WO2021036524A1 - 基于人工智能的学术关系知识图谱生成方法和机器人系统 - Google Patents

基于人工智能的学术关系知识图谱生成方法和机器人系统 Download PDF

Info

Publication number
WO2021036524A1
WO2021036524A1 PCT/CN2020/100815 CN2020100815W WO2021036524A1 WO 2021036524 A1 WO2021036524 A1 WO 2021036524A1 CN 2020100815 W CN2020100815 W CN 2020100815W WO 2021036524 A1 WO2021036524 A1 WO 2021036524A1
Authority
WO
WIPO (PCT)
Prior art keywords
relationship
user
academic
experience
users
Prior art date
Application number
PCT/CN2020/100815
Other languages
English (en)
French (fr)
Inventor
朱定局
Original Assignee
南京智慧光信息科技研究院有限公司
大国创新智能科技(东莞)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 南京智慧光信息科技研究院有限公司, 大国创新智能科技(东莞)有限公司 filed Critical 南京智慧光信息科技研究院有限公司
Publication of WO2021036524A1 publication Critical patent/WO2021036524A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/901—Indexing; Data structures therefor; Storage structures
    • G06F16/9024—Graphs; Linked lists
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/903—Querying
    • G06F16/9035—Filtering based on additional data, e.g. user or group profiles

Definitions

  • the invention relates to the field of information technology, in particular to a method and a robot system for generating an academic relationship knowledge graph based on artificial intelligence.
  • the inventor found that there are at least the following problems in the prior art: the current user academic relationship is limited to the academic relationship between users who graduated from the same school or the same unit, and did not analyze the academic relationship through academic achievements , And did not combine time with academic results for analysis.
  • an embodiment of the present invention provides a method for generating an academic relationship knowledge graph, and the method includes:
  • the experience extraction step is to extract the time period of each experience and the academic direction to which the user belongs from the experience published by each user's academic achievements;
  • intersection step find the intersection of every two experiences belonging to different users
  • intersection information acquisition step acquiring every two experiences that have an intersection by matching the experiences, and acquiring the information of the intersection part;
  • the relationship generation step is to determine the relationship between every two users according to the user's experience, and then generate the entity, label, and attribute of the relationship and associate it with the user's academic relationship knowledge base;
  • the knowledge graph generation step is used to treat each user as an entity in the user's academic relationship knowledge graph, and use the relationship label and attributes of each relationship in the user's academic relationship knowledge base as the user's academic relationship knowledge of the two users in the relationship The relationship between the corresponding entities in the map;
  • the entity acquisition step is to acquire an entity that has an association relationship with any entity in the user's academic relationship knowledge graph
  • the relationship acquisition step is to acquire a second user who has a relationship with the first user in the knowledge graph, acquire the relationship between the first user and the second user, and calculate the relationship between the first user and the second user
  • the weight sum is used to sort each user related to the first user from high to low according to the corresponding relationship weight.
  • the experience acquisition step includes:
  • the experience extraction step includes:
  • the work experience extraction step is to extract the time period of each work and the academic direction the user belongs to from the study experience of each user's academic achievements;
  • the patent experience extraction step is to extract the time period of each patent experience and the academic direction of the user in that time period from the patent experience published in academic achievements.
  • the steps for obtaining intersection information include:
  • the first intersection information acquisition step is to obtain every two experiences belonging to different users where the time period intersection is not empty and the academic direction intersection is not empty by matching the experiences, to obtain the time period and academic direction information of the intersection;
  • the second intersection information acquisition step is to obtain every two experiences belonging to different users whose time period intersection is empty and the academic direction intersection is not empty by matching the experiences to acquire the academic direction information of the intersection;
  • the relationship generation steps include:
  • Simultaneous segment relationship generation step judges the simultaneous segment relationship between every two users, generate the entity, label, and attribute of the relationship, and then join the simultaneous segment user academic relationship knowledge base;
  • the different time period relationship generation step is to determine the different time period relationship between each two users according to the user's experience, and the entities, tags, and attributes of the relationship are generated and then added to the user academic relationship knowledge base at different time periods.
  • Simultaneous segment relationship generation steps include:
  • the patent experience simultaneous segment relationship generation step If the two experiences are the patent experiences of two users, and the two users have a simultaneous relationship, then the names of the two users are regarded as the entity of the relationship, and the simultaneous segment
  • the academic peer relationship is used as the label of the relationship, and the time period and academic direction information of the intersection part are used as the attributes of the relationship, and the entities, tags, and attributes of the relationship are associated and added to the simultaneous user academic relationship knowledge base;
  • the simultaneous paragraph relationship generation step of the patent experience if the two experiences are the writing experience of one user and the patent experience of another user, and the two users have a simultaneous relationship, then the names of the two users will be used as the For the entity of the relationship, the simultaneous academic peer relationship is used as the label of the relationship, and the time period and academic direction information of the intersection part are used as the attribute of the relationship, and the entity, label, and attribute of the relationship are associated and added to the simultaneous user academic relationship knowledge base;
  • the steps for generating relationships in different periods include:
  • the patent goes through different time period relationship generation steps. If the two experiences are the patent experiences of two users, and the two users have different time period relationships, then the names of these two users are regarded as the entity of the relationship, and the different time periods
  • the academic peer relationship is used as the label of the relationship, and the time period and academic direction information of the intersection part are used as the attributes of the relationship, and the entities, labels, and attributes of the relationship are associated and added to the user academic relationship knowledge base at different time periods;
  • the process of generating the relationship between the different periods of patent experience if the two experiences are the writing experience of one user and the patent experience of another user, and the two users have different periods of relationship, then the names of the two users will be used as the For the entity of the relationship, the academic peer relationship at different time periods is used as the label of the relationship, and the academic direction information of the intersection part is used as the attribute of the relationship, and the entity, label, and attribute of the relationship are associated and added to the user academic relationship knowledge base at different time periods.
  • the steps of generating knowledge graph include:
  • each user is regarded as an entity in the simultaneous user academic relationship knowledge graph, and the relationship label and attributes of each relationship in the simultaneous user academic relationship knowledge base are used as two of the relationship.
  • the generation step of the user academic relationship knowledge graph at different time periods is to treat each user as an entity in the user academic relationship knowledge graph at different time periods, and the relationship label and attributes of each relationship in the user academic relationship knowledge base at different time periods are used as two of the relationships.
  • each user is regarded as an entity in the user academic relationship knowledge graph, and the relationship label and the relationship label of each relationship in the user academic relationship knowledge base at the same time period and the user academic relationship knowledge base in different periods
  • the attribute serves as the association relationship between the two users in the relationship between the corresponding entities in the user academic relationship knowledge graph;
  • the entity acquisition steps include:
  • Simultaneous entity acquisition step acquiring all entities in the academic relationship knowledge graph of the simultaneous users that have an association relationship with any entity, and obtaining all the academic peers of the users corresponding to the one entity at the same time;
  • the step of acquiring entities in different periods is to obtain all the entities that have an association relationship with any entity in the user's academic relationship knowledge graph at different periods, and obtain all the academic peers of users corresponding to the entity in different periods;
  • all entities in the user's academic relationship knowledge graph that have an association relationship with any entity can be obtained, and all academic peers of the user corresponding to the one entity can be obtained.
  • the relationship acquisition step specifically includes: acquiring all users who have a certain type of relationship with a user in the knowledge graph according to the above steps, acquiring the relationship between the user and each of the users, and taking the length of the time period in the relationship as P1, take the lowest academic direction level in the relationship as P2, calculate f(P1, P2) as the weight of the relationship, calculate the weight sum of all relationships between this user and each user in all users, according to the corresponding
  • the relationship weight and each user who have this type of relationship with the user are sorted from high to low. The higher the ranking, the closer the relationship between the user and the user. The user with the highest ranking and the user of the user are of this type. Closest relationship.
  • an embodiment of the present invention provides a system for generating an academic relationship knowledge graph, and the system includes:
  • the experience acquisition module is used to acquire the experience of each user in the publication of academic achievements
  • the experience extraction module is used to extract the time period of each experience and the academic direction to which the user belongs from the experience published by each user's academic achievements;
  • Experience intersection module used to find the intersection of every two experiences belonging to different users
  • intersection information acquisition module which is used to acquire every two experiences that have an intersection by matching the experiences, and obtain the information of the intersection part
  • the relationship generation module is used for judging the relationship between every two users based on the user's experience, generating the entities, tags, and attributes of the relationship and then adding them to the user academic relationship knowledge base;
  • the knowledge graph generation module is used to treat each user as an entity in the user's academic relationship knowledge graph, and use the relationship label and attributes of each relationship in the user's academic relationship knowledge base as the user's academic relationship knowledge of the two users in the relationship The relationship between the corresponding entities in the map;
  • Entity acquisition module which acquires the entity that has an association relationship with any entity in the user's academic relationship knowledge graph
  • the relationship acquisition module acquires a second user who has a relationship with the first user in the knowledge graph, acquires the relationship between the first user and the second user, and calculates the relationship between the first user and the second user
  • the weight sum is used to sort each user related to the first user from high to low according to the corresponding relationship weight.
  • an embodiment of the present invention provides a computer device, including a memory and a processor, the memory stores a computer program, and the processor implements any one of the embodiments of the first aspect when the computer program is executed. Method steps.
  • an embodiment of the present invention provides a storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method described in any one of the embodiments of the first aspect are implemented.
  • an embodiment of the present invention provides a robotic system for generating an academic relationship knowledge graph.
  • the robotic system includes a memory and a processor, the memory stores a robot program, and the processor implements the first The steps of the method described in any of the aspect embodiments.
  • the method and robot system for generating a knowledge graph of academic relationship based on artificial intelligence include: going through an acquisition step, going through an extraction step, going through an intersection step, intersection information obtaining step, relationship generating step, knowledge graph generating step, entity Get steps.
  • the above methods and systems generate and obtain more accurate user relationships through the generation method of the academic relationship knowledge map based on artificial intelligence, specific to a certain level of academic direction, such as the same journal, the same academic direction, and the same patent classification number, which can be distinguished The relationship between the same period and different time periods. For example, although some users have a relationship with the same school or the same academic direction, they belong to a certain academic direction at different times.
  • the user who has the closest academic relationship with a user can be judged according to the length of the intersection time period corresponding to the relationship and the academic direction level, and all related users of this user can be sorted according to the closeness of the relationship.
  • FIG. 1 is a flowchart of a method for generating an academic relationship knowledge graph provided by Embodiment 1 of the present invention
  • Embodiment 2 is a flowchart of a method for generating an academic relationship knowledge graph provided by Embodiment 2 of the present invention
  • Embodiment 3 is a flowchart of a method for generating an academic relationship knowledge graph provided by Embodiment 3 of the present invention
  • Embodiment 4 is a flowchart of a method for generating an academic relationship knowledge graph provided by Embodiment 4 of the present invention
  • Embodiment 5 is a schematic diagram of the knowledge graph provided by Embodiment 5 of the present invention.
  • Embodiment 6 is a schematic diagram of the knowledge graph provided by Embodiment 6 of the present invention.
  • FIG. 7 is a schematic diagram of the knowledge graph provided by Embodiment 7 of the present invention.
  • Embodiment 8 is a flowchart of a method for generating an academic relationship knowledge graph provided by Embodiment 8 of the present invention.
  • FIG. 9 is a schematic block diagram of a system for generating a user academic relationship knowledge graph provided by Embodiment 10 of the present invention.
  • a method for generating an academic relationship knowledge graph which includes a history acquisition step S100, a history extraction step S200, a history intersection step S300, an intersection information acquisition step S400, and a relationship generation step S500.
  • the experience obtaining step S100 is used to obtain each user's experience in publishing academic achievements.
  • the academic achievement publication includes the user's personal academic achievement publication, and also includes the academic achievement publication obtained from network information such as social networking sites, and also includes all information including the user's experience, and the user's filled-in experience information can also be obtained.
  • the experience extraction step S200 is used to extract the time period of each experience and the academic direction to which the user belongs from the experience published by each user's academic achievements.
  • the intersection information acquisition step S400 is used to acquire every two experiences that have an intersection by matching the experiences, and acquire the information of the intersection.
  • the relationship generation step S500 is used for judging the relationship between every two users according to the user's experience, and generating entities, tags, and attributes of the relationship to associate them with the user academic relationship knowledge base.
  • the knowledge graph generating step S600 is used to treat each user as an entity in the user academic relationship knowledge graph, and use the relationship label and attributes of each relationship in the user academic relationship knowledge base as the user academic relationship of the two users in the relationship The relationship between the corresponding entities in the knowledge graph.
  • an entity that has an association relationship with any entity in the user's academic relationship knowledge graph is acquired.
  • Relationship acquisition step S800 acquiring a second user who has a relationship with the first user in the knowledge graph, acquiring the relationship between the first user and the second user, and calculating the relationship between the first user and the second user According to the weight sum of the corresponding relationship weight, each user who has a relationship with the first user is sorted from high to low.
  • the experience obtaining step S100 includes the writing and patent experience obtaining step S110.
  • the article and patent experience obtaining step S110 is used to obtain the article experience and patent experience in the publication of academic achievements of each user.
  • the history extraction step S200 includes the writing history extraction step S210 and the patent history extraction step S220.
  • Thesis experience extraction step S210 is used to extract the time period of each thesis experience and the academic direction the user belongs to from each user's academic publication experience of the academic achievement.
  • the patent experience extraction step S220 is used to extract the time period of each patent experience and the academic direction of the user in that time period from the patent experience published in academic achievements
  • intersection information obtaining step S400 includes the first intersection information obtaining step S410.
  • the first intersection information acquisition step S410 is used to obtain every two experiences belonging to different users whose time period intersection is not empty and academic direction intersection is not empty by matching experiences, and obtain the time period and academic direction information of the intersection.
  • Extract the name of the first-level academic direction from the academic direction information experienced by user A identify and extract the first-level academic direction according to the journal or conference name of the academic achievement or the field of the patent
  • the name of the second-level academic direction according to the academic achievement Titles or keywords for the recognition and extraction of the second-level academic direction
  • the specific step of finding the time period intersection of two user experiences is to take the common time period of the two time periods as the result of the intersection.
  • the time period of the intersection is: 2015.9-2017.7
  • the relationship generation step S500 includes a simultaneous segment relationship generation step S510.
  • Simultaneous segment relationship generation step S510 is used for judging the simultaneous segment relationship between every two users according to the user's experience, and generate the entities, tags, and attributes of the relationship to associate them with the simultaneous segment user academic relationship knowledge base;
  • Simultaneous writing experience relationship generation step S511 is used for, if the two experiences are the writing experiences of two users, and the two users have a simultaneous relationship, then the names of the two users are used as the entity of the relationship, Use the simultaneous academic peer relationship as the label of the relationship, and use the time period and academic direction information of the intersection as the attribute of the relationship, and associate the entity, label, and attribute of the relationship to the simultaneous user academic relationship knowledge base;
  • Patent experience simultaneous segment relationship generation step S512 if the two experiences are the patent experiences of two users respectively, and the two users have a simultaneous relationship, then the names of the two users are regarded as the entities of the relationship, and the two experiences will be used at the same time.
  • the academic peer relationship of the segment is used as the label of the relationship, and the time period and academic direction information of the intersection part are used as the attributes of the relationship, and the entities, tags, and attributes of the relationship are associated and added to the academic relationship knowledge base of simultaneous users;
  • step S513 if the two experiences are the writing experience of one user and the patent experience of another user, and the two users have a simultaneous relationship, the names of the two users are taken as the simultaneous period relationship generation step S513.
  • the entity of the relationship uses the academic peer relationship of the same period as the label of the relationship, and the time period and academic direction information of the intersection as the attribute of the relationship, and the entity, label, and attribute of the relationship are associated and added to the user academic of the same period Relational knowledge base;
  • intersection information obtaining step S400 includes a second intersection information obtaining step S420.
  • the second intersection information acquisition step S420 is for acquiring every two experiences belonging to different users whose time period intersection is empty and the academic direction intersection is not empty by matching the experiences, to acquire the academic direction information of the intersection part
  • the relationship generation step S500 includes a relationship generation step S520 in different time periods.
  • Different time period relationship generation step S520 is used for judging the different time period relationship between every two users according to the user's experience, and the entities, tags, and attributes of the relationship are generated and associated and then added to the different time user academic relationship knowledge base.
  • step S521 The generation of the relationship between the different periods of experience in step S521 is used for if each of the two experiences is the experience of two users, and the two users have different periods of relationship, then the names of the two users are used as the entity of the relationship , Regard the academic peer relationship at different time periods as the label of the relationship, and use the academic direction information of the intersection as the attribute of the relationship, and associate the entity, label, and attribute of the relationship to the user academic relationship knowledge base at different time periods;
  • the patent experience different time period relationship generation step S522 is used for if the two experiences are the patent experience of two users and the two users have different time period relationships, then the names of the two users are used as the entity of the relationship, Use academic peer relationships at different time periods as the labels of the relationship, and use the time period and academic direction information of the intersection as the attributes of the relationship, and associate the entities, labels, and attributes of the relationship into the user academic relationship knowledge base at different time periods;
  • step S523 if the two experiences are the writing experience of one user and the patent experience of another user, and the two users have a different time relationship, then the names of the two users are taken as the step S523.
  • the entity of the relationship uses the academic peer relationship at different time periods as the label of the relationship, and the academic direction information of the intersection as the attribute of the relationship, and the entity, label, and attribute of the relationship are associated and added to the user academic relationship knowledge base at different time periods ;
  • the two users have a different time relationship, then the names of the two users "Zhang San, Li Si" are used as the entity of the relationship, the academic peer relationship at different time periods is used as the label of the relationship, and the academic direction information of the intersection is A1 Academic direction as the attribute of the relationship, join the user academic relationship knowledge base at different time periods;
  • the knowledge graph generation step S600 includes:
  • Simultaneous user academic relationship knowledge graph generation step S610 Regard each user as an entity in the simultaneous user academic relationship knowledge graph, and use the relationship label and attributes of each relationship in the simultaneous user academic relationship knowledge base as the relationship The association relationship between two users in the corresponding entity in the user academic relationship knowledge graph at the same time.
  • the entity acquisition step S700 includes:
  • Simultaneous entity acquisition step S710 Acquire all entities in the academic relationship knowledge graph of the simultaneous user that have an association relationship with any entity, and obtain all simultaneous academic peers of the user corresponding to the one entity. Obtain all entities with the name of the k-th academic direction in the association relationship with any entity in the academic relationship knowledge graph of the simultaneous users, and obtain all the academic peers of the user corresponding to the same entity with the k-th academic direction in the same period .
  • K is a natural number greater than 1.
  • the knowledge graph generation step S600 includes:
  • Step S620 of generating a knowledge graph of users' academic relations in different periods Regard each user as an entity in the knowledge graph of users' academic relations in different periods, and use the relationship label and attributes of each relationship in the user's academic relationship knowledge base in different periods as the relationship The association relationship between two users in the corresponding entities in the user academic relationship knowledge graph at different time periods,
  • the entity acquisition step S700 includes:
  • Step S720 of obtaining entities in different periods Obtain all entities in the user academic relationship knowledge graph at different periods that have an association relationship with any entity, and obtain all academic peers of users corresponding to the entity in different periods. Obtain all entities with the name of the k-th academic direction in the association relationship with any entity in the user's academic relationship knowledge graph at different time periods, and obtain all the academic peers of the user corresponding to this entity in different time periods with the k-th academic direction .
  • K is a natural number greater than 1.
  • the knowledge graph generation step S600 includes:
  • Step S630 of generating a knowledge graph of user academic relationships in mixed time periods Regarding each user as an entity in the user academic relationship knowledge graph, the relationship label and the relationship label of each relationship in the user academic relationship knowledge base at the same time period and the user academic relationship knowledge base at different time periods are generated. Its attributes serve as the association relationship between the two users in the relationship between the corresponding entities in the user academic relationship knowledge graph.
  • the entity acquisition step S700 includes:
  • Step S730 of acquiring entities in the mixed period acquiring all entities in the user's academic relationship knowledge graph that have an association relationship with any entity, and obtaining all academic peers of the user corresponding to the one entity. Obtain all entities with the name of the k-th academic direction in the association relationship with any entity in the user's academic relationship knowledge graph, and obtain all the academic peers of the user corresponding to the entity with the k-th academic direction.
  • K is a natural number greater than 1.
  • the method for generating the knowledge graph of academic relationship provided in Embodiments 5, 6 and 7 is used in combination.
  • the relationship acquisition step S800 specifically includes: acquiring all users who have a certain type of relationship with a user in the knowledge graph according to the above steps, acquiring the relationship between the user and each of the users, and calculating the length of the time period in the relationship.
  • P1 take the lowest academic direction level in the relationship as P2
  • calculate f(P1, P2) as the weight of the relationship calculate the weight sum of all the relationships between the user and each user in the user, according to the corresponding
  • Each user who has this type of relationship with the user is ranked from high to low based on the relationship weight and the relationship weight. The higher the ranking, the closer the relationship between the user and the user. The highest ranking user has the same relationship with the user.
  • the class has the closest relationship.
  • P1 takes years as the academic direction.
  • the length of the time period is not a whole year, round up to the whole year.
  • P2 is k
  • each user who has a relationship with the user is ranked as Wang Er and Li Si from high to low.
  • each user who has a contemporaneous academic peer relationship with the user is ranked as Wang Er and Li Si,
  • a more accurate relationship between users can be obtained, specific to a certain level of academic direction, such as the same journal, the same academic direction, and the same patent classification number, and the relationship between simultaneous periods and different periods can be distinguished, such as between some users Although they are related to academic directions, they belong to a certain academic direction at different times. It is possible to accurately obtain the specific time period of the relationship between the two users and the specific academic directions at all levels.
  • the user who has the closest academic relationship with a user can be judged according to the length of the intersection time period corresponding to the relationship and the academic direction level, and all related users of this user can be sorted according to the closeness of the relationship
  • an academic relationship knowledge graph generation system which includes a history acquisition module 100, a history extraction module 200, a history intersection module 300, an intersection information acquisition module 400, and a relationship generation module 500.
  • the experience obtaining module 100 is used to obtain the experience of each user in publishing academic achievements.
  • the experience extraction module 200 is used to extract the time period of each experience and the academic direction to which the user belongs from the experience published by each user's academic achievements.
  • Experience intersection module 300 used to find the intersection of every two experiences belonging to different users
  • the intersection information acquisition module 400 is configured to acquire every two experiences that have an intersection by matching the experiences, and acquire the information of the intersection.
  • the relationship generation module 500 is used for judging the relationship between every two users according to the user's experience, and generating entities, tags, and attributes of the relationship to associate them with the user academic relationship knowledge base.
  • the knowledge graph generation module 600 is used to treat each user as an entity in the user academic relationship knowledge graph, and use the relationship label and attributes of each relationship in the user academic relationship knowledge base as the user academic relationship of the two users in the relationship The relationship between the corresponding entities in the knowledge graph;
  • the entity acquisition module 700 acquires an entity that has an association relationship with any entity in the user's academic relationship knowledge graph
  • the relationship acquisition module 800 acquires a second user who has a relationship with the first user in the knowledge graph, acquires the relationship between the first user and the second user, and calculates the relationship between the first user and the second user According to the weight sum of the corresponding relationship weight, each user who has a relationship with the first user is sorted from high to low.
  • the experience acquisition module 100 includes a treatise and patent experience acquisition module 110.
  • the article and patent experience acquisition module 110 is used to obtain the article experience and patent experience in the publication of academic achievements of each user.
  • the history extraction module 200 includes a treatise history extraction module 210 and a patent history extraction module 220.
  • Thesis experience extraction module 210 which is used to extract the time period of each thesis experience and the academic direction to which the user belongs from the experience of each user’s academic publications.
  • the intersection information acquisition module 400 includes a first intersection information acquisition module 410.
  • the first intersection information acquisition module 410 is configured to acquire every two experiences belonging to different users whose time period intersection is not empty and academic direction intersection is not empty by matching experiences, and obtain the time period and academic direction information of the intersection.
  • the relationship generation module 500 includes a simultaneous segment relationship generation module 510.
  • Simultaneous segment relationship generation module 510 used for judging the simultaneous segment relationship between every two users according to the user's experience, and generate the entity, tag, and attribute of the relationship to associate and add the simultaneous segment user academic relationship knowledge base;
  • the writing experience simultaneous period relationship generation module 511 is configured to use the names of the two users as the entity of the relationship if the two experiences are the writing experience of two users, and the two users have a simultaneous relationship. Use the simultaneous academic peer relationship as the label of the relationship, and use the time period and academic direction information of the intersection as the attribute of the relationship, and associate the entity, label, and attribute of the relationship to the simultaneous user academic relationship knowledge base;
  • the patent history simultaneous segment relationship generation module 512 if the two experiences are the patent experiences of two users, and the two users have a simultaneous relationship, then the names of the two users are regarded as the entity of the relationship, and the two experiences will be simultaneously
  • the academic peer relationship of the segment is used as the label of the relationship, and the time period and academic direction information of the intersection part are used as the attributes of the relationship, and the entities, tags, and attributes of the relationship are associated and added to the academic relationship knowledge base of simultaneous users;
  • the intersection information acquisition module 400 includes a second intersection information acquisition module 420.
  • the second intersection information acquisition module 420 is used to obtain every two experiences belonging to different users where the time period intersection is empty and the academic direction intersection is not empty by matching the experiences, to obtain the academic direction information of the intersection
  • the relationship generation module 500 includes a relationship generation module 520 in different time periods.
  • the different time period relationship generation module 520 is used to determine the different time period relationship between each two users according to the user's experience, generate the entity, tag, and attribute of the relationship, and then join the different time user academic relationship knowledge base.
  • the different time period relationship generation module 521 for writing experience is used for if each of the two experiences is the writing experience of two users, and the two users have a different time period relationship, then the names of the two users are used as the entity of the relationship , Regard the academic peer relationship at different time periods as the label of the relationship, and use the academic direction information of the intersection as the attribute of the relationship, and associate the entity, label, and attribute of the relationship to the user academic relationship knowledge base at different time periods;
  • the patent experience different time period relationship generation module 522 is configured to use the names of the two users as the entity of the relationship if the two experiences are the patent experience of two users and the two users have different time periods. Use academic peer relationships at different time periods as the labels of the relationship, and use the time period and academic direction information of the intersection as the attributes of the relationship, and associate the entities, labels, and attributes of the relationship into the user academic relationship knowledge base at different time periods;
  • the different time period relationship generation module 523 on the patent experience if the two experiences are the writing experience of one user and the patent experience of the other user, and the two users have different time periods, then the names of the two users are taken as The entity of the relationship uses the academic peer relationship at different time periods as the label of the relationship, and the academic direction information of the intersection as the attribute of the relationship, and the entity, label, and attribute of the relationship are associated and added to the user academic relationship knowledge base at different time periods .
  • the knowledge graph generation module 600 includes:
  • Simultaneous user academic relationship knowledge graph generation module 610 Regard each user as an entity in the simultaneous user academic relationship knowledge graph, and use the relationship label and attributes of each relationship in the simultaneous user academic relationship knowledge base as the relationship The association relationship between two users in the corresponding entity in the user academic relationship knowledge graph at the same time.
  • the entity acquisition module 700 includes:
  • Simultaneous entity acquisition module 710 acquires all entities in the academic relationship knowledge graph of a simultaneous user that have an association relationship with any entity, and can obtain all simultaneous academic peers of the user corresponding to the one entity. Obtain all entities with the name of the k-th academic direction in the association relationship with any entity in the academic relationship knowledge graph of the simultaneous users, and obtain all the academic peers of the user corresponding to the same entity with the k-th academic direction in the same period .
  • K is a natural number greater than 1.
  • the knowledge graph generation module 600 includes:
  • the user academic relationship knowledge graph generation module 620 at different time periods each user is regarded as an entity in the user academic relationship knowledge graph at different time periods, and the relationship label and attributes of each relationship in the user academic relationship knowledge database at different time periods are used as the relationship The association relationship between two users in the corresponding entities in the user academic relationship knowledge graph at different time periods,
  • the entity acquisition module 700 includes:
  • the different period entity acquisition module 720 Obtain all entities in the user academic relationship knowledge graph of different periods that have an association relationship with any entity, and obtain all academic peers of the user corresponding to the entity in different periods. Obtain all entities with the name of the k-th academic direction in the association relationship with any entity in the user's academic relationship knowledge graph at different time periods, and obtain all the academic peers of the user corresponding to this entity in different time periods with the k-th academic direction . K is a natural number greater than 1.
  • the knowledge graph generation module 600 includes:
  • the mixed time user academic relationship knowledge graph generation module 630 Regard each user as an entity in the user academic relationship knowledge graph, the relationship label and the relationship label of each relationship in the user academic relationship knowledge base at the same time period and the user academic relationship knowledge base at different periods Its attributes serve as the association relationship between the two users in the relationship between the corresponding entities in the user academic relationship knowledge graph.
  • the entity acquisition module 700 includes:
  • Mixed-period entity acquisition module 730 acquire all entities in the user's academic relationship knowledge graph that have an associated relationship with any entity, and obtain all academic peers of the user corresponding to the one entity. Obtain all entities with the name of the k-th academic direction in the association relationship with any entity in the user's academic relationship knowledge graph, and obtain all the academic peers of the user corresponding to the entity with the k-th academic direction. K is a natural number greater than 1.
  • the user academic relationship knowledge graph generation system provided according to Embodiments 14, 15, and 16 can be used in combination.
  • the relationship acquiring module 800 specifically includes: acquiring all users who have a certain type of relationship with a user in the knowledge graph according to the above steps, acquiring the relationship between the user and each of the users, and calculating the length of the time period in the relationship.
  • P1 take the lowest academic direction level in the relationship as P2
  • calculate f(P1, P2) as the weight of the relationship calculate the weight sum of all the relationships between the user and each user in the user, according to the corresponding
  • Each user who has this type of relationship with the user is ranked from high to low based on the relationship weight and the relationship weight. The higher the ranking, the closer the relationship between the user and the user. The highest ranking user has the same relationship with the user.
  • the class has the closest relationship.
  • the methods and systems in the foregoing embodiments can be executed and deployed on computers, servers, cloud servers, supercomputers, robots, embedded devices, electronic devices, and so on.

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • Software Systems (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

基于人工智能的学术关系知识图谱生成方法和机器人系统，包括：经历获取步骤，经历抽取步骤，经历求交步骤，交集信息获取步骤，关系生成步骤，知识图谱生成步骤，实体获取步骤。上述方法和系统通过基于人工智能的用户学术关系知识图谱生成技术，提高了用户学术关系知识图谱生成的智能性、高效性。

Description

基于人工智能的学术关系知识图谱生成方法和机器人系统 技术领域
本发明涉及信息技术领域，特别是涉及一种基于人工智能的学术关系知识图谱生成方法和机器人系统。
背景技术
在实现本发明过程中，发明人发现现有技术中至少存在如下问题：现有用户学术关系只限于同一个学校毕业或同一个单位之间的用户学术关系，而没有通过学术成果来分析学术关系，更没有将时间与学术成果结合起来分析。
因此，现有技术还有待于改进和发展。
发明内容
基于此，有必要针对现有技术中的缺陷或不足，提供基于人工智能的学术关系知识图谱生成方法和机器人系统，以解决现有技术中用户学术关系生成时没有将时间、学术成果、学术方向、学术关系综合考虑的缺点。
第一方面，本发明实施例提供一种学术关系知识图谱生成方法，所述方法包括：
经历获取步骤，获取每个用户的学术成果发表中的经历；
经历抽取步骤，从每个用户的学术成果发表的经历中抽取每个经历的时间段及该时间段用户所属的学术方向；
经历求交步骤，求属于不同用户的每两个经历的交集；
交集信息获取步骤，通过对经历进行匹配获取具有交集的每两个经历，获取交集部分的信息；
关系生成步骤，根据用户的经历，判断每两个用户之间的关系，生成该关系的实体、标签、属性进行关联后加入用户学术关系知识库；
知识图谱生成步骤，用于将每个用户作为用户学术关系知识图谱中的一个实体，将用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户的用户学术关系知识图谱中对应的实体之间的关联关系；
实体获取步骤，获取用户学术关系知识图谱中与任一实体具有关联关系的实体；
关系获取步骤，获取与第一用户在知识图谱中具备关系的第二用户，获取该第一用户与该第二用户之间的关系，计算该第一用户与该第二用户之间的关系的权重和，根据对应的关系权重和从高到低对与该第一用户有关系的每一用户进行排序。
优选地，
所述经历获取步骤包括：
论著和专利经历获取步骤，获取每个用户的学术成果发表中的论著经历和专利经历；
所述经历抽取步骤包括：
论著经历抽取步骤，从每个用户的学术成果发表的论著经历中抽取每个论著经历的时间 段及该时间段用户所属的学术方向；
专利经历抽取步骤，从学术成果发表的专利经历中抽取每个专利经历的时间段及该时间段用户所属的学术方向。
优选地，
交集信息获取步骤包括：
第一交集信息获取步骤，通过对经历进行匹配获取时间段交集不为空且学术方向交集不为空的属于不同用户的每两个经历，获取交集部分的时间段和学术方向信息；
第二交集信息获取步骤，通过对经历进行匹配获取时间段交集为空且学术方向交集不为空的属于不同用户的每两个经历，获取交集部分的学术方向信息；
关系生成步骤包括：
同时段关系生成步骤，根据用户的经历，判断每两个用户之间的同时段关系，生成该关系的实体、标签、属性进行关联后加入同时段用户学术关系知识库；
不同时段关系生成步骤，根据用户的经历，判断每两个用户之间的不同时段关系，生成该关系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库。
优选地，
同时段关系生成步骤包括：
论著经历同时段关系生成步骤，如果所述两个经历分别为两个用户的论著经历，且这两个用户具有同时段关系，则将这两个用户的姓名作为该关系的实体，将同时段学术同行关系作为该关系的标签，将交集部分的时间段和学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入同时段用户学术关系知识库；
专利经历同时段关系生成步骤，如果所述两个经历分别为两个用户的专利经历，且这两个用户具有同时段关系，则将这两个用户的姓名作为该关系的实体，将同时段学术同行关系作为该关系的标签，将交集部分的时间段和学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入同时段用户学术关系知识库；
论著专利经历同时段关系生成步骤，如果所述两个经历分别为一个用户的论著经历和另一个用户的专利经历，且这两个用户具有同时段关系，则将这两个用户的姓名作为该关系的实体，将同时段学术同行关系作为该关系的标签，将交集部分的时间段和学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入同时段用户学术关系知识库；
不同时段关系生成步骤包括：
论著经历不同时段关系生成步骤，如果所述每两个经历分别为两个用户的论著经历，且这两个用户具有不同时段关系，则将这两个用户的姓名作为该关系的实体，将不同时段学术同行关系作为该关系的标签，将交集部分的学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库；
专利经历不同时段关系生成步骤，如果所述两个经历分别为两个用户的专利经历，且这两个用户具有不同时段关系，则将这两个用户的姓名作为该关系的实体，将不同时段学术同行关系作为该关系的标签，将交集部分的时间段和学术方向信息作为该关系的属性，将该关 系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库；
论著专利经历不同时段关系生成步骤，如果所述两个经历分别为一个用户的论著经历和另一个用户的专利经历，且这两个用户具有不同时段关系，则将这两个用户的姓名作为该关系的实体，将不同时段学术同行关系作为该关系的标签，将交集部分的学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库。
优选地，
知识图谱生成步骤包括：
同时段用户学术关系知识图谱生成步骤，将每个用户作为同时段用户学术关系知识图谱中的一个实体，将同时段用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户在同时段用户学术关系知识图谱中对应的实体之间的关联关系；
不同时段用户学术关系知识图谱生成步骤，将每个用户作为不同时段用户学术关系知识图谱中的一个实体，将不同时段用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户在不同时段用户学术关系知识图谱中对应的实体之间的关联关系；
混合时段用户学术关系知识图谱生成步骤，将每个用户作为用户学术关系知识图谱中的一个实体，将同时段用户学术关系知识库及不同时段用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户在用户学术关系知识图谱中对应的实体之间的关联关系；
实体获取步骤包括：
同时段实体获取步骤，获取同时段用户学术关系知识图谱中与任一实体具有关联关系的所有实体，可以得到该一实体对应的用户的所有同时段的学术同行；
不同时段实体获取步骤，获取不同时段用户学术关系知识图谱中与任一实体具有关联关系的所有实体，可以得到该一实体对应的用户的所有不同时段的学术同行；
混合时段实体获取步骤，获取用户学术关系知识图谱中与任一实体具有关联关系的所有实体，可以得到该一实体对应的用户的所有学术同行。
优选地，
关系获取步骤具体包括：根据上述步骤获取与一个用户在知识图谱中具备某类关系的所有用户，获取该个用户与该所有用户中每一用户之间的关系，将关系中时间段的长度作为P1，将关系中的最低学术方向级别作为P2，计算f(P1，P2)作为该关系的权重，计算该个用户与该所有用户中每一用户之间的所有关系的权重和，根据对应的关系权重和从高到低对与该用户有该类关系的每一用户进行排序，排序越靠前的用户与该用户的该类关系越密切，排序最靠前的用户与该用户的该类关系最密切。
第二方面，本发明实施例提供一种学术关系知识图谱生成系统，所述系统包括：
经历获取模块，用于获取每个用户的学术成果发表中的经历；
经历抽取模块，用于从每个用户的学术成果发表的经历中抽取每个经历的时间段及该时间段用户所属的学术方向；
经历求交模块，用于求属于不同用户的每两个经历的交集；
交集信息获取模块，用于通过对经历进行匹配获取具有交集的每两个经历，获取交集部分的信息；
关系生成模块，用于根据用户的经历，判断每两个用户之间的关系，生成该关系的实体、标签、属性进行关联后加入用户学术关系知识库；
知识图谱生成模块，用于将每个用户作为用户学术关系知识图谱中的一个实体，将用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户的用户学术关系知识图谱中对应的实体之间的关联关系；
实体获取模块，获取用户学术关系知识图谱中与任一实体具有关联关系的实体；
关系获取模块，获取与第一用户在知识图谱中具备关系的第二用户，获取该第一用户与该第二用户之间的关系，计算该第一用户与该第二用户之间的关系的权重和，根据对应的关系权重和从高到低对与该第一用户有关系的每一用户进行排序。
第三方面，本发明实施例提供一种计算机设备，包括存储器和处理器，所述存储器存储有计算机程序，所述处理器执行所述计算机程序时实现第一方面实施例中任一项所述方法的步骤。
第四方面，本发明实施例提供一种存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时实现第一方面实施例中任一项所述方法的步骤。
第五方面，本发明实施例提供一种学术关系知识图谱生成机器人系统，所述机器人系统包括存储器和处理器，所述存储器存储有机器人程序，所述处理器执行所述机器人程序时实现第一方面实施例中任一项所述方法的步骤。
本发明实施例提供的基于人工智能的学术关系知识图谱生成方法和机器人系统，包括：经历获取步骤，经历抽取步骤，经历求交步骤，交集信息获取步骤，关系生成步骤，知识图谱生成步骤，实体获取步骤。上述方法和系统通过基于人工智能的学术关系知识图谱生成方法生成和获得更为精准的用户之间关系，具体到某一级学术方向，例如同一期刊、同一学术方向、同一专利分类号，可以区分同时段和不同时段的关系，例如有的用户之间虽然有同校或同学术方向的关系，但是在不同时间属于某个学术方向的。可以精确地生成和得到两个用户之间的关系所在的具体时间段具体各级学术方向。可以根据关系对应的交集中时间段的长短和学术方向级别来判断与一个用户学术关系最密切的用户，并能根据关系的密切程度对这个用户的所有有关系用户进行排序。
附图说明
图1为本发明的实施例1提供的学术关系知识图谱生成方法的流程图；
图2为本发明的实施例2提供的学术关系知识图谱生成方法的流程图；
图3为本发明的实施例3提供的学术关系知识图谱生成方法的流程图；
图4为本发明的实施例4提供的学术关系知识图谱生成方法的流程图；
图5为本发明的实施例5提供的知识图谱的示意图；
图6为本发明的实施例6提供的知识图谱的示意图；
图7为本发明的实施例7提供的知识图谱的示意图；
图8为本发明的实施例8提供的学术关系知识图谱生成方法的流程图；
图9为本发明的实施例10提供的用户学术关系知识图谱生成系统的原理框图。
具体实施方式
下面结合本发明实施方式，对本发明实施例中的技术方案进行详细地描述。
(一)本发明的各种实施例中的方法包括以下步骤的各种组合：
实施例1：
如图1所示，提供一种学术关系知识图谱生成方法，包括经历获取步骤S100、经历抽取步骤S200、经历求交步骤S300、交集信息获取步骤S400、关系生成步骤S500。
经历获取步骤S100，用于获取每个用户的学术成果发表中的经历。所述学术成果发表包括用户的个人学术成果发表，也包括从社交网站等网络信息中获取的学术成果发表，还包括所有包含有用户的经历的信息，也可以获取用户填写的经历信息。
经历抽取步骤S200，用于从每个用户的学术成果发表的经历中抽取每个经历的时间段及该时间段用户所属的学术方向。
经历求交步骤S300，用于求属于不同用户的每两个经历的交集
例如：
张三的论著经历2010.9-2014.7 A1学术方向B11期刊
与
李四的专利经历2018.9-2019.7 A1学术方向B12期刊
的交集为
A1学术方向
例如，
张三的论著经历2014.9-2017.7 A21学术方向B21期刊
与
李四的论著经历2015.9-2018.7 A21学术方向B21期刊
的交集为
2015.9-2017.7 A21学术方向B21期刊
交集信息获取步骤S400，用于通过对经历进行匹配获取具有交集的每两个经历，获取交集部分的信息。
关系生成步骤S500，用于根据用户的经历，判断每两个用户之间的关系，生成该关系的实体、标签、属性进行关联后加入用户学术关系知识库。
知识图谱生成步骤S600，用于将每个用户作为用户学术关系知识图谱中的一个实体，将用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户的用户学术关系知识图谱中对应的实体之间的关联关系。
实体获取步骤S700，获取用户学术关系知识图谱中与任一实体具有关联关系的实体。
关系获取步骤S800，获取与第一用户在知识图谱中具备关系的第二用户，获取该第一用 户与该第二用户之间的关系，计算该第一用户与该第二用户之间的关系的权重和，根据对应的关系权重和从高到低对与该第一用户有关系的每一用户进行排序。
实施例2：
如图2所示，根据实施例1提供的学术关系知识图谱生成方法，
其中，经历获取步骤S100包括论著和专利经历获取步骤S110。
论著和专利经历获取步骤S110，用于获取每个用户的学术成果发表中的论著经历和专利经历。
学术成果发表可以通过用户输入，也可以从百度百科或其它网站上获取，获取从中提取论著经历和专利经历。
例如
张三
论著经历
2010.9-2014.7 A1学术方向B11期刊
2014.9-2017.7 A2学术方向B21期刊
专利经历
2017.9-2018.7 A3学术方向B31专利分类号
2018.9-2019.7 A4学术方向B41专利分类号
李四
论著经历
2011.9-2015.7 A5学术方向B51期刊
2015.9-2018.7 A2学术方向B21期刊
专利经历
2018.9-2019.7 A1学术方向B12期刊
2019.9-2021.7 A3学术方向B31专利分类号
其中，经历抽取步骤S200包括论著经历抽取步骤S210、专利经历抽取步骤S220。
论著经历抽取步骤S210，用于从每个用户的学术成果发表的论著经历中抽取每个论著经历的时间段及该时间段用户所属的学术方向
例如
从张三的论著经历中抽取
2010.9-2014.7 A1学术方向B11期刊
2014.9-2017.7 A2学术方向B21期刊
从李四的论著经历中抽取
2011.9-2015.7 A5学术方向B51期刊
2015.9-2018.7 A2学术方向B21期刊
专利经历抽取步骤S220，用于从学术成果发表的专利经历中抽取每个专利经历的时间段及该时间段用户所属的学术方向
例如
从张三的专利经历中抽取
2017.9-2018.7 A3学术方向B31专利分类号
2018.9-2019.7 A4学术方向B41专利分类号
从李四的专利经历中抽取
2018.9-2019.7 A1学术方向B12期刊
2019.9-2021.7 A3学术方向B31专利分类号
实施例3：
如图3所示，根据实施例1提供的学术关系知识图谱生成方法，
其中，交集信息获取步骤S400包括第一交集信息获取步骤S410。
第一交集信息获取步骤S410，用于通过对经历进行匹配获取时间段交集不为空且学术方向交集不为空的属于不同用户的每两个经历，获取交集部分的时间段和学术方向信息。
求两个用户经历(用户A的一个经历、用户B的一个经历)的学术方向交集的具体步骤
从用户A经历的学术方向信息中分别提取一级学术方向名称(根据学术成果所在期刊或会议名称或专利所在领域进行一级学术方向的识别和提取)、二级学术方向名称(根据学术成果的标题或关键词进行二级学术方向的识别和提取)等多级学术方向名称
同样，从用户B经历的学术方向信息中分别提取一级学术方向名称、二级学术方向名称等多级学术方向名称
如果用户A与用户B的一级学术方向名称相同但二级学术方向名称不同，则该两个用户的学术方向交集为该一级学术方向名称
如果用户A与用户B的一级学术方向名称相同且二级学术方向名称相同但三级学术方向名称不同，则该两个用户的学术方向交集为该一级学术方向名称二级学术方向名称
如果用户A与用户B的一级学术方向名称相同且二级学术方向名称相同且三级学术方向名称相同但四级学术方向名称不同，则该两个用户的学术方向交集为该一级学术方向名称二级学术方向名称三级学术方向名称
如此类推。
求两个用户经历(用户A的一个经历、用户B的一个经历)的时间段交集的具体步骤为取2个时间段的共同时间段作为交集的结果。
例如
交集 2015.9-2017.7 A21学术方向B21期刊 对应的属于张三、李四的两个经历：
张三的论著经历2014.9-2017.7 A21学术方向B21期刊
李四的论著经历2015.9-2018.7 A21学术方向B21期刊
交集部分的时间段为：2015.9-2017.7
交集部分的学术方向信息为：A21学术方向B21期刊
其中，关系生成步骤S500包括同时段关系生成步骤S510。
同时段关系生成步骤S510，用于根据用户的经历，判断每两个用户之间的同时段关系， 生成该关系的实体、标签、属性进行关联后加入同时段用户学术关系知识库；
论著经历同时段关系生成步骤S511，用于如果所述两个经历分别为两个用户的论著经历，且这两个用户具有同时段关系，则将这两个用户的姓名作为该关系的实体，将同时段学术同行关系作为该关系的标签，将交集部分的时间段和学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入同时段用户学术关系知识库；
例如
交集 2015.9-2017.7 A21学术方向B21期刊 对应的属于张三、李四的两个经历：
张三的论著经历2014.9-2017.7 A21学术方向B21期刊
李四的论著经历2015.9-2018.7 A21学术方向B21期刊
分别为两个用户的论著经历
且这两个用户具有同时段关系，则将这两个用户的姓名“张三、李四”作为该关系的实体，将同时段学术同行关系作为该关系的标签，将交集部分的时间段2015.9-2017.7和学术方向信息A21学术方向B21期刊作为该关系的属性，加入同时段用户学术关系知识库；
专利经历同时段关系生成步骤S512，如果所述两个经历分别为两个用户的专利经历，且这两个用户具有同时段关系，则将这两个用户的姓名作为该关系的实体，将同时段学术同行关系作为该关系的标签，将交集部分的时间段和学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入同时段用户学术关系知识库；
论著专利经历同时段关系生成步骤S513，如果所述两个经历分别为一个用户的论著经历和另一个用户的专利经历，且这两个用户具有同时段关系，则将这两个用户的姓名作为该关系的实体，将同时段学术同行关系作为该关系的标签，将交集部分的时间段和学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入同时段用户学术关系知识库；
实施例4：
如图4所示，根据实施例1提供的学术关系知识图谱生成方法，
其中，交集信息获取步骤S400包括第二交集信息获取步骤S420。
第二交集信息获取步骤S420，用于通过对经历进行匹配获取时间段交集为空且学术方向交集不为空的属于不同用户的每两个经历，获取交集部分的学术方向信息
例如
交集 A1学术方向 对应的属于张三、李四的两个经历：
张三的论著经历2010.9-2014.7 A1学术方向B11期刊
李四的专利经历2018.9-2019.7 A1学术方向B12专利分类号
交集部分的时间段为空
交集部分的学术方向信息为：A1学术方向
其中，关系生成步骤S500包括不同时段关系生成步骤S520。
不同时段关系生成步骤S520，用于根据用户的经历，判断每两个用户之间的不同时段关系，生成该关系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库。
论著经历不同时段关系生成步骤S521，用于如果所述每两个经历分别为两个用户的论著经历，且这两个用户具有不同时段关系，则将这两个用户的姓名作为该关系的实体，将不同时段学术同行关系作为该关系的标签，将交集部分的学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库；
专利经历不同时段关系生成步骤S522，用于如果所述两个经历分别为两个用户的专利经历，且这两个用户具有不同时段关系，则将这两个用户的姓名作为该关系的实体，将不同时段学术同行关系作为该关系的标签，将交集部分的时间段和学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库；
论著专利经历不同时段关系生成步骤S523，如果所述两个经历分别为一个用户的论著经历和另一个用户的专利经历，且这两个用户具有不同时段关系，则将这两个用户的姓名作为该关系的实体，将不同时段学术同行关系作为该关系的标签，将交集部分的学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库；
例如
交集 A1学术方向 对应的属于张三、李四的两个经历：
张三的论著经历2010.9-2014.7 A1学术方向B11期刊
李四的专利经历2018.9-2019.7 A1学术方向B12期刊
分别为一个用户的论著经历和另一个用户的专利经历
且这两个用户具有不同时段关系，则将这两个用户的姓名“张三、李四”作为该关系的实体，将不同时段学术同行关系作为该关系的标签，将交集部分的学术方向信息A1学术方向作为该关系的属性，加入不同时段用户学术关系知识库；
实施例5：
如图5所示，根据实施例1提供的学术关系知识图谱生成方法，
知识图谱生成步骤S600包括：
同时段用户学术关系知识图谱生成步骤S610：将每个用户作为同时段用户学术关系知识图谱中的一个实体，将同时段用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户在同时段用户学术关系知识图谱中对应的实体之间的关联关系。
实体获取步骤S700包括：
同时段实体获取步骤S710：获取同时段用户学术关系知识图谱中与任一实体具有关联关系的所有实体，可以得到该一实体对应的用户的所有同时段的学术同行。获取同时段用户学术关系知识图谱中与任一实体具有的关联关系中具有第k级学术方向名称的所有实体，可以得到该一实体对应的用户的所有同时段同第k级学术方向的学术同行。K为大于1的自然数。
实施例6：
如图6所示，根据实施例1提供的学术关系知识图谱生成方法，
知识图谱生成步骤S600包括：
不同时段用户学术关系知识图谱生成步骤S620：将每个用户作为不同时段用户学术关系 知识图谱中的一个实体，将不同时段用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户在不同时段用户学术关系知识图谱中对应的实体之间的关联关系，
实体获取步骤S700包括：
不同时段实体获取步骤S720：获取不同时段用户学术关系知识图谱中与任一实体具有关联关系的所有实体，可以得到该一实体对应的用户的所有不同时段的学术同行。获取不同时段用户学术关系知识图谱中与任一实体具有的关联关系中具有第k级学术方向名称的所有实体，可以得到该一实体对应的用户的所有不同时段同第k级学术方向的学术同行。K为大于1的自然数。
进一步，获取不同时段用户学术关系知识图谱中与任一实体具有的关联关系中具有第k级学术方向名称且含有“学术同行”关键词的所有实体，可以得到该一实体对应的用户的所有不同时段同第k级学术方向的学术同行。
实施例7：
如图7所示，根据实施例1提供的学术关系知识图谱生成方法，
知识图谱生成步骤S600包括：
混合时段用户学术关系知识图谱生成步骤S630：将每个用户作为用户学术关系知识图谱中的一个实体，将同时段用户学术关系知识库及不同时段用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户在用户学术关系知识图谱中对应的实体之间的关联关系。
实体获取步骤S700包括：
混合时段实体获取步骤S730：获取用户学术关系知识图谱中与任一实体具有关联关系的所有实体，可以得到该一实体对应的用户的所有学术同行。获取用户学术关系知识图谱中与任一实体具有的关联关系中具有第k级学术方向名称的所有实体，可以得到该一实体对应的用户的所有同第k级学术方向的学术同行。K为大于1的自然数。
例如，
获取用户学术关系知识图谱中与张三具有关联关系的所有实体，可以得到张三的学术同行有李四、王二，学术同行有李四、王二，学术同行有李四。
例如，
k为2时，获取用户学术关系知识图谱中与张三具有的关联关系中具有第2级学术方向名称(对应期刊、专利分类号)的所有实体，可以得到张三的所有同第2级学术方向的学术同行有李四、王二
实施例8：
如图8所示，根据实施例5、6、7提供的学术关系知识图谱生成方法进行组合使用。
实施例9：
根据实施例1提供的学术关系知识图谱生成方法，
关系获取步骤S800具体包括：根据上述步骤获取与一个用户在知识图谱中具备某类关系的所有用户，获取该个用户与该所有用户中每一用户之间的关系，将关系中时间段的长度作 为P1，将关系中的最低学术方向级别作为P2，计算f(P1，P2)作为该关系的权重，计算该个用户与该所有用户中每一用户之间的所有关系的权重和，根据对应的关系权重和从高到低对与该用户有该类关系的每一用户进行排序，排序越靠前的用户与该用户的该类关系越密切，排序最靠前的用户与该用户的该类关系最密切。
其中，P1以年为学术方向，时间段的长度不是整年时，向上取整为整年，最低学术方向为k级学术方向时P2为k，f(P1,P2)可以有多种实现方式，例如f(P1,P2)＝P1*P2+P2
例如
根据上述步骤获取与张三在知识图谱中具备关系的所有用户李四、王二，获取张三与李四之间的关系、张三与王二之间的关系，
将张三与李四之间的关系“不同时段学术同行关系、A3学术方向B31专利分类号”中时间段的长度0作为P1，将关系中的最低学术方向级别2作为P2，计算f(P1，P2)＝P1*P2+P2＝2作为该关系的权重。
将张三与李四之间的关系“不同时段学术同行关系、A1学术方向”中时间段的长度0作为P1，将关系中的最低学术方向级别1作为P2，计算f(P1，P2)＝P1*P2+P2＝1作为该关系的权重。
将张三与李四之间的关系“同时段学术同行关系、2015.9-2017.7 A21学术方向B21期刊”中时间段的长度2作为P1，将关系中的最低学术方向级别2作为P2，计算f(P1，P2)＝P1*P2+P2＝6作为该关系的权重。
计算张三与李四之间的所有关系的权重和为2+1+6＝9
将张三与王二之间的关系“同时段学术同行关系、2017.9-2018.7 A3学术方向”中时间段的长度1作为P1，将关系中的最低学术方向级别1作为P2，计算f(P1，P2)＝max(P1*P2+P2)＝2作为该关系的权重，
将张三与王二之间的关系“同时段学术同行关系、2014.9-2017.7 A2学术方向B21期刊”中时间段的长度3作为P1，将关系中的最低学术方向级别2作为P2，计算f(P1，P2)＝max(P1*P2+P2)＝8作为该关系的权重，
将张三与王二之间的关系“同时段学术同行关系、2010.9-2014.7 A1学术方向B11期刊”中时间段的长度4作为P1，将关系中的最低学术方向级别2作为P2，计算f(P1，P2)＝max(P1*P2+P2)＝10作为该关系的权重，
计算张三与王二之间的所有关系的权重和为2+8+10＝20
根据关系的权重从高到低对与该用户有关系的所述每一用户进行排序为王二、李四，
王二与张三的关系更密切，排序最靠前的王二与张三的关系最密切。
例如
根据上述步骤获取与张三在知识图谱中具备同时段学术同行关系的所有用户李四、王二，获取张三与李四之间的关系、张三与王二之间的关系，
将张三与李四之间的关系“同时段学术同行关系、2015.9-2017.7 A21学术方向B21期刊”中时间段的长度2作为P1，将关系中的最低学术方向级别2作为P2，计算f(P1，P2) ＝P1*P2+P2＝6作为该关系的权重。
计算张三与李四之间的所有关系的权重和为6＝6
将张三与王二之间的关系“同时段学术同行关系、2014.9-2017.7 A2学术方向B21期刊”中时间段的长度3作为P1，将关系中的最低学术方向级别2作为P2，计算f(P1，P2)＝max(P1*P2+P2)＝8作为该关系的权重，
将张三与王二之间的关系“同时段学术同行关系、2010.9-2014.7 A1学术方向B11期刊”中时间段的长度4作为P1，将关系中的最低学术方向级别2作为P2，计算f(P1，P2)＝max(P1*P2+P2)＝10作为该关系的权重，
计算张三与王二之间的所有关系的权重和为8+10＝18
根据关系的权重从高到低对与该用户有同时段学术同行关系的所述每一用户进行排序为王二、李四，
王二与张三的同时段学术同行关系更密切，排序最靠前的王二与张三的同时段学术同行关系最密切。
效果：可以获得更为精准的用户之间关系，具体到某一级学术方向，例如同一期刊、同一学术方向、同一专利分类号，可以区分同时段和不同时段的关系，例如有的用户之间虽然有同学术方向的关系，但是在不同时间属于某个学术方向的。可以精确地得到两个用户之间的关系所在的具体时间段具体各级学术方向。可以根据关系对应的交集中时间段的长短和学术方向级别来判断与一个用户学术关系最密切的用户，并能根据关系的密切程度对这个用户的所有有关系用户进行排序
实施例10：
如图9所示，提供一种学术关系知识图谱生成系统，包括经历获取模块100、经历抽取模块200、经历求交模块300、交集信息获取模块400、关系生成模块500。
经历获取模块100，用于获取每个用户的学术成果发表中的经历。
经历抽取模块200，用于从每个用户的学术成果发表的经历中抽取每个经历的时间段及该时间段用户所属的学术方向。
经历求交模块300，用于求属于不同用户的每两个经历的交集
交集信息获取模块400，用于通过对经历进行匹配获取具有交集的每两个经历，获取交集部分的信息。
关系生成模块500，用于根据用户的经历，判断每两个用户之间的关系，生成该关系的实体、标签、属性进行关联后加入用户学术关系知识库。
知识图谱生成模块600，用于将每个用户作为用户学术关系知识图谱中的一个实体，将用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户的用户学术关系知识图谱中对应的实体之间的关联关系；
实体获取模块700，获取用户学术关系知识图谱中与任一实体具有关联关系的实体；
关系获取模块800，获取与第一用户在知识图谱中具备关系的第二用户，获取该第一用户与该第二用户之间的关系，计算该第一用户与该第二用户之间的关系的权重和，根据对应 的关系权重和从高到低对与该第一用户有关系的每一用户进行排序。
实施例11：
根据实施例10提供的学术关系知识图谱生成系统，
其中，经历获取模块100包括论著和专利经历获取模块110。
论著和专利经历获取模块110，用于获取每个用户的学术成果发表中的论著经历和专利经历。
其中，经历抽取模块200包括论著经历抽取模块210、专利经历抽取模块220。
论著经历抽取模块210，用于从每个用户的学术成果发表的论著经历中抽取每个论著经历的时间段及该时间段用户所属的学术方向
实施例12：
根据实施例10提供的学术关系知识图谱生成系统，
其中，交集信息获取模块400包括第一交集信息获取模块410。
第一交集信息获取模块410，用于通过对经历进行匹配获取时间段交集不为空且学术方向交集不为空的属于不同用户的每两个经历，获取交集部分的时间段和学术方向信息。
其中，关系生成模块500包括同时段关系生成模块510。
同时段关系生成模块510，用于根据用户的经历，判断每两个用户之间的同时段关系，生成该关系的实体、标签、属性进行关联后加入同时段用户学术关系知识库；
论著经历同时段关系生成模块511，用于如果所述两个经历分别为两个用户的论著经历，且这两个用户具有同时段关系，则将这两个用户的姓名作为该关系的实体，将同时段学术同行关系作为该关系的标签，将交集部分的时间段和学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入同时段用户学术关系知识库；
专利经历同时段关系生成模块512，如果所述两个经历分别为两个用户的专利经历，且这两个用户具有同时段关系，则将这两个用户的姓名作为该关系的实体，将同时段学术同行关系作为该关系的标签，将交集部分的时间段和学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入同时段用户学术关系知识库；
实施例13：
根据实施例10提供的学术关系知识图谱生成系统，
其中，交集信息获取模块400包括第二交集信息获取模块420。
第二交集信息获取模块420，用于通过对经历进行匹配获取时间段交集为空且学术方向交集不为空的属于不同用户的每两个经历，获取交集部分的学术方向信息
其中，关系生成模块500包括不同时段关系生成模块520。
不同时段关系生成模块520，用于根据用户的经历，判断每两个用户之间的不同时段关系，生成该关系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库。
论著经历不同时段关系生成模块521，用于如果所述每两个经历分别为两个用户的论著经历，且这两个用户具有不同时段关系，则将这两个用户的姓名作为该关系的实体，将不同时段学术同行关系作为该关系的标签，将交集部分的学术方向信息作为该关系的属性，将该 关系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库；
专利经历不同时段关系生成模块522，用于如果所述两个经历分别为两个用户的专利经历，且这两个用户具有不同时段关系，则将这两个用户的姓名作为该关系的实体，将不同时段学术同行关系作为该关系的标签，将交集部分的时间段和学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库；
论著专利经历不同时段关系生成模块523，如果所述两个经历分别为一个用户的论著经历和另一个用户的专利经历，且这两个用户具有不同时段关系，则将这两个用户的姓名作为该关系的实体，将不同时段学术同行关系作为该关系的标签，将交集部分的学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库。
实施例14：
根据实施例10提供的用户学术关系知识图谱生成系统，
知识图谱生成模块600包括：
同时段用户学术关系知识图谱生成模块610：将每个用户作为同时段用户学术关系知识图谱中的一个实体，将同时段用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户在同时段用户学术关系知识图谱中对应的实体之间的关联关系。
实体获取模块700包括：
同时段实体获取模块710：获取同时段用户学术关系知识图谱中与任一实体具有关联关系的所有实体，可以得到该一实体对应的用户的所有同时段的学术同行。获取同时段用户学术关系知识图谱中与任一实体具有的关联关系中具有第k级学术方向名称的所有实体，可以得到该一实体对应的用户的所有同时段同第k级学术方向的学术同行。K为大于1的自然数。
实施例15：
根据实施例10提供的用户学术关系知识图谱生成系统，
知识图谱生成模块600包括：
不同时段用户学术关系知识图谱生成模块620：将每个用户作为不同时段用户学术关系知识图谱中的一个实体，将不同时段用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户在不同时段用户学术关系知识图谱中对应的实体之间的关联关系，
实体获取模块700包括：
不同时段实体获取模块720：获取不同时段用户学术关系知识图谱中与任一实体具有关联关系的所有实体，可以得到该一实体对应的用户的所有不同时段的学术同行。获取不同时段用户学术关系知识图谱中与任一实体具有的关联关系中具有第k级学术方向名称的所有实体，可以得到该一实体对应的用户的所有不同时段同第k级学术方向的学术同行。K为大于1的自然数。
实施例16：
根据实施例10提供的用户学术关系知识图谱生成系统，
知识图谱生成模块600包括：
混合时段用户学术关系知识图谱生成模块630：将每个用户作为用户学术关系知识图谱中的一个实体，将同时段用户学术关系知识库及不同时段用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户在用户学术关系知识图谱中对应的实体之间的关联关系。
实体获取模块700包括：
混合时段实体获取模块730：获取用户学术关系知识图谱中与任一实体具有关联关系的所有实体，可以得到该一实体对应的用户的所有学术同行。获取用户学术关系知识图谱中与任一实体具有的关联关系中具有第k级学术方向名称的所有实体，可以得到该一实体对应的用户的所有同第k级学术方向的学术同行。K为大于1的自然数。
实施例17：
根据实施例14、15、16提供的用户学术关系知识图谱生成系统可以进行组合使用。
实施例18：
根据实施例10提供的用户学术关系知识图谱生成系统，
关系获取模块800具体包括：根据上述步骤获取与一个用户在知识图谱中具备某类关系的所有用户，获取该个用户与该所有用户中每一用户之间的关系，将关系中时间段的长度作为P1，将关系中的最低学术方向级别作为P2，计算f(P1，P2)作为该关系的权重，计算该个用户与该所有用户中每一用户之间的所有关系的权重和，根据对应的关系权重和从高到低对与该用户有该类关系的每一用户进行排序，排序越靠前的用户与该用户的该类关系越密切，排序最靠前的用户与该用户的该类关系最密切。
上述各实施例中的方法和系统可以在计算机、服务器、云服务器、超级计算机、机器人、嵌入式设备、电子设备等上执行和部署。
以上所述实施例仅表达了本发明的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对本发明专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本发明构思的前提下，还可以做出若干变形和改进，这些都属于本发明的保护范围。因此，本发明专利的保护范围应以所附权利要求为准。

Claims (10)

  1. 一种学术关系知识图谱生成方法，其特征在于，所述方法包括：
    经历获取步骤，获取每个用户的学术成果发表中的经历；
    经历抽取步骤，从每个用户的学术成果发表的经历中抽取每个经历的时间段及该时间段用户所属的学术方向；
    经历求交步骤，求属于不同用户的每两个经历的交集；
    交集信息获取步骤，通过对经历进行匹配获取具有交集的每两个经历，获取交集部分的信息；
    关系生成步骤，根据用户的经历，判断每两个用户之间的关系，生成该关系的实体、标签、属性进行关联后加入用户学术关系知识库；
    知识图谱生成步骤，用于将每个用户作为用户学术关系知识图谱中的一个实体，将用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户的用户学术关系知识图谱中对应的实体之间的关联关系；
    实体获取步骤，获取用户学术关系知识图谱中与任一实体具有关联关系的实体；
    关系获取步骤，获取与第一用户在知识图谱中具备关系的第二用户，获取该第一用户与该第二用户之间的关系，计算该第一用户与该第二用户之间的关系的权重和，根据对应的关系权重和从高到低对与该第一用户有关系的每一用户进行排序。
  2. 根据权利要求1所述的学术关系知识图谱生成方法，其特征在于，
    所述经历获取步骤包括：
    论著和专利经历获取步骤，获取每个用户的学术成果发表中的论著经历和专利经历；
    所述经历抽取步骤包括：
    论著经历抽取步骤，从每个用户的学术成果发表的论著经历中抽取每个论著经历的时间段及该时间段用户所属的学术方向；
    专利经历抽取步骤，从学术成果发表的专利经历中抽取每个专利经历的时间段及该时间段用户所属的学术方向。
  3. 根据权利要求1所述的学术关系知识图谱生成方法，其特征在于，
    交集信息获取步骤包括：
    第一交集信息获取步骤，通过对经历进行匹配获取时间段交集不为空且学术方向交集不为空的属于不同用户的每两个经历，获取交集部分的时间段和学术方向信息；
    第二交集信息获取步骤，通过对经历进行匹配获取时间段交集为空且学术方向交集不为空的属于不同用户的每两个经历，获取交集部分的学术方向信息；
    关系生成步骤包括：
    同时段关系生成步骤，根据用户的经历，判断每两个用户之间的同时段关系，生成该关系的实体、标签、属性进行关联后加入同时段用户学术关系知识库；
    不同时段关系生成步骤，根据用户的经历，判断每两个用户之间的不同时段关系，生成该关系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库。
  4. 根据权利要求3所述的学术关系知识图谱生成方法，其特征在于，
    同时段关系生成步骤包括：
    论著经历同时段关系生成步骤，如果所述两个经历分别为两个用户的论著经历，且这两个用户具有同时段关系，则将这两个用户的姓名作为该关系的实体，将同时段学术同行关系作为该关系的标签，将交集部分的时间段和学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入同时段用户学术关系知识库；
    专利经历同时段关系生成步骤，如果所述两个经历分别为两个用户的专利经历，且这两个用户具有同时段关系，则将这两个用户的姓名作为该关系的实体，将同时段学术同行关系作为该关系的标签，将交集部分的时间段和学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入同时段用户学术关系知识库；
    论著专利经历同时段关系生成步骤，如果所述两个经历分别为一个用户的论著经历和另一个用户的专利经历，且这两个用户具有同时段关系，则将这两个用户的姓名作为该关系的实体，将同时段学术同行关系作为该关系的标签，将交集部分的时间段和学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入同时段用户学术关系知识库；
    不同时段关系生成步骤包括：
    论著经历不同时段关系生成步骤，如果所述每两个经历分别为两个用户的论著经历，且这两个用户具有不同时段关系，则将这两个用户的姓名作为该关系的实体，将不同时段学术同行关系作为该关系的标签，将交集部分的学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库；
    专利经历不同时段关系生成步骤，如果所述两个经历分别为两个用户的专利经历，且这两个用户具有不同时段关系，则将这两个用户的姓名作为该关系的实体，将不同时段学术同行关系作为该关系的标签，将交集部分的时间段和学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库；
    论著专利经历不同时段关系生成步骤，如果所述两个经历分别为一个用户的论著经历和另一个用户的专利经历，且这两个用户具有不同时段关系，则将这两个用户的姓名作为该关系的实体，将不同时段学术同行关系作为该关系的标签，将交集部分的学术方向信息作为该关系的属性，将该关系的实体、标签、属性进行关联后加入不同时段用户学术关系知识库。
  5. 根据权利要求1所述的学术关系知识图谱生成方法，其特征在于，
    知识图谱生成步骤包括：
    同时段用户学术关系知识图谱生成步骤，将每个用户作为同时段用户学术关系知识图谱中的一个实体，将同时段用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户在同时段用户学术关系知识图谱中对应的实体之间的关联关系；
    不同时段用户学术关系知识图谱生成步骤，将每个用户作为不同时段用户学术关系知识图谱中的一个实体，将不同时段用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户在不同时段用户学术关系知识图谱中对应的实体之间的关联关系；
    混合时段用户学术关系知识图谱生成步骤，将每个用户作为用户学术关系知识图谱中的一个实体，将同时段用户学术关系知识库及不同时段用户学术关系知识库中每个关系的关系 标签及其属性作为该关系中两个用户在用户学术关系知识图谱中对应的实体之间的关联关系；
    实体获取步骤包括：
    同时段实体获取步骤，获取同时段用户学术关系知识图谱中与任一实体具有关联关系的所有实体，可以得到该一实体对应的用户的所有同时段的学术同行；
    不同时段实体获取步骤，获取不同时段用户学术关系知识图谱中与任一实体具有关联关系的所有实体，可以得到该一实体对应的用户的所有不同时段的学术同行；
    混合时段实体获取步骤，获取用户学术关系知识图谱中与任一实体具有关联关系的所有实体，可以得到该一实体对应的用户的所有学术同行。
  6. 根据权利要求1所述的学术关系知识图谱生成方法，其特征在于，
    关系获取步骤具体包括：获取与一个用户在知识图谱中具备某类关系的所有用户，获取该个用户与该所有用户中每一用户之间的关系，将关系中时间段的长度作为P1，将关系中的最低学术方向级别作为P2，计算f(P1，P2)作为该关系的权重，计算该个用户与该所有用户中每一用户之间的所有关系的权重和，根据对应的关系权重和从高到低对与该用户有该类关系的每一用户进行排序，排序越靠前的用户与该用户的该类关系越密切，排序最靠前的用户与该用户的该类关系最密切。
  7. 一种学术关系知识图谱生成系统，其特征在于，所述系统包括：
    经历获取模块，用于获取每个用户的学术成果发表中的经历；
    经历抽取模块，用于从每个用户的学术成果发表的经历中抽取每个经历的时间段及该时间段用户所属的学术方向；
    经历求交模块，用于求属于不同用户的每两个经历的交集；
    交集信息获取模块，用于通过对经历进行匹配获取具有交集的每两个经历，获取交集部分的信息；
    关系生成模块，用于根据用户的经历，判断每两个用户之间的关系，生成该关系的实体、标签、属性进行关联后加入用户学术关系知识库；
    知识图谱生成模块，用于将每个用户作为用户学术关系知识图谱中的一个实体，将用户学术关系知识库中每个关系的关系标签及其属性作为该关系中两个用户的用户学术关系知识图谱中对应的实体之间的关联关系；
    实体获取模块，获取用户学术关系知识图谱中与任一实体具有关联关系的实体；
    关系获取模块，获取与第一用户在知识图谱中具备关系的第二用户，获取该第一用户与该第二用户之间的关系，计算该第一用户与该第二用户之间的关系的权重和，根据对应的关系权重和从高到低对与该第一用户有关系的每一用户进行排序。
  8. 一种计算机设备，包括存储器和处理器，所述存储器存储有计算机程序，其特征在于，所述处理器执行所述计算机程序时实现权利要求1至6中任一项所述方法的步骤。
  9. 一种存储介质，其上存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现权利要求1至6中任一项所述方法的步骤。
  10. 一种学术关系知识图谱生成机器人系统，所述机器人系统包括存储器和处理器，所述存储器存储有机器人程序，其特征在于，所述处理器执行所述机器人程序时实现权利要求1至6中任一项所述方法的步骤。
PCT/CN2020/100815 2019-08-29 2020-07-08 基于人工智能的学术关系知识图谱生成方法和机器人系统 WO2021036524A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910807251.3 2019-08-29
CN201910807251.3A CN110502671A (zh) 2019-08-29 2019-08-29 基于人工智能的学术关系知识图谱生成方法和机器人系统

Publications (1)

Publication Number Publication Date
WO2021036524A1 true WO2021036524A1 (zh) 2021-03-04

Family

ID=68590329

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/100815 WO2021036524A1 (zh) 2019-08-29 2020-07-08 基于人工智能的学术关系知识图谱生成方法和机器人系统

Country Status (2)

Country Link
CN (1) CN110502671A (zh)
WO (1) WO2021036524A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110502671A (zh) * 2019-08-29 2019-11-26 南京智慧光信息科技研究院有限公司 基于人工智能的学术关系知识图谱生成方法和机器人系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108009278A (zh) * 2017-12-20 2018-05-08 华南师范大学 一种基于时态推理的学术关系精准判别方法和系统
US10311442B1 (en) * 2007-01-22 2019-06-04 Hydrojoule, LLC Business methods and systems for offering and obtaining research services
CN110110090A (zh) * 2018-01-09 2019-08-09 鸿合科技股份有限公司 搜索方法、教育搜索引擎系统及装置
CN110502671A (zh) * 2019-08-29 2019-11-26 南京智慧光信息科技研究院有限公司 基于人工智能的学术关系知识图谱生成方法和机器人系统

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150095303A1 (en) * 2013-09-27 2015-04-02 Futurewei Technologies, Inc. Knowledge Graph Generator Enabled by Diagonal Search
CN104346446A (zh) * 2014-10-27 2015-02-11 百度在线网络技术（北京）有限公司 一种基于知识图谱的论文关联信息推荐方法及装置
CN108920556B (zh) * 2018-06-20 2021-11-19 华东师范大学 基于学科知识图谱的推荐专家方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10311442B1 (en) * 2007-01-22 2019-06-04 Hydrojoule, LLC Business methods and systems for offering and obtaining research services
CN108009278A (zh) * 2017-12-20 2018-05-08 华南师范大学 一种基于时态推理的学术关系精准判别方法和系统
CN110110090A (zh) * 2018-01-09 2019-08-09 鸿合科技股份有限公司 搜索方法、教育搜索引擎系统及装置
CN110502671A (zh) * 2019-08-29 2019-11-26 南京智慧光信息科技研究院有限公司 基于人工智能的学术关系知识图谱生成方法和机器人系统

Also Published As

Publication number Publication date
CN110502671A (zh) 2019-11-26

Similar Documents

Publication Publication Date Title
CN109918532B (zh) 图像检索方法、装置、设备及计算机可读存储介质
He et al. Category-aware next point-of-interest recommendation via listwise Bayesian personalized ranking.
CN110555112B (zh) 一种基于用户正负偏好学习的兴趣点推荐方法
Zhang et al. ORec: An opinion-based point-of-interest recommendation framework
Massimo et al. Harnessing a generalised user behaviour model for next-POI recommendation
CN110795571B (zh) 基于深度学习和知识图谱的文化旅游资源推荐方法
CN110147421B (zh) 一种目标实体链接方法、装置、设备及存储介质
CN110390352A (zh) 一种基于相似性哈希的图像暗数据价值评估方法
Zhou et al. Relevance feature mapping for content-based multimedia information retrieval
CN108986872A (zh) 用于大数据电子病历约简的多粒度属性权重Spark方法
Zhang et al. Spatiotemporal activity modeling under data scarcity: A graph-regularized cross-modal embedding approach
WO2021036524A1 (zh) 基于人工智能的学术关系知识图谱生成方法和机器人系统
Sun Personalized music recommendation algorithm based on spark platform
Zhou et al. MetaMove: On improving human mobility classification and prediction via metalearning
Setiowati et al. Point of Interest (POI) Recommendation System using Implicit Feedback Based on K-Means+ Clustering and User-Based Collaborative Filtering.
Huang et al. KG2Rec: LSH-CF recommendation method based on knowledge graph for cloud services
Xu et al. A new self-adaptive hybrid Markov topic model POI recommendation in social networks
Zuo et al. Hierarchical Prompts for Rehearsal-free Continual Learning
Zhu et al. Personalized location recommendations with local feature awareness
Li et al. From low-level geometric features to high-level semantics: An axiomatic fuzzy set clustering approach
Rao et al. A knowledge ecosystem for the food, energy, and water system
Yuan et al. PRPOIR: exploiting the region-level interest for POI recommendation
Liu et al. Bootstrapping operation-level web service ontology: A bottom-up approach
US12124522B2 (en) Search result identification using vector aggregation
US20230214432A1 (en) Search result identification using vector aggregation

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: 20858325

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: 20858325

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 050822)

122 Ep: pct application non-entry in european phase

Ref document number: 20858325

Country of ref document: EP

Kind code of ref document: A1