Knowledge graph construction method and system based on power production statistics business

Abstract

The present invention proposes a knowledge graph construction method and system based on power production statistical business, which sorts out the vertical aggregation and horizontal correlation of power production statistical indicators; constructs a linkage model of statistical indicators-detailed data-related data-related objects based on the relationship between statistical indicators; uses ontology design software to realize the graph drawing of the linkage model; and constructs intelligent analysis and intelligent answer application scenarios based on the knowledge graph. In addition to considering the statistics and aggregation of traditional power production indicators, the present invention further considers the vertical aggregation relationship of the province-city-county level, the horizontal correlation between statistical indicators at all levels, and the intelligent analysis and intelligent question and answer application scenarios, so as to achieve higher quality statistical analysis services and further promote the digitization, visualization and intelligence of statistical business.

Description

Technical Field

[0001] The present invention belongs to the technical fields of knowledge graphs, knowledge base questions and answers, and power production statistics, and in particular, relates to a knowledge graph construction method and system based on power production statistics business. Background Art

[0002] The knowledge graph was proposed by Google on May 17, 2012. It is dedicated to describing massive entities, entity attributes and entity relationships, and establishing an intelligent information retrieval model. In the knowledge graph, the basic unit of the relationship between entities is the "entity-relationship-entity" triple, that is, G = (E, R, S), where E represents the set of entities in the knowledge base E = {e1, e2, ..., en}, R represents the set of relationships between entities in the knowledge base R = {r1, r2, ..., rm}, S represents the knowledge base triple set S ⊆ E × R × E, and the attribute characteristics of the entity are represented by "attribute-value". Entity is the basic element of the knowledge graph, relationship is the relationship between different entities, attribute is the description of the entity, and value is the specific value of the entity attribute.

[0003] The full English name of knowledge base question answering is: knowledge base question answering, KBQA, given a natural language question, through semantic understanding and analysis of the question, and then use the knowledge base to query and reason to get the answer. Summary of the invention

[0004] At present, the focus of power grid statistics has gradually shifted from traditional statistics and data aggregation to analysis and research data, laying the foundation for providing efficient and high-quality statistical analysis services for future multi-business and multi-field related interactions. Knowledge graphs describe concepts, entities and their relationships in the objective world in a structured form, express information in a form closer to the human cognitive world, and provide a better ability to organize, manage and understand massive amounts of information. In combination with production statistics business, in-depth research on knowledge graph technology, and the construction of power production statistics knowledge graphs will give full play to the supervisory support role of statistics for production and operation business, further promote the digitization, visualization and intelligence of the company's statistical business, and lay the foundation for the improvement of intelligent applications of online power grid systems.

[0005] To this end, the present invention proposes a knowledge graph construction method and system based on power production statistical business, including: combing the vertical aggregation and horizontal correlation of power production statistical indicators; constructing a linkage model of statistical indicators-detailed data-related data-related objects based on the relationship between statistical indicators; using ontology design software to realize the graph drawing of the linkage model; and constructing intelligent analysis and intelligent answer application scenarios based on the knowledge graph. In addition to considering the statistics and aggregation of traditional power production indicators, the present invention further considers the vertical aggregation relationship of the province-city-county level, the horizontal correlation between statistical indicators at all levels, and intelligent analysis and intelligent question and answer application scenarios, so as to achieve higher quality statistical analysis services and further promote the digitization, visualization and intelligence of statistical business.

[0006] Its purpose is to provide a method, steps and evolution path for constructing a knowledge graph for power production statistics, and to formulate a plan for constructing a knowledge graph for production statistics, which includes building an indicator linkage model, knowledge graph drawing, intelligent search and other scenario applications, further promoting the modernization of the company's statistical work, and laying the foundation for improving the intelligence level of future online power grid systems.

[0007] The specific technical solutions are as follows:

[0008] A method for constructing a knowledge graph based on power production statistics business, characterized by comprising the following steps:

[0009] Step S1: sort out the vertical aggregation and horizontal correlation of power production statistical indicators;

[0010] Step S2: constructing a linkage model of statistical indicators-detailed data-related data-related objects based on the relationship between statistical indicators;

[0011] Step S3: Draw a graph of the linkage model using ontology design software;

[0012] Step S4: Construct application scenarios including intelligent analysis and / or intelligent answers based on the knowledge graph.

[0013] Furthermore, in step S1, electricity production statistics are classified by content to include power generation statistics, power supply statistics, power consumption statistics, power balance, load statistics, equipment statistics and energy consumption statistics; service objects include internal company statistics, as well as external statistics; statistical calibers include companies, power grids, regions, rural power grids, different voltage levels, and different industries; statistical cycles include monthly, quarterly, and annual; the relationship between power production statistical indicators is divided into three vertical levels of province-city-county, as well as horizontal correlations between each level.

[0014] Furthermore, in step S2, the model layer of the knowledge graph is constructed through ontology learning, including four steps: term extraction, concept extraction, relationship extraction and rule learning to extract and integrate production statistics related entities, namely: report, indicator, enterprise, power plant, checkpoint, administrative division, industry, month, year and report theme.

[0015] Furthermore, in step S3, the data layer of the knowledge graph of power production statistics samples is constructed, and on the basis of ontology learning, entity and relationship information in the data is extracted; the ontology design system is used to realize the graph drawing of the linkage model; and the graph drawing of the statistical indicator-detailed data linkage model is realized through the platform ontology modeling, knowledge extraction, knowledge fusion, knowledge storage, and graph drawing functions.

[0016] Furthermore, in step S4, the constructed graph is applied to three types, including intelligent search, graph algorithm research for auxiliary decision-making, and intelligent question and answer.

[0017] Furthermore, a top-down construction method is adopted to construct the knowledge graph of power production statistics: first construct the pattern layer of the knowledge graph, and then construct the data layer; that is, first construct the ontology and its relationships in the pattern layer according to the existing structured data or expert knowledge base to form the corresponding conceptual model and rule relationship, and then extract entities from the data according to this pattern to construct the data layer.

[0018] The model layer of the knowledge graph is constructed through ontology learning, including: term extraction, concept extraction, relationship extraction and rule learning. Four steps are used to integrate 10 entities to form the model layer of the knowledge graph of power production statistics samples, forming the final ontology model, namely the production statistics indicator linkage model;

[0019] The data layer of the power production statistical sample knowledge graph is constructed through two steps of entity linking and entity filling in entity learning.

[0020] And, a knowledge graph system based on power production statistics business is established on the basis of the above knowledge graph construction method based on power production statistics business, including three application modules: intelligent search, graph algorithm research for auxiliary decision-making, and intelligent question and answer.

[0021] Furthermore, the intelligent search module designs graph query conditions and conducts further analysis and research based on common problems in production statistics work, that is, explores graph search application scenarios of various types of entities, relationships, events and attributes in production statistics business, including: administrative region and unit affiliation query sub-module, equipment status query sub-module, industrial electricity consumption analysis sub-module, clean energy status analysis sub-module and industry-specific electricity consumption analysis sub-module.

[0022] Furthermore, the graph algorithm research module for auxiliary decision making is used to quickly retrieve all paths between two indicators.

[0023] Furthermore, the intelligent question and answer module is used to collect commonly used questions related to production statistics, perform technical processing on commonly used questions related to production statistics, create a knowledge base to answer questions about the current value and cumulative value of electricity production statistics, and answer questions about the year-on-year growth rate and size of the increase in electricity production statistics through the knowledge base.

[0024] And, an electronic device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that when the processor executes the program, the steps of the method for constructing a knowledge graph based on power production statistics business as described above are implemented.

[0025] And, a non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps of the method for constructing a knowledge graph based on power production statistics business as described above are implemented.

[0026] The present invention and its preferred embodiment start from production indicators and provide a specific implementation plan for constructing intelligent analysis and intelligent answer application scenarios using knowledge graphs, including methods, steps and evolution routes, and formulate a production statistics knowledge graph construction plan. The plan includes building an indicator linkage model, knowledge graph drawing, intelligent search and other scenario applications, which can give full play to the supervisory and supporting role of statistics on production and operation businesses, further promote the digitization, visualization and intelligence of the company's statistical business, and lay the foundation for improving the intelligent application of online power grid systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0028] Figure 1 is an overall architecture diagram of an embodiment of the present invention;

[0029] Figure 2 is a flow chart of knowledge base question and answer processing in an embodiment of the present invention;

[0030] Figure 3 is a production statistical indicator linkage model of an embodiment of the present invention;

[0031] Figure 4 It is a flowchart of constructing a statistical knowledge graph of power production using a top-down method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] In order to make the features and advantages of this patent more obvious and easy to understand, the following embodiments are specifically described in detail as follows:

[0033] like Figure 1 As shown, the specific steps of the method for constructing the power production statistical knowledge graph provided in this embodiment are as follows:

[0034] Step 1: Sort out the vertical aggregation and horizontal correlation of power production statistical indicators;

[0035] Step 2: Construct a linkage model of statistical indicators-detailed data-related data-related objects based on the relationship between statistical indicators;

[0036] Step 3: Use ontology design software to draw a graph of the linkage model;

[0037] Step 4: Build intelligent analysis and intelligent answer application scenarios based on the knowledge graph.

[0038] In step 1, the electricity production statistics are classified into seven aspects, including power generation statistics, power supply statistics, power consumption statistics, power balance, load statistics, equipment statistics and energy consumption statistics. The service objects include internal company statistics and external statistics (government statistics, industry statistics). The statistical caliber includes companies, power grids, regions, rural power grids; different voltage levels; different industries, etc.; statistical cycles include monthly, quarterly, and annual. The relationship between power production statistical indicators is divided into three vertical levels: province-city-county, and horizontal associations between levels.

[0039] In step 2, the model layer of the knowledge graph is constructed through "ontology learning", which mainly includes four steps: term extraction, concept extraction, relationship extraction, and rule learning. Ten production statistics-related entities are extracted and integrated, namely "report", "indicator", "enterprise", "power plant", "gateway", "administrative division", "industry", "month", "year" and "report subject".

[0040] In step three, the data layer of the knowledge graph of power production statistics samples is constructed, and the entity and relationship information in the data is extracted based on ontology learning. The graph drawing of the linkage model is realized by using the ontology design system; the graph drawing of the statistical indicator-detailed data linkage model is realized through the platform ontology modeling, knowledge extraction, knowledge fusion, knowledge storage, and graph drawing functions;

[0041] In step four, the constructed graph is applied, including three types of applications: intelligent search, graph algorithm research for auxiliary decision-making, and intelligent question and answer. This improves user work efficiency and helps new business personnel quickly understand business knowledge.

[0042] The embodiment of the present invention adopts a top-down construction method to construct a knowledge graph of power production statistics. It includes using a top-down method to first construct the model layer of the knowledge graph, and then construct the data layer. In the top-down construction, first construct the ontology and its relationship in the model layer according to the existing structured data or expert knowledge base to form a corresponding conceptual model and rule relationship, and then extract entities from the data according to this model to construct the data layer. It can be seen that the top-down construction method is a construction process from abstract to concrete, with concepts first and then concrete implementations.

[0043] The model layer of the knowledge graph is constructed through "ontology learning", which mainly includes four steps: term extraction, concept extraction, relationship extraction, and rule learning. Ten entities are integrated to form the model layer of the knowledge graph of the power production statistics sample, forming the final ontology model, namely the "production statistics indicator linkage model".

[0044] The data layer of the power production statistical sample knowledge graph is constructed through the two steps of entity linking and entity filling of "entity learning".

[0045] Specifically, Figure 1-Figure 4 As shown, the construction process includes:

[0046] Step 1: Ontology Learning

[0047] The process of building a linkage model for production statistics indicators is the process of building the pattern layer of the production statistics knowledge graph, also known as the ontology construction process. It is necessary to build the pattern layer of the knowledge graph through "ontology learning", which mainly includes four steps: term extraction, concept extraction, relationship extraction, and rule learning. Based on the "Production Statistics Report System", "Equipment Statistics Report System" and "Energy Consumption Statistics Report System", the four steps of "ontology learning" - term extraction, concept extraction, relationship extraction, and rule learning - are used to build a linkage model for production statistics indicators.

[0048] Ontology Learning Step 1: Term Extraction

[0049] Terminology is a fixed representation of entities and concepts in the knowledge graph in the field of power systems. The goal of term extraction is to obtain a set of professional tags used to represent the ontology in power production statistics. Term extraction is performed on the "Production Statistics Report System". Observe the catalog page of the "Production Statistics Report System" and extract the specific names of each report such as "Direct Report One - Power Supply Comprehensive Situation Table (Company Caliber - Municipal Power Supply Enterprise)", "Direct Report Two - Electricity Exchange Situation Table (Grid Caliber)", "Direct Report Three - Grid Load Table", "Direct Report Four - Power Sales Detail Table (Company Caliber)" as terms.

[0050] Ontology Learning Step 2: Concept Extraction

[0051] On the basis of term extraction, the concepts in the "Production Statistical Report System" are further extracted. Concepts are more abstract, representative and general terms that can represent all entities within a certain category. For example, the specific names in the catalogs such as "Direct Report 1--Comprehensive Power Supply Situation Table (Company Caliber-Municipal Power Supply Enterprise)", "Direct Report 2--Electricity Exchange Situation Table (Grid Caliber)", and "Direct Report 3--Grid Load Table" are extracted as the concept of "report".

[0052] Ontology Learning Step 3: Relation Extraction

[0053] Through the above-mentioned term extraction and concept extraction, 10 entities were integrated, namely "report", "indicator", "enterprise", "power plant", "power supply gateway", "administrative division", "industry", "month", "year" and "report theme". Various relationships between entities such as subordinate relationship, equal relationship, calculation relationship, inclusion relationship and calculation relationship were integrated.

[0054] Ontology Learning Step 4: Rule Learning

[0055] Rule learning refers to the process of learning common sentences or templates containing entities and relationships during ontology extraction, that is, a process of continuous refinement over and over again. The four steps of term extraction, concept extraction, relationship extraction and rule learning in ontology learning constitute the model layer of the knowledge graph of power production statistics samples, forming the final ontology model, namely the "production statistics indicator linkage model".

[0056] Step 2: Physical Learning

[0057] Entity learning constructs the data layer of the knowledge graph of power production statistics samples, and extracts entity and relationship information from the data based on ontology learning. These extracted entities can form structural relationships between entities according to the relationship hierarchy of the pattern layer based on their corresponding concepts. The main key technologies of entity learning include entity linking and entity filling.

[0058] Entity Learning Step 1: Entity Linking

[0059] Entity linking is the operation of linking the entity object extracted from the text to the corresponding correct entity object in the knowledge base. In the process of entity linking, in order to ensure the uniqueness of each entity, a "unique identifier" is used to represent this entity. Taking the entity "indicator" as an example, the data processing of entity linking is explained in detail. Taking the "power supply" scenario as an example, the value of the entity "indicator" involves two dimensions in the report. The horizontal dimension reflects the statistical indicator, and the vertical dimension reflects the power supply enterprise. The statistical indicator and the power supply enterprise are combined into "statistical indicator-power supply enterprise" corresponding to the specific indicator value, and the "English initials or pinyin initials abbreviation" method is used to give it a "unique identifier".

[0060] Entity Learning Step 2: Entity Filling

[0061] Entity filling adds descriptions to the acquired entities so that the entities can be better understood and distinguished. Entity attribute filling and entity relationship filling are the main contents of entity filling. The data layer of the knowledge graph of the power production statistics sample is constructed through the two steps of entity linking and entity filling of entity learning.

[0062] The constructed production statistics knowledge graph is applied in related ways. The applications of power production knowledge graph mainly include intelligent search, graph algorithm research for auxiliary decision-making, and intelligent question and answer.

[0063] Application 1: Intelligent Search

[0064] The intelligent search of the production statistics knowledge graph is guided by common problems in production statistics work (search preference) to design graph query conditions and conduct further analysis and research, that is, to explore the graph search application scenarios of various types of entities, relationships, events and attributes in production statistics business.

[0065] (1) Administrative region query

[0066] This scenario uses the method of drilling down from "province" to "city" to "county" to "power plant" to show the integrity of the hierarchical relationship between regions. The specific display effect is as follows. Search for "XX Province" in the search box, and then click on its "subordinate" relationship to clearly understand the cities in XX Province; take "XX City" as an example, click "Subordinate", expand its "Subordinate" relationship to view the districts and counties of XX City; take "XX County" of XX City as an example, click "Area" to view "XX County Power Supply Co., Ltd." and the various power plants affiliated to XX County. Similarly, this expansion method can view the relevant enterprises of the State Grid in XX Province.

[0067] (2) Query of equipment status

[0068] By using the above search method, combined with basic business electricity knowledge, you can enter common indicators related to equipment statistics in the search box to query the equipment statistics related data of XX Province. The commonly used business indicators for querying equipment status business are the number of substations, the number of main transformers, the substation capacity, the number and length of transmission lines, and the number of distribution equipment and the number of lines. Through the display of these indicators, business personnel can have some preliminary understanding of the overall situation of annual equipment statistical indicators, which lays a certain foundation for further mining and analysis of indicator data content.

[0069] (3) Analysis of industrial electricity consumption

[0070] By using the knowledge graph quick search function and combining it with electricity-related business knowledge, you can directly query the relevant indicator values ​​by selecting the relevant indicators of industrial electricity consumption, including the electricity consumption of the three industrial structures and the total electricity consumption of the whole society; at the same time, you can also query the current value, cumulative value, and the relevant value of the same period of the previous year and the value of the same month of the previous year of the electricity consumption indicator. Relevant business statistical analysts can use this search scenario to simply calculate the year-on-year and proportion of the indicators and conduct preliminary business analysis of the monthly electricity consumption indicators.

[0071] (4) Analysis of clean energy situation

[0072] By using the quick search function of the knowledge graph and combining it with business knowledge related to electricity consumption, you can see the installed capacity of clean energy, the growth rate of electricity consumption compared to the total capacity, and the growth rate of electricity consumption.

[0073] (5) Analysis of electricity consumption by industry

[0074] Use the graph to search for key indicators of electricity consumption analysis by industry, and analyze the changes in industries that correspond to the domestic / international situation based on actual electricity consumption.

[0075] Application 2: Research on graph algorithms for decision support

[0076] Research on graph algorithms for decision support. Path analysis is often used to find the path between two nodes. In the power production statistics knowledge graph, all paths between two indicators can be quickly retrieved, so that newcomers in the industry can quickly understand the business knowledge of power production statistics.

[0077] (1) Relationship analysis can be used to quickly find common topics between two reports in the knowledge graph of power production statistics; in addition, "relationship analysis" can also be applied in the scenario of "querying equipment status". In the application scenario of "equipment status", "relationship analysis" - "association analysis" can be used to quickly find the common reports and related years of two related indicators, and "highlight" the "report name" and related years. Association analysis in relationship analysis: In the graph, find all nodes and relationships within a specified degree between any two nodes. It is used to mine all nodes and relationships within a certain degree between two nodes; relationship analysis can be used to quickly find common topics between two statistical reports in the knowledge graph of power production statistics. Common neighbor analysis in relationship analysis: Get the neighbor nodes shared by two nodes, intuitively find another entity that is strongly related between any two entities in the graph, and further infer the potential relationship and similarity between the two nodes. "Relationship analysis" can be applied in the scenario of "querying equipment status". In the "Equipment Status" application scenario, you can quickly find the common report and related years of two related indicators through "Relationship Analysis" - "Correlation Analysis", and "Highlight" the "Report Name" and related years.

[0078] (2) The "betweenness centrality analysis" and "closeness centrality" in the centrality analysis can be used in the "province-city-county" power company query module of the power production statistics knowledge graph. Through the graph algorithm, the "province" and "city" nodes of the subgraph are lit up, and users can quickly find the relationship and location between provinces and cities in the complex and huge province-city-county relationship graph. Betweenness centrality analysis: In the graph, betweenness centrality is one of the measures of network graph centrality based on the shortest path. In a fully connected network graph, there is at least one shortest path between any two nodes. The betweenness centrality of each node is the number of times these shortest paths pass through the node. The "betweenness centrality analysis" and "closeness centrality" in the centrality analysis can be used in the "province-city-county" power company query module of the power production statistics knowledge graph. Through the graph algorithm, the "province" and "city" nodes of the subgraph are lit up. Closeness centrality algorithm: In the graph, the difficulty of a node to reach other nodes is the reciprocal of the average distance to all other nodes. Users can quickly find the relationship and location between provinces and cities in the complex and large province-city-county relationship map.

[0079] (3) The "Louvain algorithm" of "community discovery analysis" can be used in various reports of different topics in the knowledge graph of power production statistics. Through the algorithm, the graph divides the complete subgraph into different modules and marks them with different colors. If you zoom in on the graph, you can find that in the subgraph, the "Louvain algorithm" automatically divides the subgraph into different modules based on the internal relationship of the report "province-city-county". Community analysis found that in the application of the power production graph, it can be judged whether the two reports are completely separated or there is a certain connection. Through community discovery, business personnel related to power statistics can think about report design: analyze the subsequent optimization of the two reports or whether a new report needs to be designed to strengthen the connection between the two statistical reports. The Louvain algorithm is a community discovery algorithm based on modularity, and its optimization goal is to maximize the modularity of the entire community network. Through the algorithm, the graph divides the complete subgraph into different modules and marks them with different colors. If you zoom in on the graph, you can find that in the subgraph, the "Louvain algorithm" automatically divides the subgraph into different modules based on the internal relationship of the report "province-city-county".

[0080] (4) Research on graph algorithms can be used in "traceable display" to assist decision-making. By visualizing the key path, the reasoning and path that have the greatest impact on the result will be automatically identified and highlighted, so that power production statistics personnel can quickly find the main influencing factors. Graph algorithms are good at discovering the complex relationship characteristics between people. The application in "social scenarios" can be associated with the application of "power equipment". The algorithms that can be implemented in "crowds" can also be applied to "equipment". In the future, graph algorithms can conduct more in-depth exploration and research on the topic of "power equipment", explore more possibilities between "power equipment", and find more possibilities in complex relationships.

[0081] Application 3: Intelligent Question and Answer

[0082] The main contents may include:

[0083] (1) Collect common questions related to production statistics;

[0084] (2) Provide technical solutions to common problems related to production statistics;

[0085] (3) Create a knowledge base to answer questions about current and cumulative values ​​of electricity production statistics;

[0086] (4) Use the knowledge base to answer questions about the year-on-year growth rate, size of increase, and other comprehensive information about electricity production statistics.

[0087] For example, the general process of KBQA based on semantic parsing is: question-->entity linking-->attribute understanding (intent recognition + attribute association)-->structured query---knowledge graph--->answer. Create a knowledge base to answer questions about the current and cumulative values ​​of power production statistics.

[0088] For example, the general process of template-based KBQA is: natural language question matching ---> manually defined templates ---> intention understanding ---> manually annotating the attributes in the knowledge graph corresponding to each template ---> attribute association. Answer questions about the year-on-year growth rate of power production statistics and other comprehensive conditions through the knowledge base.

[0089] The present invention is not limited to the above-mentioned optimal implementation mode. Anyone can derive other various forms of products under the inspiration of the present invention. However, no matter what changes are made in the shape or structure, all technical solutions that are the same or similar to those of the present application fall within the protection scope of the present invention.

[0090] The above method provided in this embodiment can be stored in a computer-readable storage medium in a coded form and implemented in the form of a computer program, and the basic parameter information required for calculation is input through computer hardware, and the calculation result is output.

[0091] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0092] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0093] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0094] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0095] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

[0096] This patent is not limited to the above-mentioned optimal implementation mode. Anyone can derive other forms of knowledge graph construction methods and systems based on power production statistics business under the inspiration of this patent. All equal changes and modifications made according to the scope of the patent application of the present invention should be covered by this patent.

Claims

1. A knowledge graph construction method based on power production statistics business, characterized in that: The following steps are involved: Step S1: sort out the vertical aggregation and horizontal correlation of power production statistical indicators; Step S2: constructing a linkage model of statistical indicators-detailed data-related data-related objects based on the relationship between statistical indicators; Step S3: Draw a graph of the linkage model using ontology design software; Step S4: constructing an application scenario including intelligent analysis and / or intelligent answer based on the knowledge graph; In step S1, the power production statistics are classified by content, including power generation statistics, power supply statistics, power consumption statistics, power balance, load statistics, equipment statistics and energy consumption statistics; the service objects include internal company statistics and external statistics; the statistical caliber includes companies, power grids, regions, rural power grids, different voltage levels, and different industries; the statistical cycle includes monthly, quarterly, and annual; the relationship between power production statistical indicators is divided into three vertical levels of province-city-county, and the horizontal relationship between each level; In step S2, the model layer of the knowledge graph is constructed through ontology learning, including four steps: term extraction, concept extraction, relationship extraction, and rule learning to extract and integrate production statistics related entities. The related entities are: report, indicator, enterprise, power plant, checkpoint, administrative division, industry, month, year, and report theme; In step S3, the data layer of the knowledge graph of power production statistics samples is constructed, and based on ontology learning, entity and relationship information in the data is extracted; the ontology design system is used to realize the graph drawing of the linkage model; and the graph drawing of the statistical indicator-detailed data linkage model is realized through the platform ontology modeling, knowledge extraction, knowledge fusion, knowledge storage, and graph drawing functions.

2. The knowledge graph construction method based on power production statistics business according to claim 1 is characterized by: In step S4, the constructed graph is applied to three types, including intelligent search, graph algorithm research for auxiliary decision-making, and intelligent question and answer.

3. The knowledge graph construction method based on power production statistics business according to claim 1 is characterized by: A top-down construction method is adopted to construct the knowledge graph of power production statistics: first construct the pattern layer of the knowledge graph, and then construct the data layer; that is, first construct the ontology and relationship information in the pattern layer according to the existing structured data or expert knowledge base to form the corresponding conceptual model and rule relationship, and then extract entities from the data according to this pattern to construct the data layer; The model layer of the knowledge graph is constructed through ontology learning, including: term extraction, concept extraction, relationship extraction and rule learning. Four steps are used to integrate 10 entities to form the model layer of the knowledge graph of power production statistics samples, forming the final ontology model, namely the production statistics indicator linkage model; The data layer of the power production statistical sample knowledge graph is constructed through two steps of entity linking and entity filling in entity learning.

4. A knowledge graph system based on power production statistics business, characterized by: The knowledge graph construction method based on the power production statistics business as described in claim 1 includes three application modules: intelligent search, graph algorithm research for auxiliary decision-making, and intelligent question and answer.

5. The knowledge graph system based on power production statistics business according to claim 4 is characterized by: The intelligent search module designs graph query conditions and conducts further analysis and research based on common problems in production statistics work, that is, explores graph search application scenarios of various types of entities, relationships, events and attributes in production statistics business, including: administrative region and unit affiliation query submodule, equipment status query submodule, industrial electricity consumption analysis submodule, clean energy status analysis submodule and industry-specific electricity consumption analysis submodule.

6. The knowledge graph system based on power production statistics business according to claim 4 is characterized by: The graph algorithm research module for auxiliary decision-making is used to quickly retrieve all paths between two indicators.

7. The knowledge graph system based on power production statistics business according to claim 4 is characterized by: The intelligent question and answer module is used to collect commonly used questions related to production statistics, perform technical processing on commonly used questions related to production statistics, create a knowledge base to answer questions about the current value and cumulative value of electricity production statistics, and answer questions about the year-on-year growth rate and size of the increase in electricity production statistics through the knowledge base.

Images