TWI664539B - System, apparatus and method for monitoring internet media events based on a constructed industry knowledge graph database - Google Patents

Abstract

The invention provides a method for constructing an industry knowledge map database, including the following steps: obtaining industry data from a data source; performing data processing on the industry data to extract entities related to the industry and corresponding entity attributes and / or Entity relationship; building the industry knowledge map database based on the extracted entities, entity attributes and / or entity relationships. The invention also provides a method for monitoring specific media events related to the industry based on the constructed industry knowledge map database, which includes the following steps: obtaining Internet media data; performing event detection, event evaluation, and Screening to obtain the specific media event related to the industry; identifying directly related entities corresponding to the specific media event; accessing the industry knowledge map database based on the directly related entities to determine the specific media event An indirectly related entity corresponding to the event; and sending an early warning message to the directly related entity and / or the indirectly related entity.

Description

Method and device for monitoring internet media events based on industry knowledge map database And system

The invention relates to the field of Internet media monitoring, and in particular, to a technology for constructing an industry knowledge map database and a technology for monitoring Internet media events based on the constructed industry knowledge map database.

The rapid development of computer, communication and network technology has made the performance of terminal equipment including PCs, tablets, smart phones, Internet TVs and so on constantly improve. Correspondingly, Internet media, especially Internet social media, has gradually become one of the main channels for the public to obtain news information due to its diversity, swiftness, interactivity, easy replication, and multimedia.

However, while Internet media information has the advantages of timeliness, flexible and convenient access, and the open nature of its information sources and dissemination methods, it also leads to the following problems: Without authorization or confirmation, some sensitive Messages (for example, trade secrets) and even false messages are rapidly spread by a large number of users on the Internet media platform, which has evolved into media events that have an adverse impact on relevant individuals, businesses / institutions, industries, and society. Therefore, it is necessary to monitor media events in the Internet media and take corresponding measures after monitoring media events that meet certain conditions to reduce or eliminate their potential impact.

The existing Internet media monitoring technologies have the following disadvantages: 1) Use interest matching to provide users with Internet media monitoring. Users need to customize the content topics and related entities that are of interest, so they can only be identified and used in monitoring. Events that are directly related to the defined entity, but cannot identify events that are not defined by the user but indirectly related to the entity that the user is interested in; 2) the properties of the monitoring object are single and can only provide for a single media category and data source (for example, Specific social media, news media, forums, blogs, etc.), single data type (generally text), single language monitoring.

An object of the present invention is to provide a technology for constructing an industry knowledge map database, extracting and storing relevant data for a specific industry or field in the knowledge map database, and the constructed industry knowledge map database can be applied to Internet media monitoring In order to achieve automatic and in-depth monitoring of related Internet media events.

Another object of the present invention is to provide a technology for monitoring Internet media events based on a constructed knowledge library database of industry, which can identify non-directly related entities corresponding to specific media events during monitoring, and can detect multiple types of Internet media data for monitoring.

In order to achieve the above-mentioned object of the invention, specific technical solutions provided by the present invention are as follows.

The invention provides a method for constructing an industry knowledge map database, which includes the following steps: obtaining industry data from a data source; performing data processing on the industry data to extract entities related to the industry and corresponding entity attributes and / or Entity relationship; building the industry knowledge map database based on the extracted entities, entity attributes and / or entity relationships.

Preferably, the step of obtaining industry data is achieved by: obtaining structured industry data from a third-party industry database, the structured industry data includes a plurality of fields; and the step of data processing the industry data This is achieved by: performing data cleaning and extraction-transformation-loading (ETL) processing on the structured industry data; the steps of constructing an industry knowledge graph database are achieved by: based on the extracted entities and entity attributes And / or entity relationship to generate the row Industry knowledge map database.

Preferably, the step of obtaining industry data is achieved by using web crawler technology to obtain industry-related data from Internet data sources, which include unstructured or semi-structured data sources; The steps of processing data of industry data are described as follows: using information extraction technology in natural language processing to perform entity identification and relationship extraction on the relevant data of the industry to extract the entity, entity attributes and / or entities Relationship; the step of constructing an industry knowledge map database is implemented in the following manner: supplementing or updating the industry knowledge map database based on the extracted entities, entity attributes, and / or entity relationships. Further preferably, the above steps are performed periodically at a predetermined cycle.

Preferably, the step of obtaining industry data is achieved by: using an application program interface (API) to obtain industry-related data from an Internet data source by query, the Internet data source includes an open data source; the The step of processing data for industry data is achieved by: before extracting entities related to the industry and corresponding entity attributes and / or entity relationships, performing data cleaning and extraction-conversion on the industry-related data- Loading (ETL) processing; the step of constructing an industry knowledge graph database is implemented by: supplementing or updating the industry knowledge graph database based on the extracted entities, entity attributes, and / or entity relationships. Further preferably, the above steps are performed periodically at a predetermined cycle.

Preferably, the step of obtaining industry data is achieved by: using an application programming interface (API) or a web crawler technology to obtain industry-related Internet media data from Internet data sources; and performing data processing on industry data The steps are implemented by: performing event detection, event evaluation, and screening on the Internet media materials to extract specific media events related to the industry, and identifying corresponding directly related entities from the Internet media materials; The steps of building an industry knowledge map database are described in the following way: based on the specific media The physical event and the corresponding directly related entity supplement the industry knowledge graph database, wherein the specific media event is added to the industry knowledge graph database as an abstract entity. Further preferably, in the step of processing data of industry data, a directly related entity corresponding to the specific media event is identified by at least one of the following methods: identifying from textual data based on entity recognition in natural language processing Entities; identify entities from image or video materials based on image or video recognition processing; or identify entities from audio or video materials based on speech recognition processing. Further preferably, the specific media event includes a negative event, an emergency event, a crisis event, a group event, a public opinion event, or other events of industrial significance. Further preferably, the above steps are performed immediately and without interruption.

Preferably, the step of constructing an industry knowledge map database includes: performing semantic disambiguation and entity connection on the extracted entities. Further preferably, the step of performing semantic disambiguation and entity connection on the extracted entities is further implemented by at least one of the following methods: based on the entity knowledge, semantic disambiguation is performed independently for each of the extracted entity references. Entity connection; Based on the topic consistency hypothesis, use the association of candidate entities in the knowledge base to perform consistent semantic disambiguation and entity connection on the extracted entity references.

The present invention also provides a method for monitoring specific media events related to the industry based on the industry knowledge map database constructed in the present invention, including the following steps: obtaining Internet media data; performing event detection based on the obtained Internet media data, Event evaluation and screening to obtain the specific media event related to the industry; identify directly related entities corresponding to the specific media event; based on the directly related entities, access the industry knowledge map database to determine The non-directly related entity corresponding to the specific media event; and sending an early warning message to the directly related entity and / or the non-directly related entity.

Preferably, the event detection in the event detection, event evaluation and screening steps includes the following steps: the content in the obtained Internet media materials is updated Perform topic classification to obtain content specific to a topic; identify the entities involved from the obtained content; perform sentiment analysis on the obtained content and the identified entities, and filter the obtained content based on the results of sentiment analysis ; Event discovery based on filtered content to cluster media events and discover new media events. Further preferably, the event detection further includes the steps of analyzing the authenticity of the event based on the attributes of the media event, and sorting and / or filtering the media event according to the analysis result.

Preferably, in the step of identifying a directly related entity corresponding to a specific media event, the directly related entity corresponding to the specific media event is identified in at least one of the following ways: based on the entity recognition in natural language processing from the text Identifying entities in data; identifying entities from image or video data based on image or video recognition processing; or identifying entities from audio or video data based on speech recognition processing.

Preferably, the step of accessing the industry knowledge map database is implemented in the following manner: based on the directly related entities, querying in the industry knowledge map database to determine the non-directly related entities.

Preferably, the step of accessing the industry knowledge map database is implemented by using the data acquisition technology in the industry knowledge map database based on the directly related entities to determine the indirectly related entities.

The invention also provides a device for constructing an industry knowledge map database, which includes: a data acquisition module for acquiring industry data from a data source; a data processing module for performing data processing on the industry data to extract The industry-related entities and corresponding entity attributes and / or entity relationships are described; a database construction module is configured to construct the industry knowledge graph database based on the extracted entities, entity attributes and / or entity relationships.

Preferably, the data acquisition module obtains industry data by: obtaining structured industry data from a third-party industry database, the structured industry data includes multiple fields; and the data processing module uses the following methods: Information Office Management: Before extracting entities related to the industry and corresponding entity attributes and / or entity relationships, perform data cleaning and extract-transform-load (ETL) processing on the structured industry data; construction of the database The module constructs an industry knowledge map database by generating the industry knowledge map database based on the extracted entities, entity attributes and / or entity relationships.

Preferably, the data acquisition module obtains industry data through the following methods: using web crawler technology to obtain industry-related data from Internet data sources, the Internet data sources include unstructured or semi-structured data sources; The data processing module performs data processing in the following ways: using information extraction technology in natural language processing to perform entity identification and relationship extraction on the industry-related data to extract the entities, entity attributes, and / or entity relationships The database building module constructs an industry knowledge map database by adding or updating the industry knowledge map database based on the extracted entities, entity attributes, and / or entity relationships.

Preferably, the data acquisition module obtains industry data in the following ways: using an application program interface (API) to obtain industry-related data from Internet data sources in a query manner, and the Internet data sources include open data sources; The data processing module performs data processing in the following ways: before extracting entities related to the industry and corresponding entity attributes and / or entity relationships, performing data cleaning and extraction-transformation-loading of the industry-related data (ETL) processing; the database construction module constructs an industry knowledge graph database in the following manner: supplements or updates the industry knowledge graph database based on the extracted entities, entity attributes, and / or entity relationships.

Preferably, the data acquisition module obtains industry data in the following ways: used to obtain industry-related Internet media data from Internet data sources using an application program interface (API) or web crawler technology; the data processing module The group performs data processing in the following ways: performing event detection, event evaluation, and screening on the Internet media materials to extract specific media events related to the industry, and from the Corresponding directly related entities are identified in Internet media materials; the database building module constructs an industry knowledge map database in the following way: based on the specific media event and corresponding directly related entities, the industry knowledge map database is Supplement, wherein the specific media event is added to the industry knowledge map database as an abstract entity.

Preferably, the database building module further identifies directly related entities corresponding to the specific media event by at least one of the following methods: identifying entities from textual data based on entity recognition in natural language processing; and based on images Or video recognition processing to identify entities from images or video materials; or speech recognition processing to identify entities from audio or video materials.

Preferably, the database construction module includes a module for performing semantic disambiguation and entity connection on the extracted entities. Further preferably, the module for performing semantic disambiguation and entity connection on the extracted entities further performs semantic disambiguation and entity connection by at least one of the following methods: based on the entity knowledge, for each extracted entity References are used to perform semantic disambiguation and entity connection one by one independently. Based on the topic consistency assumption, the association of candidate entities in the knowledge base is used to perform consistent semantic disambiguation and entity connection on the extracted entity references.

Preferably, the specific media event includes a negative event, an emergency event, a crisis event, a group event, a public opinion event, or other events of industrial significance.

The invention also provides a system for monitoring specific media events related to the industry, including: a data acquisition unit for obtaining industry data from a data source; a data processing unit for performing data processing on the industry data to extract Entities related to the industry and corresponding entity attributes and / or entity relationships; a database construction unit configured to construct the industry knowledge graph database based on the extracted entities, entity attributes and / or entity relationships; and a data library storage Unit: used to store the constructed knowledge knowledge database; media event monitoring unit: used to obtain Internet media materials, and perform event detection, event evaluation and screening based on the obtained Internet media materials to obtain The specific media event related to the industry, and identifying directly related entities corresponding to the specific media event; a database access unit: for accessing the industry knowledge map database based on the directly related entity to determine the A non-directly related entity corresponding to the specific media event; a message sending unit, configured to send an early warning message to the directly related entity and / or the indirectly related entity.

Preferably, the data acquisition unit includes: a structured data acquisition unit for obtaining structured industry data from a third-party industry database, the structured industry data includes a plurality of fields; the data processing unit includes: Structured data processing unit configured to perform data cleaning and extraction-transformation-loading (ETL) processing on the structured industry data before extracting entities related to the industry and corresponding entity attributes and / or entity relationships The database building unit comprises a database generating unit for generating the industry knowledge map database based on the extracted entities, entity attributes and / or entity relationships.

Preferably, the data acquisition unit includes: an industry-related data acquisition unit, which is used to obtain industry-related data from Internet data sources using web crawler technology, and the Internet data sources include unstructured or semi-structured data Source; the data processing unit includes: an industry-related data processing unit, which uses information extraction technology in natural language processing to perform entity identification and relationship extraction on the industry-related data to extract the entities, entity attributes, and And / or entity relationship; the database construction unit includes: a database supplement / update unit for supplementing or updating the industry knowledge map database based on the extracted entities, entity attributes, and / or entity relationships.

Preferably, the data acquisition unit includes: an industry-related data acquisition unit, which is used to obtain industry-related data from an Internet data source in an inquiry manner using an application programming interface (API), and the Internet data source includes an open data source ; The data processing unit includes: an industry-related data processing unit, for extracting entities related to the industry and corresponding entity attributes and / or entity relationships, Perform data cleaning and extraction-transformation-loading (ETL) processing on the industry-related data; the database construction unit includes: a database supplement / update unit for based on the extracted entities, entity attributes, and / Or entity relationship to supplement or update the industry knowledge map database.

Preferably, the data acquisition unit includes: a media data acquisition unit configured to obtain industry-related Internet media data from an Internet data source by using an application programming interface (API) or a web crawler technology; the data processing unit includes A media data processing unit, configured to perform event detection, event evaluation, and screening on the Internet media data to extract specific media events related to the industry and identify corresponding directly related entities from the Internet media data; The database construction unit includes: a database supplement / update unit for supplementing the industry knowledge map database based on the specific media event and corresponding directly related entities, wherein the specific media event serves as an abstraction Entities are added to the industry knowledge map database.

Preferably, the database supplement / update unit is further configured to perform semantic disambiguation and entity connection on the extracted entities.

Preferably, the media event monitoring unit is further configured to: classify the content in the obtained Internet media materials to obtain content targeted to a specific topic; identify the entities involved from the obtained content; Perform sentiment analysis on the content and the identified entities, and filter the obtained content based on the results of sentiment analysis; perform event discovery based on the filtered content to cluster media events and discover new media events. Further preferably, the media event monitoring unit is further configured to analyze the authenticity of the events based on the attributes of the media events, and sort and / or filter the media events according to the analysis results.

Preferably, the database access unit is further configured to query the industry knowledge map database based on the directly related entities to determine the non-directly related entities.

Preferably, the database access unit is further configured to: based on the directly related entities, use a data acquisition technology in the industry knowledge map database to determine the non-directly related entities.

Preferably, the specific media event includes a negative event, an emergency event, a crisis event, a group event, a public opinion event, or other events of industrial significance.

By implementing the technical solution provided by the present invention, the following technical effects can be obtained: 1) for one or more target areas or industries, automatic and in-depth monitoring of related Internet media events is realized, and non-corresponding to specific media events can be identified Directly related entities; 2) Automatic monitoring of Internet media data from multiple sources, multiple data types, and multiple languages during monitoring.

S11-S15‧‧‧step

S31-S35‧‧‧step

S41-S45‧‧‧step

S421-S422‧‧‧step

S51-S53‧‧‧step

60‧‧‧Source

61‧‧‧Data Acquisition Unit

62‧‧‧Data Processing Unit

63‧‧‧Database building unit

64‧‧‧Database Storage Unit

65‧‧‧Database Access Unit

66‧‧‧Media Event Monitoring Unit

67‧‧‧Message sending unit

601‧‧‧Partner Industry

602‧‧‧ internet source

611‧‧‧ Structured Data Acquisition Unit

612‧‧‧ Industry related data acquisition unit

613‧‧‧Media data acquisition unit

621‧‧‧ structured data processing unit

622‧‧‧Industry-related data processing unit

623‧‧‧Media data processing unit

631‧‧‧Database generation unit

632‧‧‧Database Supplement / Update Unit

The first diagram is an exemplary flowchart of a method for constructing an industry knowledge map database provided by the present invention; the second diagram is an exemplary structured industry profile provided by the present invention; the third diagram is a media event provided by the present invention Exemplary flowchart of a method for monitoring; the fourth diagram is an exemplary flowchart of another method for constructing an industry knowledge map database provided by the present invention; the fifth diagram is another alternative method of constructing an industry knowledge map database provided by the present invention An exemplary flowchart of the method; FIG. 6 is an exemplary functional block diagram of a system for monitoring media events provided by the present invention.

The following describes specific implementations of the present invention in the form of embodiments with reference to the accompanying drawings, so that those skilled in the art can understand the objectives, technical solutions, and advantages of the present invention. Those skilled in the art can understand that the specific implementations described in the form of embodiments are merely exemplary, and can be implemented without these specific contents. The concept of the present invention is now shown.

The invention provides a technology for constructing an industry knowledge map database and a technology for monitoring Internet media events based on the constructed industry knowledge map database, so as to achieve the purpose of the invention.

The invention relates to the application of a Knowledge Graph database technology. Knowledge graph database is a special database for knowledge management, which is convenient for collecting, sorting and extracting knowledge in related fields. Entities, entity attributes, and entity relationships are defined in the knowledge graph database. Among them, entities correspond to things in the real world (for example, a company A, a character X), and each entity can be identified by a globally unique ID. Entity attributes are used to describe the inherent characteristics of the entity (for example, the Chinese and English names of company A and person X). Entity relationships are used to connect entities to describe the connections between entities (for example, the relationship between person X and company A). By constructing a knowledge graph database, you can more efficiently and deeply utilize knowledge composed of entities, entity attributes, and entity relationships to discover the complex relationships between things.

As a kind of database, the knowledge graph database can be stored in various forms. For example, the knowledge graph database can be stored in a traditional relational database, using the semantic network RDF (Resource Description Framework) triples, or a new type of non-relational database. Preferably, the knowledge graph database may be stored using a graph database, such as Neo4j, OrientDB, Titan-BerkeleyDB, HyperGraphDB, and the like.

Depending on the size and purpose of the knowledge graph database, the data sources used to construct the knowledge graph database can be diverse. For example, the data source can be an open encyclopedia source (for example, Baidu Encyclopedia, Wikipedia, etc.), or it can be a structured database (for example, Wikisource, DBpedia, vertical website, or professional information for a specific industry Library, etc.), or any relevant third-party semi-structured or unstructured source of information (for example, professional websites, content published in Internet media, including news, company annual reports, corporate announcements, etc.).

Those skilled in the art should understand that the knowledge map database constructed in the present invention is oriented in a specific field or industry in the construction process, but is not limited to a single industry. The constructed knowledge graph database realizes the integration and connection of entities and events related to one or more industries, the attributes of entities and events, and the relationships between entities and entities, entities and events, and events and events into a knowledge graph. .

The first figure is an exemplary flowchart of a method for constructing an industry knowledge map database provided by the present invention. The method may include steps S11-S15.

In step S11, industry data is obtained from industry data sources, and entities and corresponding entity attributes and entity relationships are extracted from the industry data to generate the industry knowledge map database.

Industry data sources are sources of basic information for one or more specific areas or industries, where these areas or industries are targeted for monitoring. In one embodiment, the industry data source may be a structured industry database to obtain the highest quality industry basic data as much as possible. The structured database can be accessed through an application programming interface (API), and the data can be obtained in a query manner (for example, by a query command).

Through "Extraction-Transform-Load (ETL)" processing, the obtained industry data can be transformed, and then entities, entity attributes and entity relationships can be extracted from the transformed data and loaded. To the industry knowledge map database proposed by the present invention. The specific implementation steps of the ETL operation can be realized through existing data integration methods. For example, in the ontology-based data integration method, the mapping relationship between various fields in different databases and various entity information is defined in a predetermined way, so as to extract entities and entity attributes according to the fields and their contents And entity relationships to complete the construction of a basic industry knowledge map database. In addition, due to the differences in the structure of the industry database and the problems of data noise, missing data, or data errors, it may be necessary to perform data cleaning operations on the industry data. The data cleaning operation may be implemented by combining technical means known in the art with ETL processing.

As an example, the second figure illustrates an exemplary structured industry profile, which may be obtained from a structured industry database, as described above. In the second figure, Table 1 is an example of the structured information of a listed company, which includes two data items of company A and company B. Each data item includes the Chinese and English names of the company, registered address, stock code, chairman of the board, etc. Multiple fields. By performing ETL operations on the structured data, you can extract the entities (i.e. company A, company B, person X, person Y), entity attributes (specific information of company A and company B), and entity relationships (i.e. A's and X's relationship with company B's and Y's), thus generating a knowledge graph database for their industry.

In another embodiment, the industry data source can also be a semi-structured or unstructured data source from the Internet, and industry data can be captured from the data source through a web crawler technology, and a natural language processing technology-based Information extraction operations to extract entities, entity attributes, and entity relationships.

In step S12, information related to the industry is obtained from an Internet data source, and entities related to the industry and corresponding entity attributes and entity relationships are extracted from the data.

In this step, the relevant information in the above specific field or industry is first obtained from the Internet data source. Internet sources can be structured, semi-structured, or unstructured sources. Therefore, according to the different structural characteristics of Internet data sources, different methods can be used to obtain industry-related data. Then, extract the entities and the corresponding entity attributes and entity relationships from the industry-related data.

For structured Internet data sources, you can query the corresponding data content and obtain entities, entity attributes, and entity relationships through APIs. For semi-structured data sources, you can analyze the content through information extraction operations in natural language processing technology after grabbing the data content, thereby extracting industry-related entities, entity attributes, and entity relationships. A semi-structured data source is a data source that contains partially structured and partially unstructured data. Approach to semi-structured data. For example, HTML and XML files are the most common semi-structured data. In the process of processing HTML and XML files, on the one hand, you can use structured information based on tags, on the other hand, you can combine information extraction technology and machine learning technology to extract the required information.

Information extraction operations include entity recognition operations and relationship extraction operations.

The entity recognition operation may use existing natural language processing tools (for example, part-of-speech tagging or named entity recognition tools), or use a machine learning method to train an entity recognition model for specific tagging data. It should be pointed out that some natural language processing tasks and processing tools are language-dependent (for example, Chinese materials require word segmentation and English materials do not). Machine learning methods digitally represent materials in different languages and formats, and then use general, language-independent algorithms (such as conditional random field algorithms and hidden Markov models) for model training.

The relation extraction operation can be implemented through a variety of existing statistical learning or machine learning methods. For example, you can use the template learning method, taking the entities that meet a certain relationship in the knowledge graph database as examples, extracting and counting the sentence patterns, contexts, etc. that appear in existing text in a large amount of text to form a relationship extraction template, and then The resulting template is used in textual materials to extract new instances. If you extract an instance that does not yet exist in the knowledge graph database, you can add it to the knowledge graph database.

In step S13, the industry knowledge map database is supplemented or updated based on the industry-related entities and corresponding entity attributes and entity relationships.

After extracting industry-related entities and corresponding entity attributes and entity relationships, you can associate and compare them with the corresponding information in the knowledge graph database, and add new entities, entity attributes, and entity relationships to the Knowledge graph database, and can update existing entity attributes and entity relationships.

As mentioned above, the industry knowledge graph database proposed by the present invention may use a traditional relational database, RDF triples database, or a new type. Non-relational database (for example, a graph database). Correspondingly, the specific operations of supplementing or updating the knowledge graph database can be implemented in a customized manner using a database query language. For example, these database query languages include SQL language for relational databases, and RDF triple query language SPARQL. Cypher language for Neo4j graph database.

Continue to explain with the example in the second figure. Assuming that the structured data of listed company executives in Table 2 are obtained from the structured Internet data source through API query, the following additions and updates can be made to the industry knowledge map database: 1) the entity Z, the entity of person Z Attributes and the relationship between person Z and company B are added to the knowledge graph database; 2) the entity attributes of person X and person Y are added; 3) the relationship between person Y and company B is updated (that is, from "current position" to "Former time").

In the process of supplementing or updating the industry knowledge graph database, entity connection operations and semantic disambiguation operations are required.

The entity link operation aims to map an entity reference (or entity reference) appearing in the data content to the related entity concept in the knowledge graph database. For example, in the sentences "Jobs was one of the founders of Apple" and "Steve Jobs founded NeXT in the United States in 1985", "Jobs" and "Steve Jo These two entities refer to the same person entity concept "Steve Jobs (ex-CEO of Apple)" in the knowledge graph database, so the entity link operation will be required to These two entities refer to the same entity. Semantic disambiguation is designed to disambiguate entity references. For example, "Apple" refers to an entity that can correspond to multiple ambiguities, such as "Apple (Fruit)", "Apple Inc.", "Apple Daily", "Apple (Movie)", etc., and The "Apple" in the first sentence of the above example should correspond to the corporate entity concept "Apple Inc." in the Knowledge Graph database instead of "Apple (fruit)", "Apple (movie)", or "Apple Daily". Entity connection and semantic disambiguation are usually together ongoing. Because semantic disambiguation is a means of entity connection, and entity connection is the purpose of semantic disambiguation, the two are often used interchangeably or represented each other on different occasions.

Any existing entity connection and semantic disambiguation techniques can be used in the present invention. For example, one type of method is based on entity knowledge to disambiguate and link entity references one by one independently. Entity knowledge includes, but is not limited to, the occurrence probability of the entity, the name distribution of the entity (full name, alias, abbreviation, etc.), the context of the entity (such as co-occurrence information of words, word distribution, etc.), and the entity's presence in the knowledge base Category information (such as corporate entities, personal entities, location entities, etc.). Can use probability-based (such as linear regression or logistic regression, etc.) or machine learning (such as Support Vector Machines, Random Forest, etc.) to learn and train semantic disambiguation based on entity knowledge and Entity Link Model. Another type of method is based on the assumption of topic consistency (that is, the entities in the article are usually related to the text topic, so these entities also have semantic relevance), using candidate entities referred to by all entities in the text content in the knowledge base (such as Wiki The association in the encyclopedia or the knowledge map constructed by the present invention disambiguates and links all entity references in an article consistently. This type of method usually uses collaborative reasoning based on the structure of the graph data in the calculation process. That is, the candidate entities referred to by all entities in the content of the article are used to construct a candidate entity graph using their relationship in the knowledge base. The dense distribution of the graph reflects The degree of semantic association between different candidate entity nodes in the graph. The process of entity connection is to synergistically enhance evidence by repeatedly passing evidence (possible degree of association between different entities) according to the dependency structure of the candidate entity graph to converge. The above two types of methods can also be combined flexibly or organically to improve the performance of disambiguation and connection.

In step S14, Internet industry data related to the industry is obtained from Internet data sources, and specific media events related to the industry and corresponding directly related entities are extracted from the Internet media data.

Internet media sources can be obtained from Internet sources in a number of ways. For example, some social media sites (e.g. Sina Weibo, Facebook, Twitter Etc.) have opened APIs for obtaining their data. You can also use web crawler technology and content extraction technology to crawl news website or industry media website data.

There are various technical implementations in the art for monitoring Internet media to obtain specific media events. For example, in one implementation, the Internet media data is first detected to discover the content of the media event and the entity involved in the specific field or industry of interest, and then the newly discovered media event is indexed by different indicators (E.g., negative, significant, unexpected, speed and scope of transmission, credibility, etc.) to screen media events that meet the requirements.

For different types of Internet media materials, different processing technologies can be used to identify directly related entities corresponding to media events. For example, entity recognition technology based on natural language processing can be used to identify entities from textual materials, image or video recognition processing technology can be used to identify entities from image or video materials, and speech recognition processing technology can be used from audio or video materials To identify entities. Those skilled in the art can understand that the present invention does not limit the media types and language types of Internet media materials.

In step S15, the industry knowledge graph database is supplemented based on the specific media event and the corresponding directly related entity, wherein the specific media event is added to the industry knowledge graph database as an abstract entity. .

After obtaining specific industry-related media events and corresponding directly related entities (for example, the corruption scandal of the chairman of a listed company and the companies, people, and locations involved in the incident), the event is added to the industry knowledge as an abstract entity In the graph database, at the same time, the entities directly related to the event are physically connected and semantically disambiguated, that is, the entity corresponding to the entity in the industry knowledge graph database is found, and it is related to the abstract entity representing the event Make an association. If it is found that the entity involved in the event does not exist in the industry knowledge map database, it can be supplemented in the manner described in step S13 above. After completing the supplement to the industry knowledge map database, the directly related entities based on the event can be compared with other entities in the knowledge map database The relationship between them is to find other entities that are not directly related to the abstract entities representing media events in the industry knowledge graph database.

After the industry knowledge map database is constructed in the above manner, Internet media events can be automated and deeply monitored based on the constructed information. Preferably, after the first construction of the industry knowledge graph database is completed, in order to maintain the integrity and validity of the information, the industry knowledge graph database can also be updated. For example, steps S12 and S13 can be performed periodically at a predetermined cycle. Steps S14 and S15 can also be performed in an immediate and uninterrupted manner.

In addition, those skilled in the art can understand that the contents of various materials such as industry materials, industry-related materials, and Internet media materials in the present invention may be in multiple languages or multiple types (for example, text, graphics Image, video, voice, etc.), the present invention does not make any restrictions on this.

The third figure is an exemplary flowchart of a method for monitoring media events provided by the present invention. The method can monitor specific media events related to the industry based on the industry knowledge map database constructed in the present invention. The method may include steps S31-S35.

In step S31, Internet media materials are acquired.

As mentioned above, there are many ways to obtain Internet media materials from Internet sources. For example, some social media websites (for example, Sina Weibo, Facebook, Twitter, etc.) have opened APIs for obtaining their materials. You can also use web crawler technology and content extraction technology to crawl news website or industry media website data.

In step S32, event detection, event evaluation, and screening are performed based on the obtained Internet media materials to obtain the specific media event related to the industry.

As mentioned above, there are various technical implementations for monitoring Internet media to obtain specific media events in the art. For example, in one implementation, first detect Internet media materials to discover specific areas of interest Or the content of media events in the industry and the entities involved in the event, and then according to different indicators for newly discovered media events (for example, the negativeness, significance, emergencies, speed and scope of transmission, credibility, etc.) Evaluate to identify media events that meet your requirements.

Specifically, in one embodiment, the technical implementation steps involved in event detection may include topic classification, entity recognition, sentiment analysis, and event discovery.

In the topic classification step, topic classification is performed on content in the obtained Internet media material to obtain content specific to a topic. The purpose of topic classification is to filter out texts that belong to a topic of interest or a category related to customer needs from the content obtained. Topic classification is a text mining technology. Generally, machine learning or deep learning methods are used to train a classification model on labeled data, and then applied to the text to determine its topic category. Any existing classification model (for example, Naive Bayes model, decision tree, support vector machine, artificial neural network, etc.) can be used in the present invention.

In the entity recognition step, the entities involved are identified from the obtained content. The purpose of entity extraction is to find out the entities involved in the article for further analysis. For example, entity recognition can include extracting entities from text information using information extraction technology in natural language processing, identifying entities from image (including video) information using image recognition technology, and using speech recognition technology to extract speech information from speech information. Recognize entities. You can also combine entities identified from text, images, and speech.

In the step of sentiment analysis, sentiment analysis is performed on the obtained content and the identified entity, and the obtained content is filtered based on the result of the sentiment analysis. Sentiment analysis is used to judge the full content of the content and the sentiment polarity expressed by different entities to find content that meets the monitoring conditions. The prior art generally implements sentiment analysis using text classification methods (eg, classifying sentiment as positive, neutral, or negative) or regression analysis methods (eg, expressing sentiment as a score between -5 and +5). To judge the sentiment of an entity in the content, you can use the context information of the entity in the text, or use the dependency syntax analysis tool to find the text part of the text that is related to the entity for targeting. Sentiment analysis of entities.

In the event discovery step, event discovery is performed based on the filtered content to cluster media events and discover new media events. The purpose of event discovery is to extract event information (such as the time and place of the event) from different texts, and then cluster and merge the relevant information into an abstract "event". By comparing with existing events to determine new occurrences And cluster events based on similarity or relevance of content.

Optionally, during the event detection process, the authenticity of the event may also be analyzed based on the attributes of the media event (for example, the time and place of the event, the publisher of the media event and its related attributes, etc.), and according to the analysis result Sort and / or filter media events.

Those skilled in the art can understand that the implementation manners listed in the above steps for each operation are merely exemplary, and some other existing manners in the art can also implement these operations, and the present invention does not do a specific way to implement the above operations. No restrictions.

In step S33, a directly related entity corresponding to the specific media event is identified.

In one embodiment, each directly related entity in each media event can be obtained through entity identification and event discovery operations in event monitoring. At the same time, as described above, each directly related entity can be related to the corresponding entity concept in the industry knowledge graph database or supplemented to the industry knowledge graph database through entity connection and semantic disambiguation processing.

In step S34, based on the directly related entities, the industry knowledge map database is accessed to determine non-directly related entities corresponding to the specific media event.

In one embodiment, the industry knowledge graph database can be directly queried on the industry knowledge graph database through preset various conditions. It is not a directly related entity. For example, the preset conditions may be: 1) an entity that is directly related to the event in the N layer (N can be 1, 2, 3 ...); 2) the degree of association with the directly related event entity meets a certain Other entities with certain conditions (such as greater than a specified threshold); 3) entities that have a specific relationship (e.g., supply relationship, investment relationship, etc.) with entities directly related to the event; 4) have a specific attribute (e.g., Entities belonging to a specified industry, located in a location, holding a position, etc.). These preset conditions can be used individually or in any combination.

In another embodiment, a method of data acquisition can be used to mine non-directly related entities of an event on the basis of an industry knowledge map database using multiple conditions. For example, the specific implementation method can use link prediction technology for graph data, that is, the problem of detecting non-directly related entities of an event is expressed as "the nodes representing the event in the database of prediction industry knowledge graph database are directly related to "There is a connected edge between other physical nodes other than the physical node". Conditions that can be used to link predictions include, but are not limited to, the characteristics of the event itself (e.g., the type of event, time and place attributes, negativeness, etc.), the relationship between the event and historical events (including the type of relationship and the strength of the relationship), and the event directly The relationship between related entities and other entities (including the type of relationship and the strength of the relationship) and all the knowledge that can be mined in the knowledge graph database, such as entity types and attributes, so as to achieve a comprehensive judgment of non-directly related entities of a specific media event .

In step S35, an early warning message is sent to the directly related entity and / or the non-directly related entity.

After identifying the direct and non-directly related entities corresponding to a specific media event, multiple channels (for example, email, mobile phone text messaging, instant chat tools, social networking platforms, etc.) can be used to send early warning messages to corresponding entity users. The early warning message can include a text description of the event itself, pictures, dissemination of relevant statistical information, event evaluation indicators, and ways in which related entities may be affected by the event, and so on.

Those skilled in the art can understand that the specific media described in the present invention The event can be various types of events that meet the conditions set by the user and can be obtained from the Internet media, such as negative events, emergencies, crisis events, mass events, or public opinion events. The invention does not place any restrictions on this.

As a preferred embodiment, the fourth figure shows an exemplary flowchart of another method for constructing an industry knowledge map database provided by the present invention. The method may include steps S41, S421 / S422, and S43-S45.

In step S41, industry data is obtained from industry data sources, and entities and corresponding entity attributes and entity relationships are extracted from the industry data to generate an industry knowledge map database.

In step S421, based on the structured data source, an application program interface is used to obtain the entities, entity attributes, and entity relationships related to the industry in a query manner. In one embodiment, the structured data source may be a structured open data platform such as Wikidata and DBPedia, and industry-related data may be obtained from it through an API.

In step S422, based on the semi-structured or unstructured data source, the data is subjected to entity recognition and relationship extraction using natural language processing technology to extract entities, entity attributes and entity relationships related to the industry. In one embodiment, the semi-structured or unstructured data source may be an open data platform such as Wikipedia or Baidu Encyclopedia, or it may be any relevant third-party data source (for example, a professional website, in Internet media Published content, etc.), and industry-relevant data can be obtained through web crawlers or content extraction technologies.

Preferably, steps S421 and / or S422 and S43 may be performed periodically at a predetermined cycle.

In step S43, an industry knowledge map database is supplemented or updated based on the industry-related entities and corresponding entity attributes and entity relationships.

In step S44, Internet media data is obtained from Internet data sources, and specific media events related to the industry are extracted from the Internet media data. And corresponding directly related entities.

In step S45, an industry knowledge map database is supplemented based on the specific media event and the corresponding directly related entity, where the specific media event is added to the industry knowledge map database as an abstract entity.

Preferably, steps S44 and S45 can be performed in an immediate and uninterrupted manner

The fifth figure is an exemplary flowchart of another method for constructing an industry knowledge map database provided by the present invention. The method may include steps S51-S53: in step S51, obtaining industry data from a data source; in step S52, performing data processing on the industry data to extract entities related to the industry and corresponding entity attributes and / Or entity relationship; in step S53, the industry knowledge map database is constructed based on the extracted entities, entity attributes and / or entity relationships.

As mentioned above, the sources of the industry knowledge graph database can be diverse, including but not limited to open encyclopedia sources, structured databases, and any related third-party semi-structured or unstructured Internet source. At the same time, as mentioned above, the data source of the industry knowledge map database can also be the Internet media source.

In one embodiment, the data source may be a structured industry database, and the method may be implemented in the following specific ways: In step S51 (1), obtaining from a third-party industry database includes multiple fields In step S52 (1), before extracting entities related to the industry and corresponding entity attributes and / or entity relationships, perform data cleaning and extraction-conversion on the structured industry data- Load (ETL) processing; in step S53 (1), generate the industry knowledge map database based on the extracted entities, entity attributes and / or entity relationships.

In another embodiment, the data source may be an unstructured or semi-structured Internet data source, and the method may be implemented in the following specific ways: Implementation: In step S51 (2), use web crawler technology to obtain industry-related data from Internet data sources, which include unstructured or semi-structured data sources; in step S52 (2) , Using information extraction technology in natural language processing to perform entity identification and relationship extraction on the industry-related data to extract the entity, entity attributes, and / or entity relationship; in step S53 (2), based on the extracted Supplement or update the industry knowledge graph database based on the entities, entity attributes, and / or entity relationships of the industry knowledge graph database.

In addition, the steps S51 (2) -S53 (2) may be performed periodically with a predetermined cycle.

In another embodiment, the data source may be an open Internet data source, and the method may be implemented in the following specific ways: In step S51 (3), an application program interface (API) is used to query from the Obtain industry-related data from Internet data sources; in step S52 (3), before extracting entities related to the industry and corresponding entity attributes and / or entity relationships, perform data cleaning on the industry-related data And extract-transform-load (ETL) processing; in step S53 (3), the industry knowledge graph database is supplemented or updated based on the extracted entities, entity attributes and / or entity relationships.

In addition, the steps S51 (3) -S53 (3) may be performed periodically with a predetermined cycle.

In another embodiment, the data source may be an Internet media data source, and the method may be implemented in the following specific ways: In step S51 (4), using an application program interface (API) or a web crawler technology, Obtain Internet media data from Internet data sources; in step S52 (4), perform event detection, event evaluation, and screening on the Internet media data to extract specific media events related to the industry, and from the Internet media The corresponding directly related entity is identified in the data; in step S53 (4), the industry knowledge map database is supplemented based on the specific media event and the corresponding directly related entity, where the specific media event is Abstract entities are added to the industry knowledge map database.

For example, in step S52 (4), directly related entities corresponding to a specific media event may be identified by at least one of the following methods: identifying entities from textual data based on entity recognition in natural language processing; based on images or videos The recognition process identifies an entity from the image or video material; or, based on the voice recognition process, identifies the entity from the audio or video material.

For example, the specific media event may include a negative event, an emergency event, a crisis event, a group event, a public opinion event, or other events of industrial significance.

In addition, the steps S51 (4) -S53 (4) may be performed immediately and without interruption.

In another embodiment, the steps of supplementing or updating the industry knowledge graph database in steps S53 (2), S53 (3), and S53 (4) may include: performing semantic disambiguation on the extracted entities And physical connection. For example, the semantic disambiguation and entity connection can be performed in at least one of the following ways: based on entity knowledge, semantic disambiguation and entity connection are performed independently for each extracted entity reference one by one; based on the topic consistency assumption Using the association of candidate entities in the knowledge base, the extracted entity references are subjected to consistent semantic disambiguation and entity connection.

The method for constructing an industry knowledge map database provided by the present invention has been described in the foregoing embodiments. Those skilled in the art can understand that various combinations of these embodiments are also included in the conception of such a method for constructing an industry knowledge map database.

The sixth figure is an exemplary functional block diagram of a system for monitoring media events provided by the present invention. The system includes a data acquisition unit 61, a data processing unit 62, a database construction unit 63, a data library storage unit 64, a media event monitoring unit 66, a database access unit 65, and a message sending unit 67.

The data obtaining unit 61 is configured to obtain industry data from a data source 60.

A data processing unit 62 is configured to perform data processing on the industry data to extract entities related to the industry and corresponding entity attributes and / or entity relationships; a database construction unit 63 is configured to, based on the extracted entities, Entity attributes and / or entity relationships to build the industry knowledge map database; a data library storage unit 64: used to store the constructed industry knowledge map database; a media event monitoring unit 66: used to obtain Internet media data, based on the obtained To perform event detection, event evaluation, and screening on the Internet media materials to obtain the specific media event related to the industry, and identify directly related entities corresponding to the specific media event; a database access unit 65: configured to A related entity, accessing the industry knowledge map database to determine a non-directly related entity corresponding to the specific media event; a message sending unit 67, configured to send to the directly related entity and / or the indirectly related entity Warning message.

In one embodiment, the data acquisition unit 61 includes: a structured data acquisition unit 611, configured to obtain structured data from a third party industry 601 database, the structured data includes a plurality of fields; the data processing The unit 62 includes: a structured data processing unit 621, configured to perform data cleaning and extraction-transformation-loading (ETL) processing on the structured data; the database construction unit 63 includes: a database generation unit 631, The industry knowledge map database is generated based on the extracted entities, entity attributes and / or entity relationships.

In another embodiment, the data acquisition unit 61 includes an industry-related data acquisition unit 612, which is used to obtain industry-related data from Internet data sources by using web crawler technology, and the Internet data source 602 includes non-structured data. Structured or semi-structured data source; the data processing unit 62 includes: industry-related information A material processing unit 622, configured to use entity information extraction technology in natural language processing to perform entity identification and relationship extraction on the industry-related materials to extract the entities, entity attributes, and / or entity relationships; the database construction The unit 63 includes a database supplement / update unit 632 for supplementing or updating the industry knowledge graph database based on the extracted entities, entity attributes and / or entity relationships.

In another embodiment, the data acquisition unit 61 includes: an industry-related data acquisition unit 612, which is used to obtain industry-related data from an Internet data source by using an application program interface (API) in a query manner, the Internet data source 602 includes an open data source; the data processing unit 62 includes: an industry-related data processing unit 622, configured to extract the entities related to the industry and the corresponding entity attributes and / or entity relationships before extracting the related industries Relevant data is subjected to data cleaning and extraction-transformation-loading (ETL) processing; the database construction unit 63 includes: a database supplement / update unit 632, which is used to extract the entities, entity attributes and / or entity relationships based on the extracted entities Supplement or update the industry knowledge map database.

In another embodiment, the data acquisition unit 61 includes: a media data acquisition unit 613, which is configured to acquire industry-related Internet media data from an Internet data source 602 using an application program interface (API) or a web crawler technology; The data processing unit 62 includes: a media data processing unit 623, configured to perform event detection, event evaluation, and screening on the Internet media data to extract specific media events related to the industry, and to remove the media media data from the Internet media data. Corresponding directly related entities are identified in the database; the database construction unit 63 includes: a database supplement / update unit 632 for supplementing the industry knowledge map database based on the specific media event and the corresponding directly related entities , Wherein the specific media event is added to the industry knowledge graph database as an abstract entity.

In one embodiment, the database supplement / update unit 632 is further configured to perform semantic disambiguation and entity connection on the extracted entities.

In one embodiment, the media event monitoring unit 66 further Used for: classifying the content in the obtained Internet media materials to obtain content specific to a topic; identifying the entities involved from the obtained content; performing sentiment analysis on the obtained content and the identified entities, And based on the results of sentiment analysis, the obtained content is filtered; event discovery is performed based on the filtered content to cluster media events and discover new media events. In another embodiment, the media event monitoring unit 66 is further configured to analyze the authenticity of the events based on the attributes of the media events, and sort and / or filter the media events according to the analysis results.

In one embodiment, the database access unit 65 is further configured to: based on the directly related entities, query in the industry knowledge map database to determine the non-directly related entities. In another embodiment, the database access unit 65 is further configured to: based on the directly related entities, use a data acquisition technology in the industry knowledge map database to determine the indirectly related entities.

The system for monitoring media events provided by the present invention is described in the foregoing embodiments. Those skilled in the art can understand that the operation steps in the various methods described above with reference to the first diagram, the third to the fifth diagrams of the accompanying drawings can be applied to the constituent units of the system, and therefore will not be repeated here.

Those skilled in the art should also understand that the various exemplary method steps and units described in connection with the various embodiments disclosed in the present invention can be implemented as electronic hardware, computer software, or a combination of the two. In order to clearly indicate the interchangeability of hardware and software, the various exemplary steps and units described above are generally described around their functions. Whether such functionality is implemented as hardware or software depends on the particular application and design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present disclosure.

The "exemplary / exemplary" used in the specification of the present invention means used as an example, illustration or explanation. Any technical solution described as “exemplary” in the specification should not be interpreted as being better or more advantageous than other technical solutions.

The present invention provides the above description of the disclosed technical content to enable those skilled in the art to implement or use the present invention. To those skilled in the art, many modifications and variations to these technical contents are obvious, and the general principle defined by the present invention can also be applied to other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the specific embodiments shown above, but should be consistent with the widest scope consistent with the inventive concept disclosed by the present invention.

Claims (37)

A method for constructing an industry knowledge map database includes the following steps: Step 101: Obtain industry data of one or more industries in a specific field from a data source; Industry data includes Internet media data; step 102, performing data processing on the industry data of the industry in the one or more specific fields to extract entities related to the industry in the one or more specific fields and corresponding entities Attributes and / or entity relationships, where the entity refers to part of the industry data of the industry in the one or more specific fields; and event detection, event evaluation, and screening of the Internet media data to extract and compare Describe one or more industry-specific specific media events in a specific field; and step 103, perform semantic disambiguation and entity connection on the extracted entities, based on the extracted specific media events and the entities, entity attributes, and / Or entity relationship to build the industry knowledge map database. The method according to item 1 of the scope of patent application, wherein the step 101 further comprises: obtaining structured industry data from a third party industry database, the structured industry data includes a plurality of fields; the step 102 further includes: before extracting entities related to the one or more specific industries and corresponding entity attributes and / or entity relationships, performing data cleaning and extraction-transformation-loading on the structured industry data ( ETL) processing; and the step 103 further comprises: generating the industry knowledge map database based on the extracted entities, entity attributes, and / or entity relationships. The method according to item 1 of the scope of patent application, wherein the step 101 further comprises: using a web crawler technology to obtain data related to the industry in one or more specific fields from an Internet data source. The Internet data source includes an unstructured or semi-structured data source; the step 102 further includes: using information extraction technology in natural language processing to perform the data related to the industry in the one or more specific fields Entity identification and relationship extraction to extract the entity, entity attribute, and / or entity relationship; and the step 103 further includes: performing a process on the industry knowledge map database based on the extracted entity, entity attribute, and / or entity relationship Supplement or update. The method according to item 1 of the scope of patent application, characterized in that the step 101 further comprises: using an application programming interface (API) to obtain information related to the industry in the one or more specific fields from an Internet data source in a query manner. The Internet data source includes an open data source; the step 102 further includes: before extracting entities related to the one or more specific industries and corresponding entity attributes and / or entity relationships, Performing data cleaning and extract-transform-load (ETL) processing on the data related to the industry in the one or more specific fields; and the step 103 further includes: based on the extracted entities, entity attributes, and / or The entity relationship supplements or updates the industry knowledge map database. The method according to item 1 of the scope of patent application, wherein the step 101 further comprises: using an application program interface (API) or a web crawler technology to obtain from the Internet data source and the one or more specific fields Industry-related Internet media materials; the step 102 further includes: identifying the corresponding directly related entities from the Internet media materials; and the step 103 further includes: based on the specific media event and the corresponding directly related entities, The industry knowledge map database is supplemented, wherein the specific media event is added to the industry knowledge map database as an abstract entity. The method according to item 5 of the scope of patent application, characterized in that in step 102, a directly related entity corresponding to the specific media event is further extracted in at least one of the following ways: based on an entity in natural language processing Recognize entities from textual materials; identify entities from image or video materials based on image or video recognition processing; or identify entities from audio or video materials based on speech recognition processing. The method according to any one of claims 3 to 5, wherein the step 103 includes: performing semantic disambiguation and entity connection on the extracted entities to identify the entities in the industry The corresponding entity in the knowledge graph database. The method according to item 7 of the scope of patent application, wherein the steps of performing semantic disambiguation and entity connection on the extracted entities are further implemented by at least one of the following methods: based on the entity knowledge, each extracted Based on the subject consistency assumption, using the association of candidate entities in the knowledge base to perform consistent semantic disambiguation and entity connection on the extracted entity references. The method according to item 1 of the scope of the patent application, wherein the specific media event includes a negative event, an emergency event, a crisis event, a group event, a public opinion event, or other events of industrial significance. The method according to item 3 or 4 of the scope of patent application, wherein the steps 101-103 are performed periodically at a predetermined period. The method according to item 5 of the scope of patent application, wherein the steps 101-103 are performed immediately and without interruption. A method for monitoring a specific media event related to an industry in one or more specific fields according to the industry knowledge map database according to any one of the claims 1 to 11 in the scope of patent applications, which comprises the following steps: Step 1201, obtaining Internet media materials; step 1202, performing event detection, event evaluation, and screening based on the obtained Internet media materials to obtain the specific media event related to the industry; step 1203, extracting the corresponding media event Step 1204, based on the association relationship between the directly related entity and the industry knowledge map database, determine a non-directly related entity corresponding to the specific media event; and step 1205, report to the directly related entity A related entity and / or the indirectly related entity sends an early warning message. The method according to item 12 of the scope of patent application, wherein the event detection in step 1202 includes the following steps: classifying content in the obtained Internet media materials to obtain content targeted to a specific topic; Identify the involved entities from the obtained content; perform sentiment analysis on the obtained content and the identified entities, and filter the obtained content based on the results of the sentiment analysis; and perform event discovery based on the filtered content to Cluster media events and discover new media events. The method according to item 13 of the scope of patent application, wherein the event detection in step 1202 further includes the steps of analyzing the authenticity of the event based on the attributes of the media event, and performing a media event analysis based on the analysis result. Sort and / or filter. The method according to item 12 of the scope of patent application, characterized in that in step 1203, directly related entities corresponding to the specific media event are extracted by at least one of the following methods: based on entity recognition in natural language processing Identify entities from textual data; identify entities from image or video data based on image or video recognition processing; or identify entities from audio or video data based on speech recognition processing. The method according to item 12 of the scope of patent application, wherein the step 1204 is implemented by: performing semantic disambiguation and entity connection on the directly related entities to identify the information in the industry knowledge map database. Corresponding entities, and based on the directly related entities, query other entities and / or entity relationships in the industry knowledge map database with preset conditions to determine the non-directly related entities. The method according to item 12 of the scope of patent application, wherein the step 1204 is implemented by: performing semantic disambiguation and entity connection on the directly related entities to identify the information in the industry knowledge map database. Corresponding entities, and based on the directly related entities, use data extraction technology on other entities and / or entity relationships in the industry knowledge map database to determine the non-directly related entities. An apparatus for constructing an industry knowledge map database is characterized in that it includes: a data acquisition module for obtaining industry data of one or more industries in a specific field from a data source, the industry information of one or more industries in a specific field The industry data includes Internet media data; a data processing module is configured to perform data processing on the industry data of the industry in the one or more specific fields to extract information related to the industry in the one or more specific fields The entity and corresponding entity attributes and / or entity relationships, where the entity refers to part of the industry data of the industry in the one or more specific fields; and event detection, event evaluation, and Screening to extract specific media events related to the industry in the one or more specific fields; and a database building module for semantic disambiguation and entity linking of the extracted entities, based on the extracted specific Media events and the entities, entity attributes, and / or entity relationships build the industry knowledge map database. The device according to item 18 of the scope of patent application, wherein the data acquisition module is further configured to: obtain structured industry data from a third-party industry database, where the structured industry data includes multiple fields; The data processing module is further configured to perform data cleaning and extraction on the structured industry data before extracting entities related to the one or more specific industries and corresponding entity attributes and / or entity relationships. -Conversion-loading (ETL) processing; and the database building module is further configured to generate the industry knowledge map database based on the extracted entities, entity attributes, and / or entity relationships. The device according to item 18 of the scope of application for a patent, wherein the data acquisition module is further configured to use a web crawler technology to obtain information related to the one or more industries in a specific field from an Internet data source. Data, said Internet data sources include unstructured or semi-structured data sources; said data processing module is further configured to: use information extraction technology in natural language processing to Industry-related data for entity identification and relationship extraction to extract the entities, entity attributes, and / or entity relationships; and the database building module is further configured to: based on the extracted entities, entity attributes, and / or entity relationships Supplement or update the industry knowledge map database. The device according to item 18 of the scope of patent application, wherein the data acquisition module is further configured to: use an application programming interface (API) to obtain information from the Internet data source and the one or more specific fields in a query manner. Industry-related data, the Internet data source includes open data sources; the data processing module is further configured to: extract entities related to the industry in the one or more specific fields and corresponding entity attributes and / Before the entity relationship, the data cleaning and extraction-transformation-loading (ETL) processing is performed on the data related to the industry in the one or more specific fields; and the database building module is further used for: The extracted entities, entity attributes and / or entity relationships supplement or update the industry knowledge map database. The device according to item 18 of the scope of the patent application, wherein the data acquisition module is further configured to: use an application programming interface (API) or a web crawler technology to obtain information from the Internet data source and the one or more Industry-specific Internet media data in a specific field; the data processing module is further configured to: identify the corresponding directly related entity from the Internet media data; and the database construction module is further configured to: The specific media event and the corresponding directly related entity supplement the industry knowledge graph database, wherein the specific media event is added to the industry knowledge graph database as an abstract entity. The device according to item 22 of the scope of patent application, wherein the data processing module further extracts directly related entities corresponding to the specific media event in at least one of the following ways: based on entities in natural language processing Recognize entities from textual materials; identify entities from image or video materials based on image or video recognition processing; or identify entities from audio or video materials based on speech recognition processing. The device according to any one of claims 20 to 22, wherein the database building module includes: a module for performing semantic disambiguation and entity linking on the extracted entities, and A corresponding entity of the entity in the industry knowledge map database is identified. The device according to item 24 of the scope of patent application, wherein the module for performing semantic disambiguation and entity connection on the extracted entity further performs semantic disambiguation and entity connection by at least one of the following methods: : Semantic disambiguation and entity connection for each extracted entity reference independently based on entity knowledge; and based on the topic consistency assumption, use the association of candidate entities in the knowledge base to unify the extracted entity references Sexual disambiguation and entity connection. The device according to item 22 of the scope of patent application, wherein the specific media event includes a negative event, an emergency event, a crisis event, a group event, a public opinion event, or other events of industrial significance. A system for monitoring specific media events related to an industry, which is characterized in that it includes: a data acquisition unit for obtaining industry data from one or more industries in a specific field from a data source; a data processing unit for Perform data processing on the industry data of the industry in one or more specific fields to extract entities related to the industry in the one or more specific fields and corresponding entity attributes and / or entity relationships, where the entities are Refers to part of the industry data of the industry in the one or more specific fields; a database building unit that performs semantic disambiguation and entity connection on the extracted entities for use based on the extracted entities, entity attributes and / or entities Relation construction industry knowledge map database; database storage unit: used to store the constructed industry knowledge map database; media event monitoring unit: used to obtain Internet media materials, perform event detection based on the obtained Internet media materials, Event evaluation and screening to obtain the specific media events related to the industry, and to identify A direct related entity corresponding to a specific media event; a database access unit: configured to access the industry knowledge map database based on the directly related entity to determine a non-directly related entity corresponding to the specific media event; and message transmission A unit, configured to send an early warning message to the directly related entity and / or the non-directly related entity. The system according to item 27 of the scope of patent application, wherein the data acquisition unit includes: a structured data acquisition unit for obtaining structured industry data from a third party industry database, and the structured industry data includes A plurality of fields; the data processing unit includes: a structured data processing unit, configured to perform data on the structured industry data before extracting entities related to the industry and corresponding entity attributes and / or entity relationships Cleaning and extraction-transformation-loading (ETL) processing; and the database construction unit includes: a database generation unit for generating the industry knowledge map database based on the extracted entities, entity attributes and / or entity relationships . The system according to item 27 of the scope of application for a patent, wherein the data acquisition unit includes: an industry-related data acquisition unit, which is used to obtain information related to the one or more specific data from an Internet data source using a web crawler technology. Industry-related data in the field, the Internet data sources include unstructured or semi-structured data sources; the data processing unit includes: industry-related data processing units, which use information extraction technology in natural language processing to Entity identification and relationship extraction of the industry-related data in one or more specific fields to extract the entities, entity attributes and / or entity relationships; and the database building unit includes: a database supplement / update unit, Supplementing or updating the industry knowledge map database based on the extracted entities, entity attributes and / or entity relationships. The system according to item 27 of the scope of patent application, wherein the data acquisition unit includes: an industry-related data acquisition unit, which is used to obtain an information source from an Internet data source in an inquiry manner using an application program interface (API) and the first data acquisition unit. Or multiple industry-specific data, the Internet data source includes open data sources; the data processing unit includes: industry-related data processing unit, used to extract industries related to the one or more specific fields Before relevant entities and corresponding entity attributes and / or entity relationships, the data cleaning and extraction-transformation-loading (ETL) processing is performed on the data related to the industry in the one or more specific fields; and The database construction unit includes: a database supplement / update unit for supplementing or updating the industry knowledge graph database based on the extracted entities, entity attributes and / or entity relationships. The system according to item 27 of the scope of patent application, wherein the data acquisition unit includes: a media data acquisition unit, which is used to obtain and obtain information from Internet data sources using an application program interface (API) or a web crawler technology. Said one or more industry-related Internet media materials in a specific field; the data processing unit includes: a media data processing unit for performing event detection, event evaluation, and screening on the Internet media materials to extract information related to the industry Related specific media events, and identifying corresponding directly related entities from the internet media materials; and the database building unit includes: a database supplement / update unit for based on the specific media events and corresponding direct correlations An entity to supplement the industry knowledge map database, wherein the specific media event is added to the industry knowledge map database as an abstract entity. The system according to any one of claims 29 to 31, wherein the database supplement / update unit is further configured to perform semantic disambiguation and entity connection on the extracted entities. The system according to item 27 of the scope of patent application, wherein the media event monitoring unit is further configured to: classify content in the obtained Internet media materials to obtain content targeted to a specific topic; Identify the entities involved in the obtained content; perform sentiment analysis on the obtained content and the identified entities, and filter the obtained content based on the results of the sentiment analysis; and perform event discovery based on the filtered content to media Cluster events and discover new media events. The system according to item 33 of the scope of patent application, wherein the media event monitoring unit is further configured to analyze the authenticity of events based on the attributes of the media events, and sort and / or media events according to the analysis results. Or filtered. The system according to item 27 of the scope of patent application, wherein the database access unit is further configured to: based on the directly related entities, preset other conditions and other entities in the industry knowledge map database and And / or an entity relationship query to determine the non-directly related entity. The system according to item 27 of the scope of patent application, wherein the database access unit is further configured to: based on the directly related entity, perform an analysis on other entities and / or entity relationships in the industry knowledge map database Data acquisition techniques are used to identify the non-directly related entities. The system according to item 27 of the scope of the patent application, wherein the specific media event includes a negative event, an emergency event, a crisis event, a group event, a public opinion event, or other events of industrial significance.