Public opinion processing method and system for electric power service system based on knowledge graph

Abstract

The present invention belongs to the field of electric power service systems, and specifically relates to a method and system for processing public opinion in an electric power service system based on a knowledge graph. The method for processing public opinion in an electric power service system based on a knowledge graph provided by the present invention includes the following steps: S1: collecting relevant information on public opinion, impact and processing strategies of the electric power service system; S2: analyzing public opinion characteristics, standardizing related concepts, defining and extracting public opinion types, attributes and associations; S3: using a distributed storage method to store the constructed network public opinion knowledge graph; S4: accessing the public opinion guidance request of the electric power service system; S5: querying the network public opinion knowledge graph of the distributed electric power service system. The present invention solidifies the public opinion characteristics of the electric power service system in the form of a knowledge graph, cooperates with the public opinion processing strategy, realizes rapid and accurate public opinion control, and achieves the purpose of effective prevention before the impact of public opinion, and timely response and loss prevention after the occurrence of public opinion.

Description

Technical Field

[0001] The present invention belongs to the field of electric power service systems, and specifically relates to a method and system for processing public opinion in electric power service systems based on knowledge graphs. Background Art

[0002] With the vigorous development of the self-media industry, it is difficult to distinguish the truth from the false on the Internet, and the recipients of information are easily guided by the news to produce various emotions, which will affect public opinion. Self-media has greatly reduced the cost of making news and accelerated the speed of news dissemination. For various emerging public opinions, especially for industries related to the power service system, it affects individuals and enterprises at a small level, and affects people's livelihood and the economy at a large level, but there is currently no better way to deal with it. Summary of the invention

[0003] The knowledge graph-based public opinion processing method and system for electric power service systems provided by the present invention can effectively solve the problems of background technologies.

[0004] The method for processing public opinion of a power service system based on a knowledge graph provided by the present invention comprises the following steps:

[0005] S1: Collect relevant information on public opinion, impact and handling strategies of the power service system;

[0006] S2: Analyze the characteristics of public opinion, standardize related concepts, define and extract the types, attributes and relationships of public opinion, including:

[0007] Define concepts; clarify the types and quantities of concepts from three aspects: public opinion topology, public opinion response, and public opinion monitoring; public opinion topology includes three types: institutional public opinion, corporate public opinion, and personal public opinion; public opinion response includes four types: understanding the background, analyzing trends, identifying key information, and formulating strategies; public opinion monitoring includes platform monitoring and manual monitoring;

[0008] Define the relationship; the internal logic of public opinion information in the power service system between public opinion topology, public opinion response, and public opinion monitoring;

[0009] Define attributes; describe the connotation of the concept;

[0010] Establish a public opinion topology sub-network, extract concept types from institutional public opinion, corporate public opinion, and personal public opinion as nodes and relationships of public opinion topology; the public opinion topology sub-network is described as G tp = {V tp +E tp +W tp}, where V tp is the node set of the public opinion topology sub-network, V tp = {V 11 ,V 12 ,V 13}, where V 11 is the public opinion type node of the entire power system, V 12 is the refined type node, V 13 is a further refined type node, E tp is the relationship set of the public opinion topology sub-network, W tp It is the relationship attribute description of the public opinion topology sub-network;

[0011] Establish a public opinion response sub-network, extract concept types from understanding the background, analyzing trends, identifying key information, and formulating strategies as nodes and relationships for public opinion response; the public opinion response sub-network is described as G yd = {V yd +E yd +W yd}, where V yd is the node set of the public opinion response sub-network, V yd = {V 21 ,V 22 ,V 23 ,V 24}, where V 21 is the public opinion type node to be resolved, V 22 is the node of public opinion impact to be resolved, V 23 is the node of public opinion analysis to be solved, V 24 is the node of the public opinion response strategy to be solved, E yd is the relationship set of the public opinion response sub-network, W yd It is the relationship attribute description of the public opinion response sub-network;

[0012] Establish a public opinion monitoring sub-network, extract concept types from platform monitoring and manual monitoring as nodes and relationships for public opinion monitoring; the public opinion monitoring sub-network is described as G jc = {V jc +E jc +W jc}, where V jc is the node set of the public opinion monitoring sub-network, V jc = {V 31 ,V 32}, where V 31 is the set of monitoring method nodes, V 32 is the node set of the monitoring object, E jc is the relationship set of the public opinion monitoring sub-network, W jc It is the relationship attribute description of the public opinion monitoring sub-network;

[0013] Merge each sub-network to build a network public opinion knowledge graph, which is described as G = {V + E + W}, where V = V tp ∪V yd ∪V jc ; E=E tp∪E yd ∪E jc ; W = W tp ∪W yd ∪W jc ;

[0014] S3: uses a distributed storage method to store the constructed network public opinion knowledge graph;

[0015] S4: Access to the public opinion guidance request of the power service system;

[0016] S5: Query the network public opinion knowledge graph of the distributed power service system.

[0017] As a further optimization of the present invention, step S5 includes the following steps:

[0018] A1: The process of representing the network public opinion knowledge graph is represented by an undirected graph, and the subgraphs of the knowledge graph and query are defined as G = {V + E + W}, Q = {V q +E q +W q};

[0019] A2: Introducing the domain vector P Q (q n )=[γ Q (q i ,q1),γ Q (q i ,q2),…,γ Q (q i ,q n )], where q∈V, n is the number of nodes in Q, γ Q The calculation method is as follows:

[0020]

[0021] Among them, α and β are nodes of the network public opinion knowledge graph, l α,β is the shortest path length between two nodes, m α,β is the number of shortest paths, λ is a regulation factor ranging from 0 to 1;

[0022] A3: Query nodes, match the similarity between node domain vectors, and obtain the optimal matching point:

[0023] c(f,q i )=∑H{λ Q (q i ,q j ),λ G [f(q i ),f(q j )]}, where f is an embedding of the graph and H is defined as follows:

[0024]

[0025] The present invention also provides a public opinion processing system for a power service system based on a knowledge graph, which is characterized by comprising:

[0026] Public opinion information collection module, which collects information related to public opinion, impact and handling strategies of the power service system;

[0027] The knowledge graph management module establishes the network public opinion knowledge graph in step S2;

[0028] Storage module, distributed storage network public opinion knowledge graph;

[0029] The request module includes a client and a receiving end; the client sends a public opinion guidance request; the receiving end receives the public opinion guidance request sent by the client;

[0030] The query module receives the public opinion guidance request from the receiving end, matches the optimal result in the network public opinion knowledge graph in the storage module and outputs it.

[0031] As a further optimization of the present invention, several storage servers are used to save the original data of the network public opinion knowledge graph; the client communicates with the storage server through an API interface.

[0032] As a further optimization of the present invention, several tracking servers are also included to monitor the communication ports between the client and the storage server.

[0033] As a further optimization of the present invention, the public opinion information collection module collects network public opinion information, document public opinion information and oral public opinion information.

[0034] As a further optimization of the present invention, collecting online public opinion information includes crawling the original page of public opinion data according to the HTML source code of the web page, crawling parameters, depth and related rules according to the URL of the original page, and obtaining the source code of the original page.

[0035] As a further optimization of the present invention, it is characterized in that it also includes a preprocessing module; the preprocessing module parses the source code through JSoup, obtains the DOM object, extracts the public opinion content, and performs word segmentation and denoising on the public opinion content.

[0036] The present invention also provides an electronic device, comprising the electric power service system public opinion processing system based on knowledge graph.

[0037] The present invention also provides a computer-readable storage medium that stores executable instructions. When the executable instructions are run on a computer, the computer executes the method for processing public opinion in an electric power service system based on a knowledge graph.

[0038] The present invention solidifies the public opinion characteristics of the power service system in the form of a knowledge graph, and cooperates with the public opinion processing strategy to achieve rapid and accurate public opinion control, so as to achieve the purpose of effective prevention before the impact of public opinion, and timely response and loss prevention after the occurrence of public opinion.

[0039] Instruction Manual

[0040] Figure 1 This is a partial schematic diagram of the logical relationship of merging sub-networks to construct a network public opinion knowledge graph in this embodiment. DETAILED DESCRIPTION

[0041] The method for processing public opinion of a power service system based on a knowledge graph provided by the present invention comprises the following steps:

[0042] S1: Collect relevant information on public opinion, public opinion impact and public opinion handling strategies related to the power service system. In view of the rapid development of Internet information, public opinion collection is mainly based on network collection, supplemented by various documents and word of mouth.

[0043] S2: Based on the collected data related to public opinion in the power system, analyze the characteristics of public opinion, clarify and standardize related concepts, define and extract the types, attributes and correlations of public opinion, and prepare for the establishment of a network public opinion knowledge graph.

[0044] The definition of concepts mainly clarifies the types and quantities of concepts from three aspects: public opinion topology, public opinion response, and public opinion monitoring. Public opinion topology refers to the topological relationship of various types of public opinion. The public opinion topology in this embodiment establishes three types of institutional public opinion, corporate public opinion, and personal public opinion, and considers the vertical and horizontal relationships between the three categories.

[0045] Institutional public opinion can be subdivided into three parts: regional public opinion, economic public opinion, and people's livelihood public opinion. For example, regional public opinion can be divided into international public opinion, national public opinion, local public opinion, etc.; economic public opinion can be subdivided into economic public opinion, financial public opinion, etc.; people's livelihood public opinion can be divided into cultural public opinion, environmental public opinion, etc. Specifically, the detailed categories and correlations can be clarified based on the actual degree of refinement.

[0046] Corporate public opinion can be subdivided into three parts: product public opinion, service public opinion, and management public opinion. For example, product public opinion can be further divided into quality public opinion, price public opinion, etc.; service public opinion can be subdivided into positive evaluation public opinion, neutral evaluation public opinion, and negative evaluation public opinion; management public opinion can be divided into orderly public opinion, disorderly public opinion, etc.

[0047] Personal public opinion can be subdivided into three parts: subjective public opinion, objective public opinion and malicious public opinion. For example, subjective public opinion can be divided into strong public opinion and non-strong public opinion; objective public opinion can be divided into consumer public opinion and bystander public opinion; malicious public opinion can be divided into opponent public opinion and follow-up public opinion.

[0048] Public opinion response is classified according to the actual response process, including four types: understanding the background, analyzing trends, identifying key information, and formulating strategies.

[0049] Public opinion monitoring is based on specific monitoring methods, including platform monitoring and manual monitoring.

[0050] The relationship is defined by first clarifying the public opinion topology of the public opinion information of the power service system, conducting targeted public opinion monitoring based on the topological structure, and clarifying detailed public opinion response strategies based on the monitoring situation and specific topological nodes. This is the internal logic between the description of public opinion topology, public opinion response, and public opinion monitoring specifications, such as Figure 1 shown.

[0051] The defined attributes describe the original information of the public opinion type, which is the basic parameter that distinguishes one type from another. In the modeling of the graph, the attribute definition is mainly based on the connotation description of the type and does not exist in the form of network nodes.

[0052] Afterwards, the public opinion topology sub-network is established, and the concept types are extracted from institutional public opinion, corporate public opinion, and personal public opinion as nodes and relationships of the public opinion topology; the public opinion topology sub-network is described as G tp = {V tp +E tp +W tp}, where V tp is the node set of the public opinion topology sub-network, E tp is the relationship set of the public opinion topology sub-network, W tp is the relational attribute description of the public opinion topology sub-network; in addition, V tp = {V 11 ,V 12 ,V 13}, where V 11 is the public opinion type node of the entire power system, V 12 It is a detailed type node, such as institutional public opinion, economic public opinion, and people's livelihood public opinion. 13 It is a further refined type node, such as other local public opinions, cultural public opinions, etc.

[0053] Establish a public opinion response sub-network, extract concept types from understanding the background, analyzing trends, identifying key information, and formulating strategies as nodes and relationships for public opinion response; the public opinion response sub-network is described as G yd = {V yd +Eyd +W yd}, where V yd is the node set of the public opinion response sub-network, E yd is the relationship set of the public opinion response sub-network, W yd is the relational attribute description of the public opinion response sub-network; in addition, V yd = {V 21 ,V 22 ,V 23 ,V 24}, where V 21 is the public opinion type node to be resolved, V 22 is the node of public opinion impact to be resolved, V 23 is the node of public opinion analysis to be solved, V 24 It is the node of public opinion response strategy to be resolved.

[0054] Establish a public opinion monitoring sub-network, extract concept types from platform monitoring and manual monitoring as nodes and relationships for public opinion monitoring; the public opinion monitoring sub-network is described as G jc = {V jc +E jc +W jc}, where V jc is the node set of the public opinion monitoring sub-network, E jc is the relationship set of the public opinion monitoring sub-network, W jc is the relational attribute description of the public opinion monitoring sub-network; in addition, V jc = {V 31 ,V 32}, where V 31 is the set of monitoring method nodes, V 32 It is used to describe the set of nodes of the monitored object.

[0055] Merge each sub-network to build a network public opinion knowledge graph, which is described as G = {V + E + W}, where V = V tp ∪V yd ∪V jc ; E=E tp ∪E yd ∪E jc ; W = W tp ∪W yd ∪W jc .

[0056] S3: Adopt the network public opinion knowledge graph of the distributed power service system. Specifically, adopt a distributed storage method to store the network public opinion knowledge graph.

[0057] S4: During the application process, when there is a query demand, access the public opinion guidance request of the power service system. Specifically, the following steps are adopted:

[0058] A1: The process of representing the network public opinion knowledge graph is represented by an undirected graph, and the subgraphs of the knowledge graph and query are defined as G = {V + E + W}, Q = {V q +E q +W q};

[0059] A2: Introducing the domain vector P Q (q n )=[γ Q (q i ,q1),γ Q (q i ,q2),…,γ Q (q i ,q n )], where q∈V, n is the number of nodes in Q, γ Q The calculation method is as follows:

[0060]

[0061] Among them, α and β are nodes of the network public opinion knowledge graph, l α,β is the shortest path length between two nodes, m α,β is the number of shortest paths, λ is a regulation factor ranging from 0 to 1;

[0062] A3: Query nodes, match the similarity between node domain vectors, and obtain the optimal matching point:

[0063] c(f,q i )=∑H{λ Q (q i ,q j ),λ G [f(q i ),f(q j )]}, where f is an embedding of the graph and H is defined as follows:

[0064]

[0065] S5: Finally, query the network public opinion knowledge graph of the distributed power service system public opinion and output the query results.

[0066] This embodiment also provides a public opinion processing system for a power service system based on a knowledge graph, including a public opinion information collection module, a knowledge graph management module, a request module and a query module.

[0067] Among them, the public opinion information collection module collects relevant information on public opinion, impact and processing strategies of the power service system. The public opinion information collection module can collect network public opinion information, literature public opinion information and oral public opinion information. In this embodiment, the public opinion information collection module mainly collects network public opinion information. Specifically, the original page of the public opinion data is captured according to the HTML source code of the web page, and the crawling parameters, depth and related rules are performed according to the URL of the original page to obtain the source code of the original page.

[0068] The web page source code obtained by the public opinion information collection module usually contains a lot of worthless content, which needs to be processed. Therefore, this embodiment also provides a preprocessing module. The preprocessing module parses the source code through JSoup to obtain the DOM object, and processes the tags therein, and uses relevant algorithms to extract the text of the network public opinion information of the power system service, obtain important information such as content reports, time, and title, and finally extract the public opinion content from these information contents, and perform word segmentation and denoising on the content. The word segmentation and denoising processing can specifically use MapReduce, word segmentation tools and introduce user dictionaries.

[0069] The knowledge graph management module establishes the network public opinion knowledge graph in step S2 of the above-mentioned knowledge graph-based public opinion processing method for the electric power service system.

[0070] The storage module stores the network public opinion knowledge graph in a distributed manner. Specifically, several storage servers can be used to store the original data of the network public opinion knowledge graph; the client communicates with the storage server through an API interface. In addition, this embodiment also has several tracking servers that monitor the communication port between the client and the storage server.

[0071] The request module includes a client and a receiving end; the client sends a public opinion guidance request; the receiving end receives the public opinion guidance request sent by the client.

[0072] The query module receives the public opinion guidance request from the receiving end, matches the optimal result in the network public opinion knowledge graph in the storage module and outputs it.

[0073] This embodiment also provides an electronic device, including the above-mentioned knowledge graph-based public opinion processing system for the electric service system.

[0074] This embodiment also provides a computer-readable storage medium that stores executable instructions. When the executable instructions are run on a computer, the computer executes the method for processing public opinion in a power service system based on a knowledge graph.

[0075] By adopting the method and system of this embodiment, when facing public opinion, response measures can be quickly and effectively queried and actively responded. In addition, since the power service system is aimed at people including others, enterprises and individuals, the network public opinion knowledge map constructed for the power service system in this embodiment is highly targeted and reasonable. Strategic planning is carried out based on the collected relevant information, based on practice, to reduce the risk of improper response.

[0076] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention, rather than to limit the scope of protection of the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the essence and scope of the technical solution of the present invention.

Claims

1. A method for processing public opinion in a power service system based on knowledge graph, characterized in that: The following steps are included: S1: Collect relevant information on public opinion, impact and handling strategies of the power service system; S2: Analyze the characteristics of public opinion, standardize related concepts, define and extract the types, attributes and relationships of public opinion, including: Define concepts; clarify the types and quantities of concepts from three aspects: public opinion topology, public opinion response, and public opinion monitoring; public opinion topology includes three types: institutional public opinion, corporate public opinion, and personal public opinion; public opinion response includes four types: understanding the background, analyzing trends, identifying key information, and formulating strategies; public opinion monitoring includes platform monitoring and manual monitoring; Define the relationship; the internal logic of public opinion information in the power service system between public opinion topology, public opinion response, and public opinion monitoring; Define attributes; describe the connotation of concepts; describe the original information of public opinion types, which does not exist in the form of network nodes; Establish a public opinion topology sub-network, extract concept types from institutional public opinion, corporate public opinion, and personal public opinion as nodes and relationships of public opinion topology; the public opinion topology sub-network is described as G tp = {V tp +E tp +W tp }, where V tp is the node set of the public opinion topology sub-network, V tp = {V 11 ,V 12 ,V 13 }, where V 11 is the public opinion type node of the entire power service system, V 12 is the refined type node, V 13 is a further refined type node, E tp is the relationship set of the public opinion topology sub-network, W tp It is the relationship attribute description of the public opinion topology sub-network; Establish a public opinion response sub-network, extract concept types from understanding the background, analyzing trends, identifying key information, and formulating strategies as nodes and relationships for public opinion response; the public opinion response sub-network is described as G yd = {V yd +E yd +W yd }, where V yd is the node set of the public opinion response sub-network, V yd = {V 21 ,V 22 ,V 23 ,V 24 }, where V 21 is the public opinion type node to be resolved, V 22 is the node of public opinion impact to be resolved, V 23 is the node of public opinion analysis to be solved, V 24 is the node of the public opinion response strategy to be solved, E yd is the relationship set of the public opinion response sub-network, W yd It is the relationship attribute description of the public opinion response sub-network; Establish a public opinion monitoring sub-network, extract concept types from platform monitoring and manual monitoring as nodes and relationships for public opinion monitoring; the public opinion monitoring sub-network is described as G jc = {V jc +E jc +W jc }, where V jc is the node set of the public opinion monitoring sub-network, V jc = {V 31 ,V 32 }, where V 31 is the set of monitoring method nodes, V 32 is the node set of the monitoring object, E jc is the relationship set of the public opinion monitoring sub-network, W jc It is the relationship attribute description of the public opinion monitoring sub-network; Merge each sub-network to build a network public opinion knowledge graph, which is described as G = {V + E + W}, where V = V tp ∪V yd ∪V jc ; E=E tp ∪E yd ∪E jc ; W = W tp ∪W yd ∪W jc ; S3: uses a distributed storage method to store the constructed network public opinion knowledge graph; S4: Access to the public opinion guidance request of the power service system; S5: Query the network public opinion knowledge graph of the power service system; Step S5 includes the following steps: A1: The process of representing the network public opinion knowledge graph with an undirected graph is defined as G = {V + E + W}, Q = {V q +E q +W q }; A2: Introducing the domain vector P Q (q n )=[γ Q (q i ,q1),γ Q (q i ,q2),…,γ Q (q i ,q n )], where q∈V, n is the number of nodes in Q, γ Q The calculation method is as follows: Among them, α and β are nodes of the network public opinion knowledge graph, l α,β is the shortest path length between two nodes, m α,β is the number of shortest paths, λ is a regulation factor ranging from 0 to 1; A3: Query nodes, match the similarity between node domain vectors, and obtain the optimal matching point: c(f,q i )=∑H{λ Q (q i ,q j ),λ G [f(q i ),f(q j )]}, where f is an embedding of the graph and H is defined as follows: 。 2. The public opinion processing system of the power service system based on knowledge graph is characterized by: include: Public opinion information collection module, which collects information related to public opinion, impact and handling strategies of the power service system; A knowledge graph management module, which establishes the network public opinion knowledge graph in step S2 of the public opinion processing method for the electric power service system based on the knowledge graph as claimed in claim 1; Storage module, distributed storage network public opinion knowledge graph; The request module includes a client and a receiving end; the client sends a public opinion guidance request; The receiving end receives the public opinion guidance request sent by the client; The query module receives the public opinion guidance request from the receiving end, matches the optimal result in the network public opinion knowledge graph in the storage module and outputs it.

3. The public opinion processing system for electric service system based on knowledge graph according to claim 2 is characterized in that: Use several storage servers to save the original data of the online public opinion knowledge graph; the client communicates with the storage server through the API interface.

4. The public opinion processing system for electric service system based on knowledge graph according to claim 2 is characterized in that: It also includes several tracking servers that monitor the communication ports between the client and the storage server.

5. The public opinion processing system for electric service system based on knowledge graph according to claim 2 is characterized in that: The public opinion information collection module collects public opinion information including network public opinion information, literature public opinion information and oral public opinion information.

6. The public opinion processing system for electric power service system based on knowledge graph according to claim 5 is characterized in that: Collect online public opinion information, including crawling the original page of public opinion data based on the HTML source code of the web page, crawling parameters, depth and related rules based on the URL of the original page, and obtaining the source code of the original page.

7. The public opinion processing system for electric service system based on knowledge graph according to claim 6 is characterized in that: It also includes a preprocessing module; the preprocessing module parses the source code through JSoup, obtains the DOM object, extracts the public opinion content, and performs word segmentation and denoising on the public opinion content.

8. An electronic device, characterized in that: Including the public opinion processing system for power service system based on knowledge graph as described in any one of claims 2-7.

9. A computer-readable storage medium, characterized in that: Executable instructions are stored. When the executable instructions are run on a computer, the computer executes the method for processing public opinion in a power service system based on a knowledge graph as described in claim 1.

Images