calls the information visualization program based on the knowledge graph stored in the memory 1005 through the processor 1001, and executes the information visualization method based on the knowledge graph provided in the embodiment of the present application.
  • FIG. 2 is a schematic flowchart of a first embodiment of an information visualization method based on a knowledge graph of this application.
  • the information visualization method based on the knowledge graph includes the following steps:
  • Step S10 Determine the initial knowledge element entity according to the knowledge element entity information input by the user.
  • a preset knowledge graph before the step of determining the initial knowledge meta-entity according to the knowledge meta-entity information input by the user, a preset knowledge graph needs to be constructed, where constructing the preset knowledge graph includes obtaining corpus information of preset topic dimensions;
  • the corpus information is input into the TransR model corresponding to the preset topic dimension to obtain the corresponding knowledge element entity, the feature vector of the knowledge element entity, and the preset map distance between the knowledge element entities;
  • the knowledge element entity, the feature vector of the knowledge element entity, and the predetermined map distance construct a predetermined knowledge map.
  • the corpus information is input into the TransR model corresponding to the preset topic dimension.
  • the preferred model in this embodiment is the TransR model, and other models can also be used to construct the knowledge graph, which is not limited in this embodiment.
  • the initial knowledge meta-entity is determined according to the knowledge meta-entity information input by the user.
  • the movie domain is used as an example to construct a preset knowledge graph. It is assumed that movie A is a movie that the user has watched, and the user does not have a specific goal. For quick information retrieval and viewing, the user decides to use movie A as the starting point of the search, and then determines the initial knowledge element entity based on the knowledge element entity information input by the user. At this time, the initial knowledge element entity can be regarded as movie A.
  • Step S20 Find the associated knowledge element entity corresponding to the initial knowledge element entity in the preset knowledge graph, and determine the graph distance between the associated knowledge element entity and the initial knowledge element entity.
  • the step of searching for the associated knowledge element entity corresponding to the initial knowledge element entity in the preset knowledge graph, and determining the graph distance between the associated knowledge element entity and the initial knowledge element entity specifically includes : According to the initial knowledge element entity, extract the feature vector corresponding to the initial knowledge element entity from the preset knowledge graph; according to the feature vector and the knowledge element entities included in the preset knowledge graph Search for the associated knowledge element entity corresponding to the initial knowledge element entity in the preset map distance of, and determine the map distance between the associated knowledge element entity and the initial knowledge element entity according to the search result.
  • the initial knowledge meta-entity can be regarded as movie A
  • the feature vector corresponding to the initial knowledge meta-entity, namely movie A, extracted from the preset knowledge graph can be denoted as h.
  • the feature vector h and the prediction Set the preset map distance between the knowledge element entities included in the knowledge graph to find the associated knowledge element entity corresponding to the initial knowledge element entity, wherein the preset knowledge element entities included in the preset knowledge graph
  • the expression form of the preset map distance can be adjusted according to actual needs.
  • This embodiment is not limited, assuming that the initial knowledge element entity corresponds to movie A
  • the associated knowledge meta-entity includes movie B, movie C, movie D, movie E, movie F, movie G, movie H, movie I, movie J, and movie K. According to the search result, it is determined that the associated knowledge meta-entity and the initial The map distance of the knowledge element entity.
  • Step S30 Determine that the map distance belongs to a preset target distance range, and use the associated knowledge element entity corresponding to the map distance as the target knowledge element entity.
  • the associated knowledge element entity corresponding to the map distance is used as the target knowledge element entity.
  • the preset target distance range is artificially defined, and the size of the preset target distance range can be adjusted according to actual needs. This embodiment is not limited to this. It is assumed that the initial knowledge element entity is the associated knowledge corresponding to movie A
  • the meta-entity includes movie B, movie C, movie D, movie E, movie F, movie G, movie H, movie I, movie J, movie K, where the map distance belongs to the preset target distance range including movie B, movie C, movie D, movie E, movie F, movie G, movie H, movie I.
  • the method further includes: sorting the map distance by size, and obtaining a sorting result; and selecting a preset number of target knowledge element entities from the associated knowledge element entities according to the sorting result.
  • the preset number is artificially defined, and the size of the preset number can be adjusted according to actual needs. This embodiment is not limited to this. It is assumed that the initial knowledge element entity, that is, the associated knowledge element entity corresponding to movie A, includes movie B.
  • the target knowledge meta-entities with a preset number of 8 are selected, and the target knowledge meta-entities are obtained as movie B, movie C, movie D, movie E, movie F, movie G, movie H, and movie I.
  • Step S40 Display the target knowledge element entity and the initial knowledge element entity through the current interface.
  • the step of displaying the target knowledge element entity and the initial knowledge element entity through the current interface specifically includes: taking the initial knowledge element entity as the display center, and displaying the target knowledge element entity Arrange evenly around the display center, and display the arrangement result through the current interface.
  • the display mode of the current interface can be artificially defined, and the display mode of the current interface can be adjusted according to actual needs, which is not limited in this embodiment.
  • Fig. 3 is a schematic diagram of general dimensional information visualization according to an embodiment of this application; wherein, the initial knowledge element entity is movie A (A in Fig. 3), and the target knowledge element entity is movie B (such as B in Figure 3), Movie C (C in Figure 3), Movie D (D in Figure 3), Movie E (E in Figure 3), Movie F (F in Figure 3), Movie G (such as In Fig. 3, G), movie H (H in Fig. 3), and movie I (I in Fig. 3) are evenly arranged around the display center, that is, movie A (A in Fig. 3).
  • the initial knowledge element entity is determined according to the knowledge element entity information input by the user; the associated knowledge element entity corresponding to the initial knowledge element entity is searched in the preset knowledge graph, and the associated knowledge element entity and the initial knowledge element entity are determined The map distance of the knowledge element entity; determine that the map distance belongs to the preset target distance range, and use the associated knowledge element entity corresponding to the map distance as the target knowledge element entity; compare the target knowledge element entity and the initial The knowledge element entity is displayed.
  • FIG. 4 is a schematic flowchart of a second embodiment of a method for information visualization based on a knowledge graph of this application.
  • the information visualization method based on the knowledge graph of this embodiment further includes:
  • Step S101 Obtain corpus information of preset topic dimensions.
  • the corpus information of at least one domain of the knowledge graph to be constructed is acquired, and the corpus information may include: multiple pieces of knowledge text content.
  • the knowledge graph can be constructed through the knowledge graph construction system.
  • the knowledge graph construction system can be hardware devices such as computers, servers, or software installed on hardware devices.
  • the field can refer to the professional field, such as the "metallurgy” field, the "economics” field, the "medicine” field, etc.
  • the field can have multiple subfields, for example, the "medicine” field includes the "pediatric medicine” field.
  • the corpus information refers to the content fragments of the knowledge text, and the corpus information involved in the knowledge graph to be constructed in this embodiment may be the movie "Chasing the Dragon". Knowledge text content fragments such as "Chasing the Dragon movie starred by Andy Lau”.
  • Step S102 Input the corpus information into the TransR model corresponding to the preset topic dimension to obtain the corresponding knowledge element entity, the feature vector of the knowledge element entity, and the preset map distance between the knowledge element entities .
  • the corpus information is input into the TransR model corresponding to the preset topic dimension.
  • the preferred model in this embodiment is the TransR model, and other models can also be used to construct the knowledge graph, which is not included in this embodiment. limit.
  • the representative work of knowledge graph representation methods at home and abroad mainly includes the embedded model TransE based on translation, the embedded model TransH based on hyperplane, the embedded model TransR based on entity relationship space, and the embedding based on clustering and entity relationship space. CTransR and the embedded model TransD based on the dynamic mapping matrix.
  • the above methods are collectively referred to as the knowledge representation model based on translation.
  • the embedded model TransR based on the entity relationship space represents each relationship r by establishing a mapping matrix Mr and a vector r. Specifically, TransR maps the head entity vector h and the tail entity vector t to the relationship vector r through a matrix.
  • the corpus information is input into the TransR model corresponding to the preset topic dimension to obtain the corresponding knowledge element entity, the feature vector of the knowledge element entity, and the preset atlas between the knowledge element entities distance.
  • the corpus information may include: multiple knowledge text content fragments, word segmentation and part-of-speech tagging are performed on the knowledge text content fragments, the keywords in the knowledge text content fragments are obtained, and the keywords are preset Match the rules of the domain ontology to obtain the knowledge element instance, the attribute of the knowledge element instance and the association relationship between the knowledge element instance in the knowledge text content fragment;
  • the domain ontology includes: the domain topic, the domain topic includes at least one model, model The attributes of and the relationship between the models; the model includes at least one knowledge element instance.
  • the domain theme can refer to a professional field or a sub-field.
  • the professional field can have multiple sub-fields.
  • Each field will have multiple models.
  • Each model will have its own unique attributes, and there will be various relationships between the models.
  • the corpus information is translated into the corresponding knowledge element entity, the feature vector of the knowledge element entity, and the preset map distance between each knowledge element entity through the TransR model.
  • the preset map distance between the knowledge element entities included in the preset knowledge map is artificially defined when the preset knowledge map is constructed according to the TransR model, and the expression form of the preset map distance can be adjusted according to actual needs.
  • the embodiment does not impose restrictions on this,
  • Step S103 Construct a preset knowledge graph according to the knowledge element entity, the feature vector of the knowledge element entity, and the preset graph distance.
  • the corpus information of a preset topic dimension is acquired; the corpus information is input into the TransR model corresponding to the preset topic dimension to obtain the corresponding knowledge element entity, the feature vector of the knowledge element entity, and the corresponding knowledge element entity.
  • the preset atlas distance between the knowledge element entities; the preset knowledge atlas is constructed according to the knowledge element entities, the feature vectors of the knowledge element entities, and the preset atlas distances.
  • FIG. 5 is a schematic flowchart of a third embodiment of a method for information visualization based on a knowledge graph of this application.
  • the method for information visualization based on the knowledge graph of this embodiment further includes:
  • Step S51 Determine the jump knowledge element entity according to the jump knowledge element entity information input by the user based on the current interface.
  • the target knowledge element entity and the initial knowledge element entity are displayed through the current interface.
  • the initial knowledge meta-entity movie A as the display center, surround the target knowledge meta-entity, namely movie B, movie C, movie D, movie E, movie F, movie G, movie H, and movie I.
  • the display center is evenly arranged, and the result of the arrangement is displayed through the current interface.
  • the jump knowledge element entity is determined according to the jump knowledge element entity information input by the user based on the current interface.
  • the dimension to be jumped needs to be selected, and the user inputs the jump knowledge based on the current interface
  • the meta-entity information determines the meta-entity of the jump knowledge, assuming that the meta-entity of the jump knowledge is the "protagonist" relationship dimension.
  • Step S52 Find the projection matrix corresponding to the jumped knowledge element entity in the preset knowledge graph, and update the preset graph distance between the knowledge element entities in the preset knowledge graph according to the projection matrix to obtain a new Knowledge graph.
  • the dimension to be jumped needs to be selected, and the user determines the jumped knowledge element entity based on the jumped knowledge element entity information entered in the current interface, assuming that the jumped knowledge element entity is the "leader" Relationship dimension.
  • Set the map distance that is, to multiply each knowledge element entity in the preset knowledge map by the projection matrix Mr to obtain a new knowledge map.
  • Step S53 Find the jump-related knowledge element entity corresponding to the initial knowledge element entity in the new knowledge graph, and determine the jump graph between the jump-related knowledge element entity and the initial knowledge element entity distance.
  • the jump-related knowledge element entity corresponding to the initial knowledge element entity is searched, and the relationship between the jump-related knowledge element entity and the initial knowledge element entity is determined
  • the step of jumping to the map distance specifically includes: extracting a feature vector corresponding to the initial knowledge element entity from the new knowledge map according to the initial knowledge element entity; according to the feature vector and the new knowledge
  • the preset graph distance between the knowledge element entities included in the graph searches for the jump-related knowledge element entity corresponding to the initial knowledge element entity; determining the jump-related knowledge element entity and the initial knowledge element entity according to the search result The jump map distance.
  • the corresponding projection matrix when the jump knowledge element entity is searched for the "leading actor” relationship dimension in the preset knowledge graph is marked as Mr.
  • the projection matrix updates the preset map distance between the knowledge element entities in the preset knowledge graph, that is, the projection matrix Mr is left multiplied by each knowledge element entity in the preset knowledge graph to obtain a new knowledge graph.
  • the feature vector corresponding to the initial knowledge element entity extracted from the new knowledge graph can be denoted as h1, and the search is based on the feature vector h1 and the preset graph distance between the knowledge element entities included in the new knowledge graph
  • the jump-related knowledge meta-entity corresponding to the initial knowledge meta-entity assuming that the jump-related knowledge meta-entity is the "protagonist" relationship dimension, the jump-related knowledge meta-entity corresponding to the initial knowledge meta-entity includes movie B, movie J, and movie M. Movie E, Movie K, Movie L, Movie H, Movie I, Movie X, Movie Y, and determine the jump map distance between the jump associated knowledge element entity and the initial knowledge element entity according to the search result.
  • Step S54 When the jump map distance belongs to the preset target distance range, use the jump associated knowledge element entity corresponding to the jump map distance as the jump target knowledge element entity.
  • the jump associated knowledge element entity corresponding to the jump map distance is used as the jump target knowledge element entity.
  • the preset target distance range is artificially defined, and the size of the preset target distance range can be adjusted according to actual needs. This embodiment does not impose restrictions on this.
  • the jump-related knowledge meta-entities corresponding to the initial knowledge meta-entity include movie B, movie J, movie M, movie E, movie K, movie L, movie H, movie I, movie X, and movie Y, where the jump map
  • the distances belonging to the preset target distance range include movie B, movie J, movie M, movie E, movie K, movie L, movie H, and movie I.
  • the method further includes: sorting the jump map distances by size and obtaining the sorting result; and selecting a preset number of jump targets from the jump-related knowledge element entities according to the sorting result Knowledge meta-entity.
  • the preset number is artificially defined, and the size of the preset number can be adjusted according to actual needs.
  • the jump knowledge element entity is the "protagonist" relationship dimension, and the initial knowledge element
  • the jump-related knowledge meta-entity corresponding to the entity includes movie B, movie J, movie M, movie E, movie K, movie L, movie H, movie I, movie X, and movie Y.
  • the preset number is 8, right
  • the jump map distances of the jump-related knowledge element entities are sorted by size, and the jump sort results are obtained as movie B, movie J, movie M, movie E, movie K, movie L, movie H, movie I, movie X , Movie Y, select a preset number of 8 jump target knowledge element entities from the jump related knowledge element entities according to the jump sorting result, and obtain the jump target knowledge element entities as movie B, movie J, and movie M, movie E, movie K, movie L, movie H, movie I.
  • Step S55 Display the initial knowledge element entity and the jump target knowledge element entity with the initial knowledge element entity as the center.
  • the initial knowledge element entity and the jump target knowledge element entity are displayed through the current interface.
  • the display mode of the current interface can be artificially defined, and the display mode of the current interface can be adjusted according to actual needs, which is not limited in this embodiment.
  • FIG. 6 is a schematic diagram of the visualization of jump dimension information according to an embodiment of the application; wherein, the initial knowledge element entity is movie A (A in FIG.
  • the jump target knowledge element entity is movie B (B in Figure 6), Movie J (J in Figure 6), Movie M (M in Figure 6), Movie E (E in Figure 6), Movie K (K in Figure 6), Movie L (L in Figure 6), movie H (H in Figure 6), and movie I (I in Figure 6) are evenly arranged around the display center, that is, movie A (A in Figure 6).
  • the map distance between employee A and employee B is the correlation between employee A and employee B, and if employee A and employee B are considered in the dimension of the work relationship
  • the correlation may be due to the fact that there is a high degree of correlation between employee A and employee B in the same leadership boss. If the relationship between employee A and employee B is considered in the dimension of blood relationship, that is, the distance between employee A and employee B.
  • the correlation between employees A and B may have a low degree of correlation. But it will not list the blood relatives of employee A and the blood relatives of employee B just because the dimension jumps to the blood relationship dimension. What is still considered is the map distance between employee A and employee B, that is, the distance between employee A and employee B. Correlation.
  • the input jumping knowledge element entity information determines the jumping knowledge element entity; searching the projection matrix corresponding to the jumping knowledge element entity in the preset knowledge graph, and updating each knowledge in the preset knowledge graph according to the projection matrix
  • the preset map distance between meta entities is used to obtain a new knowledge map; in the new knowledge map, the map distance calculation is performed for the employee A and the employee B.
  • attribute values such as the number of user ratings, user views, and user comments can be collected.
  • attribute values such as the number of user ratings, user views, and user comments.
  • use the number of user ratings, user views, Attribute values such as the number of user comments construct a preset knowledge graph, where each movie is still an entity, and “user views”, “user ratings”, “user comments”, etc. are new relationships in the preset knowledge graph.
  • the triple (h, r, t) array of the knowledge meta-entity can be parsed as such that user ⁇ has watched movie A, plus the corpus information of the preset topic dimensions to construct the preset knowledge Atlas, the preset knowledge atlas is updated, and the projection matrix of the relationship dimension between the feature vector of the user entity and the behavior (such as the number of user ratings, user page views, and user comments) is added. Therefore, according to step S51 to step S55, a target of the jump operation can be added, which is called the user dimension.
  • the user dimension corresponds to the actual physical meaning. The relationship is closer to actual application scenarios.
  • the jump knowledge element entity is determined according to the jump knowledge element entity information input by the user based on the current interface; the projection matrix corresponding to the jump knowledge element entity is searched in the preset knowledge graph, and the projection matrix is determined according to the projection.
  • the matrix updates the preset graph distance between the knowledge element entities in the preset knowledge graph to obtain a new knowledge graph; in the new knowledge graph, search for the jump-related knowledge element corresponding to the initial knowledge element entity Entity, and determine the jump map distance between the jump-associated knowledge element entity and the initial knowledge element entity; when the jump map distance belongs to the preset target distance range, the jump map
  • the jump associated knowledge element entity corresponding to the distance is used as the jump target knowledge element entity; the initial knowledge element entity and the jump target knowledge element entity are displayed with the initial knowledge element entity as the center.
  • FIG. 7 is a schematic flowchart of a fourth embodiment of a method for information visualization based on a knowledge graph of this application.
  • the method for information visualization based on the knowledge graph of this embodiment further includes:
  • Step S61 Determine the mobile knowledge element entity according to the mobile knowledge element entity information input by the user based on the current interface.
  • the target knowledge element entity and the initial knowledge element entity are displayed through the current interface.
  • the initial knowledge meta-entity movie A as the display center, surround the target knowledge meta-entity, namely movie B, movie C, movie D, movie E, movie F, movie G, movie H, and movie I.
  • the display center is evenly arranged, and the result of the arrangement is displayed through the current interface.
  • the mobile knowledge element entity is determined according to the mobile knowledge element entity information input by the user based on the current interface.
  • the knowledge element entity of the center needs to be reselected, and the mobile knowledge element entity entered by the user based on the current interface
  • the information determines the mobile knowledge meta-entity, assuming that the mobile knowledge meta-entity is movie B.
  • Step S62 Search for the mobile-related knowledge element entity corresponding to the mobile knowledge element entity in the preset knowledge graph, and determine the mobile map distance between the mobile-related knowledge element entity and the mobile knowledge element entity.
  • the step of searching for the mobile knowledge element entity corresponding to the mobile knowledge element entity in the preset knowledge graph, and determining the distance of the mobile knowledge element entity and the mobile knowledge element entity from the mobile knowledge element entity Specifically including: according to the mobile knowledge element entity, extracting the feature vector corresponding to the mobile knowledge element entity from a preset knowledge graph; according to the feature vector and one of the knowledge element entities included in the preset knowledge graph Search for the mobile-related knowledge element entity corresponding to the mobile knowledge element entity by the preset map distance between the two; and determine the mobile map distance between the mobile-related knowledge element entity and the mobile knowledge element entity according to the search result.
  • the mobile knowledge element entity can be regarded as movie B, and the feature vector corresponding to the mobile knowledge element entity, namely movie B, extracted from the preset knowledge graph can be denoted as h2, and according to the feature vector h2 and the prediction Set the preset map distance between the knowledge element entities included in the knowledge graph to search for the mobile associated knowledge element entity corresponding to the mobile knowledge element entity, where the distance between the knowledge element entities included in the preset knowledge graph
  • the preset atlas distance is artificially defined when the preset knowledge atlas is constructed according to the TransR model.
  • the expression form of the preset atlas distance can be adjusted according to actual needs.
  • This embodiment does not limit this, assuming that the mobile knowledge element entity is movie B
  • the corresponding mobile-associated knowledge meta-entities include movie D, movie J, movie P, movie E, movie Q, movie O, movie N, movie A, movie Z, and movie V. According to the search results, it is determined that the mobile-associated knowledge meta-entity and The mobile map distance of the mobile knowledge element entity.
  • Step S63 When the moving atlas distance belongs to the preset target distance range, use the moving associated knowledge element entity corresponding to the moving atlas distance as the moving target knowledge element entity.
  • the moving associated knowledge element entity corresponding to the moving atlas distance is taken as the moving target knowledge element entity.
  • the preset target distance range is artificially defined, and the size of the preset target distance range can be adjusted according to actual needs. This embodiment is not limited to this.
  • the mobile knowledge element entity is the mobile association corresponding to movie B
  • the knowledge meta-entity includes movie D, movie J, movie P, movie E, movie Q, movie O, movie N, movie A, movie Z, movie V, where the moving map distance belongs to the preset target distance range including movie D , Movie J, Movie P, Movie E, Movie Q, Movie O, Movie N, Movie A.
  • the method further includes: sorting the distances of the mobile atlases by size, and obtaining a result of the mobile sorting; and selecting a preset number of mobile target knowledge from the mobile-related knowledge element entities according to the result of the mobile sorting Meta entity.
  • the preset number is artificially defined, and the size of the preset number can be adjusted according to actual needs. This embodiment is not limited to this.
  • the mobile knowledge element entity that is, the mobile associated knowledge element entity corresponding to movie B includes Movie D, Movie J, Movie P, Movie E, Movie Q, Movie O, Movie N, Movie A, Movie Z, Movie V, wherein the preset number is 8, and the mobile map of the mobile-associated knowledge element entity
  • the distance is sorted by size, and the result of moving sorting is obtained as movie D, movie J, movie P, movie E, movie Q, movie O, movie N, movie A, movie Z, movie V, according to the mobile ranking result from the Select a preset number of 8 mobile target knowledge element entities from the mobile associated knowledge element entities, and obtain the mobile target knowledge element entities as movie D, movie J, movie P, movie E, movie Q, movie O, movie N, movie A.
  • Step S64 Display the moving target knowledge element entity and the mobile knowledge element entity with the mobile knowledge element entity as the center.
  • the moving target knowledge element entity and the mobile knowledge element entity are displayed through the current interface. Specifically, it includes: taking the mobile knowledge element entity as the display center, arranging the moving target knowledge element entity evenly around the display center, and displaying the arrangement result through the current interface.
  • the display mode of the current interface can be artificially defined, and the display mode of the current interface can be adjusted according to actual needs, which is not limited in this embodiment.
  • Fig. 8 is a schematic diagram of the visualization of mobile center information according to an embodiment of the application; wherein, the mobile knowledge element entity is movie B, and the mobile target knowledge element entity is movie D (D in Fig.
  • Movie J J in Figure 8
  • Movie P P in Figure 8
  • Movie E E in Figure 8
  • Movie Q Q in Figure 8
  • Movie O O in Figure 8
  • Movie N N in Figure 8
  • movie A A in Figure 8
  • the user can perform mixed operations of moving and jumping any number of times. Moreover, in theory, the user can reach any entity that is a knowledge element through two operations: move and jump. The user can understand various situations intuitively and from multiple angles through two operation modes: move and jump, and according to the user's existing knowledge, can more conveniently retrieve the search that meets the user's imagination but is inconvenient to express through text or language as a search The result of words.
  • the user can also choose to use the user information as the display center.
  • the user can choose the user information such as the number of user ratings, the number of user views, and the number of user comments.
  • user ⁇ selects 10 movies as the central point from the user's views, extracts the knowledge element entity vectors of these 10 movies, and calculates the average value of the knowledge element entity vectors of these 10 movies as the central knowledge element entity selected by user ⁇ .
  • the central knowledge element entity selected by user ⁇ is used as the display center, and the central knowledge element entity is used as the initial knowledge element entity.
  • Steps S10-S40 can be performed according to the flowchart of the first embodiment of the knowledge graph-based information visualization method of this application. It is shown that, in addition, user ⁇ can also perform any number of dimensional jump and move operations. Refer to the flow diagrams of the third embodiment and the fourth embodiment of the knowledge graph-based information visualization method of this application, or you can re-select the central knowledge Re-search the meta-entity.
  • a mobile knowledge element entity is determined based on the mobile knowledge element entity information input by the user based on the current interface; the mobile knowledge element entity corresponding to the mobile knowledge element entity is searched in a preset knowledge graph, and the mobile knowledge element entity is determined.
  • an embodiment of the present application also proposes a storage medium that stores an information visualization program based on a knowledge graph, and when the information visualization program based on the knowledge graph is executed by a processor, the above-mentioned knowledge-based The steps of the information visualization method of the atlas.
  • FIG. 9 is a structural block diagram of a first embodiment of an information visualization device based on a knowledge graph of this application.
  • the information visualization device based on the knowledge graph proposed in the embodiment of the present application includes:
  • the input module 10 is used to determine the initial knowledge element entity according to the knowledge element entity information input by the user.
  • a preset knowledge graph before the step of determining the initial knowledge meta-entity according to the knowledge meta-entity information input by the user, a preset knowledge graph needs to be constructed, where constructing the preset knowledge graph includes obtaining corpus information of preset topic dimensions;
  • the corpus information is input into the TransR model corresponding to the preset topic dimension to obtain the corresponding knowledge element entity, the feature vector of the knowledge element entity, and the preset map distance between the knowledge element entities;
  • the knowledge element entity, the feature vector of the knowledge element entity, and the predetermined map distance construct a predetermined knowledge map.
  • the corpus information is input into the TransR model corresponding to the preset topic dimension.
  • the preferred model in this embodiment is the TransR model, and other models can also be used to construct the knowledge graph, which is not limited in this embodiment.
  • the initial knowledge meta-entity is determined according to the knowledge meta-entity information input by the user.
  • the movie domain is used as an example to construct a preset knowledge graph. It is assumed that movie A is a movie that the user has watched, and the user does not have a specific goal. For quick information retrieval and viewing, the user decides to use movie A as the starting point of the search, and then determines the initial knowledge element entity based on the knowledge element entity information input by the user. At this time, the initial knowledge element entity can be regarded as movie A.
  • the searching module 20 is configured to search for the associated knowledge element entity corresponding to the initial knowledge element entity in a preset knowledge graph, and determine the graph distance between the associated knowledge element entity and the initial knowledge element entity.
  • the step of searching for the associated knowledge element entity corresponding to the initial knowledge element entity in the preset knowledge graph, and determining the graph distance between the associated knowledge element entity and the initial knowledge element entity specifically includes : According to the initial knowledge element entity, extract the feature vector corresponding to the initial knowledge element entity from the preset knowledge graph; according to the feature vector and the knowledge element entities included in the preset knowledge graph Search for the associated knowledge element entity corresponding to the initial knowledge element entity in the preset map distance of, and determine the map distance between the associated knowledge element entity and the initial knowledge element entity according to the search result.
  • the initial knowledge meta-entity can be regarded as movie A
  • the feature vector corresponding to the initial knowledge meta-entity, namely movie A, extracted from the preset knowledge graph can be denoted as h.
  • the feature vector h and the prediction Set the preset map distance between the knowledge element entities included in the knowledge graph to find the associated knowledge element entity corresponding to the initial knowledge element entity, wherein the preset knowledge element entities included in the preset knowledge graph
  • the expression form of the preset map distance can be adjusted according to actual needs.
  • This embodiment is not limited, assuming that the initial knowledge element entity corresponds to movie A
  • the associated knowledge meta-entity includes movie B, movie C, movie D, movie E, movie F, movie G, movie H, movie I, movie J, and movie K. According to the search result, it is determined that the associated knowledge meta-entity and the initial The map distance of the knowledge element entity.
  • the target module 30 is configured to determine that the map distance belongs to a preset target distance range, and use the associated knowledge element entity corresponding to the map distance as the target knowledge element entity.
  • the associated knowledge element entity corresponding to the map distance is used as the target knowledge element entity.
  • the preset target distance range is artificially defined, and the size of the preset target distance range can be adjusted according to actual needs. This embodiment does not limit this, assuming that the initial knowledge element entity is the associated knowledge corresponding to movie A
  • the meta-entity includes movie B, movie C, movie D, movie E, movie F, movie G, movie H, movie I, movie J, movie K, where the map distance belongs to the preset target distance range including movie B, movie C, movie D, movie E, movie F, movie G, movie H, movie I.
  • the method further includes: sorting the map distance by size, and obtaining a sorting result; and selecting a preset number of target knowledge element entities from the associated knowledge element entities according to the sorting result.
  • the preset number is artificially defined, and the size of the preset number can be adjusted according to actual needs. This embodiment is not limited to this. It is assumed that the initial knowledge element entity, that is, the associated knowledge element entity corresponding to movie A, includes movie B.
  • the target knowledge meta-entities with a preset number of 8 are selected, and the target knowledge meta-entities are obtained as movie B, movie C, movie D, movie E, movie F, movie G, movie H, and movie I.
  • the display module 40 is configured to display the target knowledge element entity and the initial knowledge element entity through the current interface.
  • the step of displaying the target knowledge element entity and the initial knowledge element entity through the current interface specifically includes: taking the initial knowledge element entity as the display center, and displaying the target knowledge element entity Arrange evenly around the display center, and display the arrangement result through the current interface.
  • the display mode of the current interface can be artificially defined, and the display mode of the current interface can be adjusted according to actual needs, which is not limited in this embodiment.
  • Fig. 3 is a schematic diagram of general dimensional information visualization according to an embodiment of this application; wherein, the initial knowledge element entity is movie A (A in Fig. 3), and the target knowledge element entity is movie B (such as B in Figure 3), Movie C (C in Figure 3), Movie D (D in Figure 3), Movie E (E in Figure 3), Movie F (F in Figure 3), Movie G (such as In Fig. 3, G), movie H (H in Fig. 3), and movie I (I in Fig. 3) are evenly arranged around the display center, that is, movie A (A in Fig. 3).
  • the input module 10 is used to determine the initial knowledge element entity according to the knowledge element entity information input by the user;
  • the search module 20 is used to search for the associated knowledge element entity corresponding to the initial knowledge element entity in the preset knowledge graph, And determine the map distance between the associated knowledge element entity and the initial knowledge element entity;
  • the target module 30 is used to determine that the map distance belongs to a preset target distance range, and the associated knowledge element entity corresponding to the map distance is used as the target Knowledge element entity;
  • the display module 40 is used to display the target knowledge element entity and the initial knowledge element entity through the current interface.
  • the technical solution of this application essentially or the part that contributes to the existing technology can be embodied in the form of a software product, and the computer software product is stored in a storage medium (such as a read-only memory (Read Only Memory)). , ROM)/RAM, magnetic disk, optical disk), including several instructions to make a terminal device (can be a mobile phone, computer, server, or network device, etc.) execute the method described in each embodiment of this application.
  • a storage medium such as a read-only memory (Read Only Memory)
  • ROM read-only memory
  • RAM magnetic disk
  • optical disk including several instructions to make a terminal device (can be a mobile phone, computer, server, or network device, etc.) execute the method described in each embodiment of this application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Animal Behavior & Ethology (AREA)
  • Computational Linguistics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

一种基于知识图谱的信息可视化方法、装置、设备及存储介质，属于知识图谱技术领域。该方法包括根据用户输入的知识元实体信息确定初始知识元实体（S10）；在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离（S20）；确定所述图谱距离属于预设目标距离范围，将所述图谱距离对应的关联知识元实体作为目标知识元实体（S30）；通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示（S40）。

Description

基于知识图谱的信息可视化方法、装置、设备及存储介质
本申请要求2019年12月30日申请的，申请号为201911424386.8，名称为“基于知识图谱的信息可视化方法、装置、设备及存储介质”的中国专利申请的优先权，在此将其全文引入作为参考。
技术领域
本申请涉及知识图谱技术领域，尤其涉及一种基于知识图谱的信息可视化方法、装置、设备及存储介质。
 背景技术
知识图谱技术由Google于2012年5月正式提出，其初衷在于提高搜索引擎的能力，增强用户的搜索质量与体验。目前，随着智能信息服务应用的不断发展，知识图谱技术已被广泛应用于智能搜索、智能问答、个性化推荐等领域。知识图谱技术涉及的数据存储及使用形式，为机器理解人类世界的知识起到了重要的推进作用。知识图谱技术把复杂的知识领域通过数据挖掘、信息处理、知识计量和图形绘制而显示出来，揭示知识领域的动态发展规律，为学科研究提供切实的、有价值的参考。
知识图谱技术产生自搜索引擎，也是在这一领域得到最大的发展与应用。现在，更多的应用产品会考虑借助知识图谱技术带来的信息，优化推荐、智能问答等其它信息系统。但是现有应用产品借助知识图谱技术优化推荐、智能问答时，需要用户提供明确目标化的信息，因此用户在并无特定目标时的信息检索与查看十分不便，造成用户在资源检索方面的体验感较差。
上述内容仅用于辅助理解本申请的技术方案，并不代表承认上述内容是现有技术。
 技术解决方案
本申请的主要目的在于提供一种基于知识图谱的信息可视化方法、装置、设备及存储介质，旨在解决现有技术用户在并无特定目标时的信息检索与查看十分不便以及体验感差的技术问题。
为实现上述目的，本申请提供了一种基于知识图谱的信息可视化方法，所述方法包括以下步骤：
根据用户输入的知识元实体信息确定初始知识元实体；
在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离；
确定所述图谱距离属于预设目标距离范围，将所述图谱距离对应的关联知识元实体作为目标知识元实体；
通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示。
此外，为实现上述目的，本申请还提出一种电子设备，所述设备包括：存储器、处理器及存储在所述存储器上并可在所述处理器上运行的基于知识图谱的信息可视化程序，所述基于知识图谱的信息可视化程序配置为实现如上文所述的基于知识图谱的信息可视化方法的步骤。
此外，为实现上述目的，本申请还提出一种存储介质，所述存储介质上存储有基于知识图谱的信息可视化程序，所述基于知识图谱的信息可视化程序被处理器执行时实现如上文所述的基于知识图谱的信息可视化方法的步骤。
本申请通过根据用户输入的知识元实体信息确定初始知识元实体；在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离；确定所述图谱距离属于预设目标距离范围，将所述图谱距离对应的关联知识元实体作为目标知识元实体；通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示。
通过上述方式方便用户在并无特定目标时的信息快速检索与查看，提升用户在资源检索方面的体验感，解决了现有技术用户在并无特定目标时的信息检索与查看十分不便以及体验感差的技术问题。
 附图说明
图1是本申请实施例方案涉及的硬件运行环境的电子设备的结构示意图；
图2为本申请基于知识图谱的信息可视化方法第一实施例的流程示意图；
图3为本申请实施例的通用维度信息可视化示意图；
图4为本申请基于知识图谱的信息可视化方法第二实施例的流程示意图；
图5为本申请基于知识图谱的信息可视化方法第三实施例的流程示意图；
图6为本申请实施例的跳转维度信息可视化示意图；
图7为本申请基于知识图谱的信息可视化方法第四实施例的流程示意图；
图8为本申请实施例的移动中心信息可视化示意图；
图9为本申请基于知识图谱的信息可视化装置第一实施例的结构框图。
本申请目的的实现、功能特点及优点将结合实施例，参照附图做进一步说明。
 本发明的实施方式
应当理解，此处所描述的具体实施例仅用以解释本申请并不用于限定本申请。
参照图1，图1为本申请实施例涉及的硬件运行环境的电子设备结构示意图。
如图1所示，该电子设备可以包括：处理器1001，例如中央处理器（Central Processing Unit，CPU），通信总线1002、用户接口1003，网络接口1004，存储器1005。其中，通信总线1002用于实现这些组件之间的连接通信。用户接口1003可以包括显示屏（Display）、输入单元比如键盘（Keyboard），可选用户接口1003还可以包括标准的有线接口、无线接口。网络接口1004可选的可以包括标准的有线接口、无线接口（如无线保真（WIreless-FIdelity，WI-FI）接口）。存储器1005可以是高速的随机存取存储器（Random Access Memory，RAM）存储器，也可以是稳定的非易失性存储器（Non-Volatile Memory，NVM），例如磁盘存储器。存储器1005可选的还可以是独立于前述处理器1001的存储装置。
本领域技术人员可以理解，图1中示出的结构并不构成对电子设备的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置。
如图1所示，作为一种存储介质的存储器1005中可以包括操作系统、网络通信模块、用户接口模块以及基于知识图谱的信息可视化程序。
在图1所示的电子设备中，网络接口1004主要用于与网络服务器进行数据通信；用户接口1003主要用于与用户进行数据交互；本申请电子设备中的处理器1001、存储器1005可以设置在电子设备中，所述电子设备通过处理器1001调用存储器1005中存储的基于知识图谱的信息可视化程序，并执行本申请实施例提供的基于知识图谱的信息可视化方法。
本申请实施例提供了一种基于知识图谱的信息可视化方法，参照图2，图2为本申请一种基于知识图谱的信息可视化方法第一实施例的流程示意图。
本实施例中，所述基于知识图谱的信息可视化方法包括以下步骤：
步骤S10：根据用户输入的知识元实体信息确定初始知识元实体。
在本实施例中，所述根据用户输入的知识元实体信息确定初始知识元实体的步骤之前，需要构建预设知识图谱，其中，构建预设知识图谱包括获取预设主题维度的语料信息；将所述语料信息输入至所述预设主题维度对应的TransR模型中，以获得对应的知识元实体、所述知识元实体的特征向量以及所述知识元实体之间的预设图谱距离；根据所述知识元实体、所述知识元实体的特征向量及所述预设图谱距离构建预设知识图谱。其中，将所述语料信息输入至所述预设主题维度对应的TransR模型中，本实施例优选模型为TransR模型，也可以使用其他模型构建知识图谱，本实施例对此并不加以限制。
易于理解的是，根据用户输入的知识元实体信息确定初始知识元实体，例如以电影域为例构建预设知识图谱，假设电影A是用户观看过的一部电影，用户并无特定目标时需要进行信息快速检索与查看，用户决定以电影A作为搜索的起点，则根据用户输入的知识元实体信息确定初始知识元实体，此时初始知识元实体可以视为电影A。
步骤S20：在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离。
需要说明的是，所述在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离的步骤，具体包括：根据所述初始知识元实体，在所述预设知识图谱中提取所述初始知识元实体对应的特征向量；根据所述特征向量及所述预设知识图谱中包含的各知识元实体之间的预设图谱距离查找所述初始知识元实体对应的关联知识元实体；根据查找结果确定所述关联知识元实体与所述初始知识元实体的图谱距离。
在具体实现中，初始知识元实体可以视为电影A，在预设知识图谱中提取所述初始知识元实体即电影A对应的特征向量可以记为h，根据所述特征向量h及所述预设知识图谱中包含的各知识元实体之间的预设图谱距离查找所述初始知识元实体对应的关联知识元实体，其中，所述预设知识图谱中包含的各知识元实体之间的预设图谱距离为根据TransR模型构建预设知识图谱时人为定义，预设图谱距离的表现形式可以根据实际需求进行调整，本实施例对此不加以限制，假设所述初始知识元实体即电影A对应的关联知识元实体包括电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I、电影J、电影K，根据查找结果确定所述关联知识元实体与所述初始知识元实体的图谱距离。
步骤S30：确定所述图谱距离属于预设目标距离范围，将所述图谱距离对应的关联知识元实体作为目标知识元实体。
具体地，在所述图谱距离属于预设目标距离范围时，将所述图谱距离对应的关联知识元实体作为目标知识元实体。其中，所述预设目标距离范围为人为定义，预设目标距离范围的大小可以根据实际需求进行调整，本实施例对此不加以限制，假设所述初始知识元实体即电影A对应的关联知识元实体包括电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I、电影J、电影K，其中所述图谱距离属于预设目标距离范围的包括电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I。
易于理解地是，所述在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离的步骤之后，所述方法还包括：对所述图谱距离进行大小排序，并获取排序结果；根据所述排序结果从所述关联知识元实体中选取预设数量的目标知识元实体。其中，所述预设数量为人为定义，预设数量的大小可以根据实际需求进行调整，本实施例对此不加以限制，假设所述初始知识元实体即电影A对应的关联知识元实体包括电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I、电影J、电影K，其中所述预设数量为8，对所述关联知识元实体的图谱距离进行大小排序，并获取排序结果为电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I、电影J、电影K，根据所述排序结果从所述关联知识元实体中选取预设数量为8的目标知识元实体，获得所述目标知识元实体为电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I。
步骤S40：通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示。
需要说明的是，所述通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示的步骤，具体包括：以所述初始知识元实体为显示中心，将所述目标知识元实体围绕所述显示中心进行均匀排布，并通过当前界面对排布结果进行显示。其中，所述当前界面的显示方式可以人为定义，所述当前界面的显示方式可以根据实际需求进行调整，本实施例对此不加以限制。
具体地，以所述初始知识元实体电影A为显示中心，将所述目标知识元实体即电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I围绕所述显示中心进行均匀排布，并通过当前界面对排布结果进行显示。如图3所示，图3为本申请实施例的通用维度信息可视化示意图；其中，所述初始知识元实体即电影A（如图3中A），所述目标知识元实体如电影B（如图3中B）、电影C（如图3中C）、电影D（如图3中D）、电影E（如图3中E）、电影F（如图3中F）、电影G（如图3中G）、电影H（如图3中H）、电影I（如图3中I）围绕所述显示中心即电影A（如图3中A）进行均匀排布。
本实施例通过根据用户输入的知识元实体信息确定初始知识元实体；在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离；确定所述图谱距离属于预设目标距离范围，将所述图谱距离对应的关联知识元实体作为目标知识元实体；通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示。通过上述方式方便用户在并无特定目标时的信息快速检索与查看，提升用户在资源检索方面的体验感，解决了现有技术用户在并无特定目标时的信息检索与查看十分不便以及体验感差的技术问题。
参考图4，图4为本申请一种基于知识图谱的信息可视化方法第二实施例的流程示意图。
基于上述第一实施例，本实施例基于知识图谱的信息可视化方法在所述步骤S10之前，还包括：
步骤S101：获取预设主题维度的语料信息。
需要说明的是，获取待构建知识图谱的至少一个领域的语料信息，语料信息可以包括：多个知识文本内容片段。可以通过知识图谱构建系统构建知识图谱，知识图谱构建系统可以为计算机、服务器等硬件设备或者安装在硬件设备上的软件。其中，领域可以指专业领域，例如“冶金”领域、“经济”领域、“医学”领域等，领域可以有多个子领域，例如“医学”领域下面有“儿科医学”领域。语料信息是指知识文本内容片段，本实施例中涉及到的待构建知识图谱的语料信息可以为电影“追龙”。知识文本内容片段例如“追龙这部电影是由刘德华主演的”。
步骤S102：将所述语料信息输入至所述预设主题维度对应的TransR模型中，获得对应的知识元实体、所述知识元实体的特征向量以及所述知识元实体之间的预设图谱距离。
易于理解地是，将所述语料信息输入至所述预设主题维度对应的TransR模型中，本实施例优选模型为TransR模型，也可以使用其他模型构建知识图谱，本实施例对此并不加以限制。例如，国内外的知识图谱表示方法代表性工作主要包括基于翻译的嵌入式模型TransE，基于超平面的嵌入式模型TransH，基于实体关系空间的嵌入式模型TransR，基于聚类和实体关系空间的嵌入式模型CTransR和基于动态映射矩阵的嵌入式模型TransD等方法，上述方法被统称为基于翻译的知识表示模型。其中，基于实体关系空间的嵌入式模型TransR通过建立一个映像矩阵Mr和一个向量r来表示每一个关系r，具体地，TransR将头实体向量h和尾实体向量t通过矩阵映射到关系向量r的层次上，得到Mrh+r＝Mrt，也即基于实体关系空间的嵌入式模型TransR的优化目标。
具体地，将所述语料信息输入至所述预设主题维度对应的TransR模型中，以获得对应的知识元实体、所述知识元实体的特征向量以及所述知识元实体之间的预设图谱距离。将所述语料信息输入至TransR模型中，语料信息可以包括：多个知识文本内容片段，对知识文本内容片段进行分词以及词性标注，获取知识文本内容片段中的关键词，将关键词按照预设的规则与领域本体进行匹配，获取知识文本内容片段中的知识元实例、知识元实例的属性以及知识元实例之间的关联关系；领域本体包括：领域主题，领域主题包括的至少一个模型，模型的属性以及模型之间的关联关系；模型包括至少一个知识元实例。其中，领域主题可以指专业领域或者子领域，专业领域可以有多个子领域，每个领域内会有多个模型，每个模型会有自己特有的属性，模型之间会有各种关联关系。通过TransR模型将所述语料信息翻译为对应的知识元实体、所述知识元实体的特征向量及各知识元实体之间的预设图谱距离。其中，所述预设知识图谱中包含的各知识元实体之间的预设图谱距离为根据TransR模型构建预设知识图谱时人为定义，预设图谱距离的表现形式可以根据实际需求进行调整，本实施例对此不加以限制，
步骤S103：根据所述知识元实体、所述知识元实体的特征向量及所述预设图谱距离构建预设知识图谱。
本实施例通过获取预设主题维度的语料信息；将所述语料信息输入至所述预设主题维度对应的TransR模型中，以获得对应的知识元实体、所述知识元实体的特征向量以及所述知识元实体之间的预设图谱距离；根据所述知识元实体、所述知识元实体的特征向量及所述预设图谱距离构建预设知识图谱。通过上述方式借助预设知识图谱自身的可视化特性，为用户提供非明确目标化的信息检索，方便用户在并无特定目标时的信息快速检索与查看，提升用户在资源检索方面的体验感。
参考图5，图5为本申请一种基于知识图谱的信息可视化方法第三实施例的流程示意图。
基于上述第一实施例，本实施例基于知识图谱的信息可视化方法在所述步骤S40之后，还包括：
步骤S51：根据用户基于所述当前界面输入的跳转知识元实体信息确定跳转知识元实体。
易于理解的是，通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示。具体地，以所述初始知识元实体电影A为显示中心，将所述目标知识元实体即电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I围绕所述显示中心进行均匀排布，并通过当前界面对排布结果进行显示。根据用户基于所述当前界面输入的跳转知识元实体信息确定跳转知识元实体，具体地，如果用户选择维度跳转，需要选择跳转的维度，用户基于所述当前界面输入的跳转知识元实体信息确定跳转知识元实体，假设跳转知识元实体为“主演”关系维度。
步骤S52：在预设知识图谱中查找所述跳转知识元实体对应的投影矩阵，并根据所述投影矩阵更新所述预设知识图谱中各知识元实体之间的预设图谱距离，获得新的知识图谱。
需要说明的是，如果用户选择维度跳转，需要选择跳转的维度，用户基于所述当前界面输入的跳转知识元实体信息确定跳转知识元实体，假设跳转知识元实体为“主演”关系维度。在预设知识图谱中查找所述跳转知识元实体为“主演”关系维度时对应的投影矩阵记为Mr，根据所述投影矩阵更新所述预设知识图谱中各知识元实体之间的预设图谱距离，即对所述预设知识图谱中各知识元实体左乘投影矩阵Mr，以获得新的知识图谱。
步骤S53：在所述新的知识图谱中查找所述初始知识元实体对应的跳转关联知识元实体，并确定所述跳转关联知识元实体与所述初始知识元实体之间的跳转图谱距离。
需要说明的是，所述在所述新的知识图谱中查找所述初始知识元实体对应的跳转关联知识元实体，并确定所述跳转关联知识元实体与所述初始知识元实体之间的跳转图谱距离的步骤，具体包括：根据所述初始知识元实体，在所述新的知识图谱中提取所述初始知识元实体对应的特征向量；根据所述特征向量及所述新的知识图谱中包含的各知识元实体之间的预设图谱距离查找所述初始知识元实体对应的跳转关联知识元实体；根据查找结果确定所述跳转关联知识元实体与所述初始知识元实体的跳转图谱距离。
在具体实现中，若跳转知识元实体为“主演”关系维度，在预设知识图谱中查找所述跳转知识元实体为“主演”关系维度时对应的投影矩阵记为Mr，根据所述投影矩阵更新所述预设知识图谱中各知识元实体之间的预设图谱距离，即对所述预设知识图谱中各知识元实体左乘投影矩阵Mr，以获得新的知识图谱，在所述新的知识图谱中提取所述初始知识元实体对应的特征向量可以记为h1，根据所述特征向量h1及所述新的知识图谱中包含的各知识元实体之间的预设图谱距离查找所述初始知识元实体对应的跳转关联知识元实体，假设跳转知识元实体为“主演”关系维度，所述初始知识元实体对应的跳转关联知识元实体包括电影B、电影J、电影M、电影E、电影K、电影L、电影H、电影I、电影X、电影Y，根据查找结果确定所述跳转关联知识元实体与所述初始知识元实体之间的跳转图谱距离。
步骤S54：在所述跳转图谱距离属于所述预设目标距离范围时，将所述跳转图谱距离对应的跳转关联知识元实体作为跳转目标知识元实体。
具体地，在所述跳转图谱距离属于预设目标距离范围时，将所述跳转图谱距离对应的跳转关联知识元实体作为跳转目标知识元实体。其中，所述预设目标距离范围为人为定义，预设目标距离范围的大小可以根据实际需求进行调整，本实施例对此不加以限制，假设跳转知识元实体为“主演”关系维度，所述初始知识元实体对应的跳转关联知识元实体包括电影B、电影J、电影M、电影E、电影K、电影L、电影H、电影I、电影X、电影Y，其中所述跳转图谱距离属于预设目标距离范围的包括电影B、电影J、电影M、电影E、电影K、电影L、电影H、电影I。
易于理解地是，所述方法还包括：对所述跳转图谱距离进行大小排序，并获取排序结果；根据所述排序结果从所述跳转关联知识元实体中选取预设数量的跳转目标知识元实体。其中，所述预设数量为人为定义，预设数量的大小可以根据实际需求进行调整，本实施例对此不加以限制，假设跳转知识元实体为“主演”关系维度，所述初始知识元实体对应的跳转关联知识元实体包括电影B、电影J、电影M、电影E、电影K、电影L、电影H、电影I、电影X、电影Y，其中所述预设数量为8，对所述跳转关联知识元实体的跳转图谱距离进行大小排序，并获取跳转排序结果为电影B、电影J、电影M、电影E、电影K、电影L、电影H、电影I、电影X、电影Y，根据跳转排序结果从所述跳转关联知识元实体中选取预设数量为8的跳转目标知识元实体，获得所述跳转目标知识元实体为电影B、电影J、电影M、电影E、电影K、电影L、电影H、电影I。
步骤S55：以所述初始知识元实体为中心，对所述初始知识元实体及所述跳转目标知识元实体进行显示。
需要说明的是，通过当前界面对对所述初始知识元实体及所述跳转目标知识元实体进行显示。其中，所述当前界面的显示方式可以人为定义，所述当前界面的显示方式可以根据实际需求进行调整，本实施例对此不加以限制。
具体地，以所述初始知识元实体电影A为显示中心，将所述跳转目标知识元实体即电影B、电影J、电影M、电影E、电影K、电影L、电影H、电影I围绕所述显示中心进行均匀排布，并通过当前界面对排布结果进行显示。如图6所示，图6为本申请实施例的跳转维度信息可视化示意图；其中，所述初始知识元实体即电影A（如图6中A），所述跳转目标知识元实体如电影B（如图6中B）、电影J（如图6中J）、电影M（如图6中M）、电影E（如图6中E）、电影K（如图6中K）、电影L（如图6中L）、电影H（如图6中H）、电影I（如图6中I）围绕所述显示中心即电影A（如图6中A）进行均匀排布。
易于理解的是，若现实中考虑一家公司两位员工，员工A及员工B之间的图谱距离即员工A及员工B之间的相关性，若在工作的关系维度去考虑员工A与员工B之间的相关性，可能因为有相同的领导上司存在员工A与员工B的相关性程度高的情况，若在血缘的关系维度去考虑员工A与员工B的图谱距离即员工A与员工B之间的相关性，员工A与员工B可能存在相关性程度低的情况。但是不会因为维度跳转到血缘的关系维度就列出员工A的血缘亲属与员工B的血缘亲属，考虑的仍然是员工A与员工B之间的图谱距离即员工A与员工B之间的相关性。提到员工A与员工B的领导上司及血缘亲属只是为了更好理解员工A与员工B之间距离远近及相关性程度高低，实际上在进行维度跳转时，通过根据用户基于所述当前界面输入的跳转知识元实体信息确定跳转知识元实体；在预设知识图谱中查找所述跳转知识元实体对应的投影矩阵，并根据所述投影矩阵更新所述预设知识图谱中各知识元实体之间的预设图谱距离，以获得新的知识图谱；在所述新的知识图谱中，针对员工A与员工B进行图谱距离计算。
需要说明的是，构建预设知识图谱时为了提升个性化能力，可以采集诸如用户评分数、用户浏览量、用户评论数等属性值，以电影领域为例，使用用户评分数、用户浏览量、用户评论数等属性值构建预设知识图谱，其中，每部电影依然是实体，“用户浏览量”，“用户评分数”，“用户评论数”等为预设知识图谱中新的关系，每个用户是新的实体，则知识元实体的三元组(h，r，t)数组可以解析为诸如用户α曾经观看过电影A，再加上获取预设主题维度的语料信息构建预设知识图谱，更新预设知识图谱，增加了用户实体的特征向量和行为（如用户评分数、用户浏览量、用户评论数）的关系维度的投影矩阵。因此根据步骤S51至步骤S55，可以增加一个跳转操作的目标，称之为用户维度，所述用户维度对应着实际的物理含义，表示用户群体对产品不同的喜恶感受带来的产品的内在关系，更贴近于实际应用场景。
本实施例通过根据用户基于所述当前界面输入的跳转知识元实体信息确定跳转知识元实体；在预设知识图谱中查找所述跳转知识元实体对应的投影矩阵，并根据所述投影矩阵更新所述预设知识图谱中各知识元实体之间的预设图谱距离，以获得新的知识图谱；在所述新的知识图谱中查找所述初始知识元实体对应的跳转关联知识元实体，并确定所述跳转关联知识元实体与所述初始知识元实体之间的跳转图谱距离；在所述跳转图谱距离属于所述预设目标距离范围时，将所述跳转图谱距离对应的跳转关联知识元实体作为跳转目标知识元实体；以所述初始知识元实体为中心，对所述初始知识元实体及所述跳转目标知识元实体进行显示。通过上述方式进行维度跳转，解决了现有技术用户在并无特定目标时的信息检索与查看十分不便以及体验感差的技术问题。
参考图7，图7为本申请一种基于知识图谱的信息可视化方法第四实施例的流程示意图。
基于上述第一实施例，本实施例基于知识图谱的信息可视化方法在所述步骤S40之后，还包括：
步骤S61：根据用户基于所述当前界面输入的移动知识元实体信息确定移动知识元实体。
易于理解的是，通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示。具体地，以所述初始知识元实体电影A为显示中心，将所述目标知识元实体即电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I围绕所述显示中心进行均匀排布，并通过当前界面对排布结果进行显示。根据用户基于所述当前界面输入的移动知识元实体信息确定移动知识元实体，具体地，如果用户选择中心移动，需要重新选择中心的知识元实体，用户基于所述当前界面输入的移动知识元实体信息确定移动知识元实体，假设移动知识元实体为电影B。
步骤S62：在预设知识图谱中查找所述移动知识元实体对应的移动关联知识元实体，并确定所述移动关联知识元实体与所述移动知识元实体的移动图谱距离。
需要说明的是，所述在预设知识图谱中查找所述移动知识元实体对应的移动关联知识元实体，并确定所述移动关联知识元实体与所述移动知识元实体的移动图谱距离的步骤，具体包括：根据所述移动知识元实体，在预设知识图谱中提取所述移动知识元实体对应的特征向量；根据所述特征向量及所述预设知识图谱中包含的各知识元实体之间的预设图谱距离查找所述移动知识元实体对应的移动关联知识元实体；根据查找结果确定所述移动关联知识元实体与所述移动知识元实体的移动图谱距离。
在具体实现中，移动知识元实体可以视为电影B，在预设知识图谱中提取所述移动知识元实体即电影B对应的特征向量可以记为h2，根据所述特征向量h2及所述预设知识图谱中包含的各知识元实体之间的预设图谱距离查找所述移动知识元实体对应的移动关联知识元实体，其中，所述预设知识图谱中包含的各知识元实体之间的预设图谱距离为根据TransR模型构建预设知识图谱时人为定义，预设图谱距离的表现形式可以根据实际需求进行调整，本实施例对此不加以限制，假设所述移动知识元实体即电影B对应的移动关联知识元实体包括电影D、电影J、电影P、电影E、电影Q、电影O、电影N、电影A、电影Z、电影V，根据查找结果确定所述移动关联知识元实体与所述移动知识元实体的移动图谱距离。
步骤S63：在所述移动图谱距离属于所述预设目标距离范围时，将所述移动图谱距离对应的移动关联知识元实体作为移动目标知识元实体。
具体地，在所述移动图谱距离属于预设目标距离范围时，将所述移动图谱距离对应的移动关联知识元实体作为移动目标知识元实体。其中，所述预设目标距离范围为人为定义，预设目标距离范围的大小可以根据实际需求进行调整，本实施例对此不加以限制，假设所述移动知识元实体即电影B对应的移动关联知识元实体包括电影D、电影J、电影P、电影E、电影Q、电影O、电影N、电影A、电影Z、电影V，其中所述移动图谱距离属于预设目标距离范围的包括电影D、电影J、电影P、电影E、电影Q、电影O、电影N、电影A。
易于理解地是，所述方法还包括：对所述移动图谱距离进行大小排序，并获取移动排序结果；根据所述移动排序结果从所述移动关联知识元实体中选取预设数量的移动目标知识元实体。其中，所述预设数量为人为定义，预设数量的大小可以根据实际需求进行调整，本实施例对此不加以限制，假设所述移动知识元实体即电影B对应的移动关联知识元实体包括电影D、电影J、电影P、电影E、电影Q、电影O、电影N、电影A、电影Z、电影V，其中所述预设数量为8，对所述移动关联知识元实体的移动图谱距离进行大小排序，并获取移动排序结果为电影D、电影J、电影P、电影E、电影Q、电影O、电影N、电影A、电影Z、电影V，根据所述移动排序结果从所述移动关联知识元实体中选取预设数量为8的移动目标知识元实体，获得所述移动目标知识元实体为电影D、电影J、电影P、电影E、电影Q、电影O、电影N、电影A。
步骤S64：以所述移动知识元实体为中心，对所述移动目标知识元实体及所述移动知识元实体进行显示。
需要说明的是，通过当前界面对所述移动目标知识元实体及所述移动知识元实体进行显示。具体包括：以所述移动知识元实体为显示中心，将所述移动目标知识元实体围绕所述显示中心进行均匀排布，并通过当前界面对排布结果进行显示。其中，所述当前界面的显示方式可以人为定义，所述当前界面的显示方式可以根据实际需求进行调整，本实施例对此不加以限制。
具体地，以所述移动知识元实体电影B为显示中心，将所述移动目标知识元实体即电影D、电影J、电影P、电影E、电影Q、电影O、电影N、电影A围绕所述显示中心进行均匀排布，并通过当前界面对排布结果进行显示。如图8所示，图8为本申请实施例的移动中心信息可视化示意图；其中，所述移动知识元实体即电影B，所述移动目标知识元实体如电影D（如图8中D）、电影J（如图8中J）、电影P（如图8中P）、电影E（如图8中E）、电影Q（如图8中Q）、电影O（如图8中O）、电影N（如图8中N）、电影A（如图8中A）围绕所述显示中心即电影B（如图8中B）进行均匀排布。
需要说明的是，用户可以进行任意次的移动与跳转混合操作。并且，理论上用户可以通过移动与跳转两种操作抵达任意一个是知识元实体。用户通过移动与跳转两种操作形式，可以直观且多角度地了解各项情况，并且根据用户已有认知，可以更为便捷地检索到符合用户想象但是不方便通过文字或者语言表达为搜索话术的结果。
此外，用户也可以选择将用户信息作为显示中心，例如用户可以选择用户信息如用户评分数、用户浏览量、用户评论数等。例如用户α从用户浏览量中选择了10部电影作为中心点，提取这10部电影的知识元实体向量，计算这10部电影的知识元实体向量的均值作为用户α选择的中心知识元实体，也可以请用户提供这10部电影的权重计算加权均值作为用户α选择的中心知识元实体。将用户α选择的中心知识元实体作为显示中心，将所述中心知识元实体作为初始知识元实体，可以根据本申请基于知识图谱的信息可视化方法第一实施例的流程示意图中步骤S10-S40进行显示，此外，用户α还可以进行任意次的维度跳转与移动混合操作，参见本申请基于知识图谱的信息可视化方法第三实施例及第四实施例的流程示意图，或者也可以重新选择中心知识元实体重新进行检索。
本实施例通过根据用户基于所述当前界面输入的移动知识元实体信息确定移动知识元实体；在预设知识图谱中查找所述移动知识元实体对应的移动关联知识元实体，并确定所述移动关联知识元实体与所述移动知识元实体的移动图谱距离；在所述移动图谱距离属于所述预设目标距离范围时，将所述移动图谱距离对应的移动关联知识元实体作为移动目标知识元实体；以所述移动知识元实体为中心，对所述移动目标知识元实体及所述移动知识元实体进行显示。通过上述方式进行中心移动，解决了现有技术用户在并无特定目标时的信息检索与查看十分不便以及体验感差的技术问题。
此外，本申请实施例还提出一种存储介质，所述存储介质上存储有基于知识图谱的信息可视化程序，所述基于知识图谱的信息可视化程序被处理器执行时实现如上文所述的基于知识图谱的信息可视化方法的步骤。
参照图9，图9为本申请基于知识图谱的信息可视化装置第一实施例的结构框图。
如图9所示，本申请实施例提出的基于知识图谱的信息可视化装置包括：
输入模块10，用于根据用户输入的知识元实体信息确定初始知识元实体。
在本实施例中，所述根据用户输入的知识元实体信息确定初始知识元实体的步骤之前，需要构建预设知识图谱，其中，构建预设知识图谱包括获取预设主题维度的语料信息；将所述语料信息输入至所述预设主题维度对应的TransR模型中，以获得对应的知识元实体、所述知识元实体的特征向量以及所述知识元实体之间的预设图谱距离；根据所述知识元实体、所述知识元实体的特征向量及所述预设图谱距离构建预设知识图谱。其中，将所述语料信息输入至所述预设主题维度对应的TransR模型中，本实施例优选模型为TransR模型，也可以使用其他模型构建知识图谱，本实施例对此并不加以限制。
易于理解的是，根据用户输入的知识元实体信息确定初始知识元实体，例如以电影域为例构建预设知识图谱，假设电影A是用户观看过的一部电影，用户并无特定目标时需要进行信息快速检索与查看，用户决定以电影A作为搜索的起点，则根据用户输入的知识元实体信息确定初始知识元实体，此时初始知识元实体可以视为电影A。
查找模块20，用于在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离。
需要说明的是，所述在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离的步骤，具体包括：根据所述初始知识元实体，在所述预设知识图谱中提取所述初始知识元实体对应的特征向量；根据所述特征向量及所述预设知识图谱中包含的各知识元实体之间的预设图谱距离查找所述初始知识元实体对应的关联知识元实体；根据查找结果确定所述关联知识元实体与所述初始知识元实体的图谱距离。
在具体实现中，初始知识元实体可以视为电影A，在预设知识图谱中提取所述初始知识元实体即电影A对应的特征向量可以记为h，根据所述特征向量h及所述预设知识图谱中包含的各知识元实体之间的预设图谱距离查找所述初始知识元实体对应的关联知识元实体，其中，所述预设知识图谱中包含的各知识元实体之间的预设图谱距离为根据TransR模型构建预设知识图谱时人为定义，预设图谱距离的表现形式可以根据实际需求进行调整，本实施例对此不加以限制，假设所述初始知识元实体即电影A对应的关联知识元实体包括电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I、电影J、电影K，根据查找结果确定所述关联知识元实体与所述初始知识元实体的图谱距离。
目标模块30，用于确定所述图谱距离属于预设目标距离范围，将所述图谱距离对应的关联知识元实体作为目标知识元实体。
具体地，在所述图谱距离属于预设目标距离范围时，将所述图谱距离对应的关联知识元实体作为目标知识元实体。其中，所述预设目标距离范围为人为定义，预设目标距离范围的大小可以根据实际需求进行调整，本实施例对此不加以限制，假设所述初始知识元实体即电影A对应的关联知识元实体包括电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I、电影J、电影K，其中所述图谱距离属于预设目标距离范围的包括电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I。
易于理解地是，所述在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离的步骤之后，所述方法还包括：对所述图谱距离进行大小排序，并获取排序结果；根据所述排序结果从所述关联知识元实体中选取预设数量的目标知识元实体。其中，所述预设数量为人为定义，预设数量的大小可以根据实际需求进行调整，本实施例对此不加以限制，假设所述初始知识元实体即电影A对应的关联知识元实体包括电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I、电影J、电影K，其中所述预设数量为8，对所述关联知识元实体的图谱距离进行大小排序，并获取排序结果为电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I、电影J、电影K，根据所述排序结果从所述关联知识元实体中选取预设数量为8的目标知识元实体，获得所述目标知识元实体为电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I。
显示模块40，用于通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示。
需要说明的是，所述通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示的步骤，具体包括：以所述初始知识元实体为显示中心，将所述目标知识元实体围绕所述显示中心进行均匀排布，并通过当前界面对排布结果进行显示。其中，所述当前界面的显示方式可以人为定义，所述当前界面的显示方式可以根据实际需求进行调整，本实施例对此不加以限制。
具体地，以所述初始知识元实体电影A为显示中心，将所述目标知识元实体即电影B、电影C、电影D、电影E、电影F、电影G、电影H、电影I围绕所述显示中心进行均匀排布，并通过当前界面对排布结果进行显示。如图3所示，图3为本申请实施例的通用维度信息可视化示意图；其中，所述初始知识元实体即电影A（如图3中A），所述目标知识元实体如电影B（如图3中B）、电影C（如图3中C）、电影D（如图3中D）、电影E（如图3中E）、电影F（如图3中F）、电影G（如图3中G）、电影H（如图3中H）、电影I（如图3中I）围绕所述显示中心即电影A（如图3中A）进行均匀排布。
本实施例通过输入模块10，用于根据用户输入的知识元实体信息确定初始知识元实体；查找模块20，用于在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离；目标模块30，用于确定所述图谱距离属于预设目标距离范围，将所述图谱距离对应的关联知识元实体作为目标知识元实体；显示模块40，用于通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示。通过上述方式方便用户在并无特定目标时的信息快速检索与查看，提升用户在资源检索方面的体验感，解决了现有技术用户在并无特定目标时的信息检索与查看十分不便以及体验感差的技术问题。
应当理解的是，以上仅为举例说明，对本申请的技术方案并不构成任何限定，在具体应用中，本领域的技术人员可以根据需要进行设置，本申请对此不做限制。
需要说明的是，以上所描述的工作流程仅仅是示意性的，并不对本申请的保护范围构成限定，在实际应用中，本领域的技术人员可以根据实际的需要选择其中的部分或者全部来实现本实施例方案的目的，此处不做限制。
另外，未在本实施例中详尽描述的技术细节，可参见本申请任意实施例所提供的基于知识图谱的信息可视化方法，此处不再赘述。
此外，需要说明的是，在本文中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者系统不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者系统所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、方法、物品或者系统中还存在另外的相同要素。
上述本申请实施例序号仅仅为了描述，不代表实施例的优劣。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质（如只读存储器（Read Only Memory，ROM）/RAM、磁碟、光盘）中，包括若干指令用以使得一台终端设备（可以是手机，计算机，服务器，或者网络设备等）执行本申请各个实施例所述的方法。
以上仅为本申请的优选实施例，并非因此限制本申请的专利范围，凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换，或直接或间接运用在其他相关的技术领域，均同理包括在本申请的专利保护范围内。

Claims (20)

  1. 一种基于知识图谱的信息可视化方法，其中，所述方法包括：
    根据用户输入的知识元实体信息确定初始知识元实体；
    在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离；
    确定所述图谱距离属于预设目标距离范围，将所述图谱距离对应的关联知识元实体作为目标知识元实体；以及
    通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示。
  2. 如权利要求1所述的方法，其中，所述根据用户输入的知识元实体信息确定初始知识元实体的步骤之前，还包括：
    获取预设主题维度的语料信息；
    将所述语料信息输入至所述预设主题维度对应的TransR模型中，获得对应的知识元实体、所述知识元实体的特征向量以及所述知识元实体之间的预设图谱距离；以及
    根据所述知识元实体、所述知识元实体的特征向量及所述预设图谱距离构建预设知识图谱。
  3. 如权利要求2所述的方法，其中，所述在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离的步骤，具体包括：
    根据所述初始知识元实体，在所述预设知识图谱中提取所述初始知识元实体对应的特征向量；
    根据所述特征向量及所述预设知识图谱中包含的各知识元实体之间的预设图谱距离查找所述初始知识元实体对应的关联知识元实体；以及
    根据查找结果确定所述关联知识元实体与所述初始知识元实体的图谱距离。
  4. 如权利要求3所述的方法，其中，所述在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离的步骤之后，所述方法还包括：
    对所述图谱距离进行大小排序，并获取排序结果；以及
    根据所述排序结果从所述关联知识元实体中选取预设数量的目标知识元实体。
  5. 如权利要求4所述的方法，其中，所述通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示的步骤，具体包括：以所述初始知识元实体为显示中心，将所述目标知识元实体围绕所述显示中心进行均匀排布，并通过当前界面对排布结果进行显示。
  6. 如权利要求2所述的方法，其中，所述通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示的步骤之后，还包括：
    根据用户基于所述当前界面输入的跳转知识元实体信息确定跳转知识元实体；
    在预设知识图谱中查找所述跳转知识元实体对应的投影矩阵，并根据所述投影矩阵更新所述预设知识图谱中各知识元实体之间的预设图谱距离，获得新的知识图谱；
    在所述新的知识图谱中查找所述初始知识元实体对应的跳转关联知识元实体，并确定所述跳转关联知识元实体与所述初始知识元实体之间的跳转图谱距离；
    在所述跳转图谱距离属于所述预设目标距离范围时，将所述跳转图谱距离对应的跳转关联知识元实体作为跳转目标知识元实体；以及
    以所述初始知识元实体为中心，对所述初始知识元实体及所述跳转目标知识元实体进行显示。
  7. 如权利要求2所述的方法，其中，所述通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示的步骤之后，还包括：
    根据用户基于所述当前界面输入的移动知识元实体信息确定移动知识元实体；
    在预设知识图谱中查找所述移动知识元实体对应的移动关联知识元实体，并确定所述移动关联知识元实体与所述移动知识元实体的移动图谱距离；
    在所述移动图谱距离属于所述预设目标距离范围时，将所述移动图谱距离对应的移动关联知识元实体作为移动目标知识元实体；以及
    以所述移动知识元实体为中心，对所述移动目标知识元实体及所述移动知识元实体进行显示。
  8. 一种基于知识图谱的信息可视化装置，其特征在于，所述装置包括：
    输入模块，用于根据用户输入的知识元实体信息确定初始知识元实体；
    查找模块，用于在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离；
    目标模块，用于在所述图谱距离属于预设目标距离范围时，将所述图谱距离对应的关联知识元实体作为目标知识元实体；
    显示模块，用于通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示。
  9. 一种电子设备，其中，所述设备包括：存储器、处理器及存储在所述存储器上并可在所述处理器上运行的基于知识图谱的信息可视化程序，所述基于知识图谱的信息可视化程序配置为实现以下步骤：
    根据用户输入的知识元实体信息确定初始知识元实体；
    在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离；
    确定所述图谱距离属于预设目标距离范围，将所述图谱距离对应的关联知识元实体作为目标知识元实体；以及
    通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示。
  10. 如权利要求9所述的电子设备，其中，所述根据用户输入的知识元实体信息确定初始知识元实体的步骤之前，所述基于知识图谱的信息可视化程序还配置为实现：
    获取预设主题维度的语料信息；
    将所述语料信息输入至所述预设主题维度对应的TransR模型中，获得对应的知识元实体、所述知识元实体的特征向量以及所述知识元实体之间的预设图谱距离；以及
    根据所述知识元实体、所述知识元实体的特征向量及所述预设图谱距离构建预设知识图谱。
  11. 如权利要求10所述的电子设备，其中，所述基于知识图谱的信息可视化程序配置为实现的所述在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离的步骤，具体包括：
    根据所述初始知识元实体，在所述预设知识图谱中提取所述初始知识元实体对应的特征向量；
    根据所述特征向量及所述预设知识图谱中包含的各知识元实体之间的预设图谱距离查找所述初始知识元实体对应的关联知识元实体；以及
    根据查找结果确定所述关联知识元实体与所述初始知识元实体的图谱距离。
  12. 如权利要求11所述的电子设备，其中，在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离的步骤之后，所述基于知识图谱的信息可视化程序还配置为实现：
    对所述图谱距离进行大小排序，并获取排序结果；以及
    根据所述排序结果从所述关联知识元实体中选取预设数量的目标知识元实体。
  13. 如权利要求12所述的电子设备，其中，所述通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示的步骤，具体包括：以所述初始知识元实体为显示中心，将所述目标知识元实体围绕所述显示中心进行均匀排布，并通过当前界面对排布结果进行显示。
  14. 如权利要求10所述的电子设备，其中，所述通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示的步骤之后，所述基于知识图谱的信息可视化程序还配置为实现：
    根据用户基于所述当前界面输入的跳转知识元实体信息确定跳转知识元实体；
    在预设知识图谱中查找所述跳转知识元实体对应的投影矩阵，并根据所述投影矩阵更新所述预设知识图谱中各知识元实体之间的预设图谱距离，获得新的知识图谱；
    在所述新的知识图谱中查找所述初始知识元实体对应的跳转关联知识元实体，并确定所述跳转关联知识元实体与所述初始知识元实体之间的跳转图谱距离；
    在所述跳转图谱距离属于所述预设目标距离范围时，将所述跳转图谱距离对应的跳转关联知识元实体作为跳转目标知识元实体；以及
    以所述初始知识元实体为中心，对所述初始知识元实体及所述跳转目标知识元实体进行显示。
  15. 一种存储介质，其中，所述存储介质上存储有基于知识图谱的信息可视化程序，所述基于知识图谱的信息可视化程序被处理器执行时实现以下步骤：
    根据用户输入的知识元实体信息确定初始知识元实体；
    在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离；
    确定所述图谱距离属于预设目标距离范围，将所述图谱距离对应的关联知识元实体作为目标知识元实体；以及
    通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示。
  16. 如权利要求15所述的存储介质，其中，所述根据用户输入的知识元实体信息确定初始知识元实体的步骤之前，所述基于知识图谱的信息可视化程序被处理器执行时还实现：
    获取预设主题维度的语料信息；
    将所述语料信息输入至所述预设主题维度对应的TransR模型中，获得对应的知识元实体、所述知识元实体的特征向量以及所述知识元实体之间的预设图谱距离；以及
    根据所述知识元实体、所述知识元实体的特征向量及所述预设图谱距离构建预设知识图谱。
  17. 如权利要求16所述的存储介质，其中，所述基于知识图谱的信息可视化程序被处理器执行时实现的所述在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离的步骤，具体包括：
    根据所述初始知识元实体，在所述预设知识图谱中提取所述初始知识元实体对应的特征向量；
    根据所述特征向量及所述预设知识图谱中包含的各知识元实体之间的预设图谱距离查找所述初始知识元实体对应的关联知识元实体；以及
    根据查找结果确定所述关联知识元实体与所述初始知识元实体的图谱距离。
  18. 如权利要求17所述的存储介质，其中，在预设知识图谱中查找所述初始知识元实体对应的关联知识元实体，并确定所述关联知识元实体与所述初始知识元实体的图谱距离的步骤之后，所述基于知识图谱的信息可视化程序被处理器执行时还实现：
    对所述图谱距离进行大小排序，并获取排序结果；以及
    根据所述排序结果从所述关联知识元实体中选取预设数量的目标知识元实体。
  19. 如权利要求18所述的存储介质，其中所述通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示的步骤，具体包括：以所述初始知识元实体为显示中心，将所述目标知识元实体围绕所述显示中心进行均匀排布，并通过当前界面对排布结果进行显示。
  20. 如权利要求16所述的存储介质，其中，所述通过当前界面对所述目标知识元实体及所述初始知识元实体进行显示的步骤之后，所述基于知识图谱的信息可视化程序被处理器执行时还实现：
    根据用户基于所述当前界面输入的跳转知识元实体信息确定跳转知识元实体；
    在预设知识图谱中查找所述跳转知识元实体对应的投影矩阵，并根据所述投影矩阵更新所述预设知识图谱中各知识元实体之间的预设图谱距离，获得新的知识图谱；
    在所述新的知识图谱中查找所述初始知识元实体对应的跳转关联知识元实体，并确定所述跳转关联知识元实体与所述初始知识元实体之间的跳转图谱距离；
    在所述跳转图谱距离属于所述预设目标距离范围时，将所述跳转图谱距离对应的跳转关联知识元实体作为跳转目标知识元实体；
    以所述初始知识元实体为中心，对所述初始知识元实体及所述跳转目标知识元实体进行显示。
PCT/CN2020/111114 2019-12-30 2020-08-25 基于知识图谱的信息可视化方法、装置、设备及存储介质 WO2021135290A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201911424386.8A CN111159431A (zh) 2019-12-30 2019-12-30 基于知识图谱的信息可视化方法、装置、设备及存储介质
CN201911424386.8 2019-12-30

Publications (1)

Publication Number Publication Date
WO2021135290A1 true WO2021135290A1 (zh) 2021-07-08

Family

ID=70560774

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/111114 WO2021135290A1 (zh) 2019-12-30 2020-08-25 基于知识图谱的信息可视化方法、装置、设备及存储介质

Country Status (2)

Country Link
CN (1) CN111159431A (zh)
WO (1) WO2021135290A1 (zh)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111159431A (zh) * 2019-12-30 2020-05-15 深圳Tcl新技术有限公司 基于知识图谱的信息可视化方法、装置、设备及存储介质
CN111694929B (zh) * 2020-05-29 2023-04-07 平安科技（深圳）有限公司 基于数据图谱的搜索方法、智能终端和可读存储介质
CN111858784A (zh) * 2020-07-21 2020-10-30 广东科杰通信息科技有限公司 一种基于transH的人员亲属关系预测方法
CN112464627B (zh) * 2020-11-05 2021-12-03 北京科技大学 一种面向共指关系的手动文本标注工具及方法
CN112395506A (zh) * 2020-12-04 2021-02-23 上海帜讯信息技术股份有限公司 一种资讯推荐方法、装置、电子设备和存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108460083A (zh) * 2018-01-16 2018-08-28 浙江大学 一种知识图谱可视化查询工具
CN109271525A (zh) * 2018-08-08 2019-01-25 北京百度网讯科技有限公司 用于生成知识图谱的方法、装置、设备以及计算机可读存储介质
CN109345399A (zh) * 2018-10-23 2019-02-15 平安科技（深圳）有限公司 理赔风险评估方法、装置、计算机设备及存储介质
CN110334219A (zh) * 2019-07-12 2019-10-15 电子科技大学 基于注意力机制融入文本语义特征的知识图谱表示学习方法
CN111159431A (zh) * 2019-12-30 2020-05-15 深圳Tcl新技术有限公司 基于知识图谱的信息可视化方法、装置、设备及存储介质

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108694177B (zh) * 2017-04-06 2022-02-18 北大方正集团有限公司 知识图谱构建方法及系统
CN109992673A (zh) * 2019-04-10 2019-07-09 广东工业大学 一种知识图谱生成方法、装置、设备及可读存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108460083A (zh) * 2018-01-16 2018-08-28 浙江大学 一种知识图谱可视化查询工具
CN109271525A (zh) * 2018-08-08 2019-01-25 北京百度网讯科技有限公司 用于生成知识图谱的方法、装置、设备以及计算机可读存储介质
CN109345399A (zh) * 2018-10-23 2019-02-15 平安科技（深圳）有限公司 理赔风险评估方法、装置、计算机设备及存储介质
CN110334219A (zh) * 2019-07-12 2019-10-15 电子科技大学 基于注意力机制融入文本语义特征的知识图谱表示学习方法
CN111159431A (zh) * 2019-12-30 2020-05-15 深圳Tcl新技术有限公司 基于知识图谱的信息可视化方法、装置、设备及存储介质

Also Published As

Publication number Publication date
CN111159431A (zh) 2020-05-15

Similar Documents

Publication Publication Date Title
WO2021135290A1 (zh) 基于知识图谱的信息可视化方法、装置、设备及存储介质
WO2021139325A1 (zh) 媒体信息推荐方法、装置、电子设备及存储介质
JP5612731B2 (ja) 関心領域についての関連情報の決定
WO2018072071A1 (zh) 知识图谱构建系统及方法
US10698956B2 (en) Active knowledge guidance based on deep document analysis
US7769771B2 (en) Searching a document using relevance feedback
KR20190037300A (ko) 키워드의 의미론적 관련성을 측정하기 위한 시스템 및 방법
US20220083874A1 (en) Method and device for training search model, method for searching for target object, and storage medium
JP2013191046A (ja) 構造化文書管理装置、構造化文書検索方法
US20220300543A1 (en) Method of retrieving query, electronic device and medium
CN107315833A (zh) 基于应用程序的检索与下载的方法和装置
JP2007528531A (ja) カテゴリ別のキーワードの入力順位を提供するための検索サービスシステムおよびその方法
CN110008396B (zh) 对象信息推送方法、装置、设备及计算机可读存储介质
US9547701B2 (en) Method of discovering and exploring feature knowledge
JP2020004217A (ja) 情報表示方法、情報表示プログラムおよび情報表示装置
US9251264B2 (en) Systems and methods for enabling an electronic graphical search space of a database
JP6163143B2 (ja) 情報提供装置、情報提供方法、および情報提供プログラム
JP2022500766A (ja) 化学構造の性質を予測するための方法およびシステム
CN113625884A (zh) 一种输入词推荐方法、装置、服务器及存储介质
JP2017072964A (ja) 情報分析装置及び情報分析方法
JP3967230B2 (ja) 画像情報表示システム
JP6176107B2 (ja) 情報収集プログラム、情報収集方法及び情報収集装置
JP7493195B1 (ja) プログラム、方法、情報処理装置、システム
EP4270291A1 (en) Method and computer device to recommend related product technical field
US10909127B2 (en) Method and server for ranking documents on a SERP

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20908902

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20908902

Country of ref document: EP

Kind code of ref document: A1