A power grid information question and answer method and device based on a knowledge graph

Abstract

The specification relates to the technical field of computers, in particular to a power grid information question and answer method and device based on a knowledge graph. The power grid information question and answer method based on the knowledge graph comprises extracting question elements for received question information to obtain address region information, question types and question representation information; determining a target knowledge graph based on the address region information; determining sub-answer information from the target knowledge graph based on the question representation information; and processing the sub-answer information using a target information processing model corresponding to the question type to obtain answer information corresponding to the question information. According to the embodiments of the specification, since only the target knowledge graph corresponding to the address region is determined, the waste of computer resources when determining the sub-answer information is reduced. In addition, since the target information processing model corresponding to the question type is used to process the determined sub-answer information to obtain the answer information, the accuracy of the answer information is improved.

Description

Technical Field

[0001] This specification relates to the field of computer technology, and in particular to a knowledge graph-based question-and-answer method and apparatus for power grid information. Background Technology

[0002] Currently, in the power grid sector, when the power grid system interacts with user terminals via question-and-answer mechanisms, the power grid system determines the corresponding answer information from a pre-built knowledge graph based on the question information received from the user terminal. However, due to the large size of the knowledge graph, determining the corresponding answer information using the entire knowledge graph requires enormous computational resources. Furthermore, because the knowledge graph is constructed in a way that the same entity node may have different node identifiers, the accuracy of the obtained answer information is low if the node identifiers are not corrected or if the answer information obtained from the knowledge graph is not intervened.

[0003] How to improve the accuracy of answer information while reducing resource waste is a problem that urgently needs to be solved in existing technologies. Summary of the Invention

[0004] To address the problems in existing technologies, embodiments of this specification provide a knowledge graph-based method, apparatus, computer device, and readable medium for power grid information question answering. By determining the target knowledge graph corresponding to the address region, the waste of computer resources when determining sub-answer information is reduced. Furthermore, by processing the determined sub-answer information based on a target information processing model corresponding to the question type to obtain the answer information, the accuracy of the answer information is improved.

[0005] To solve the above-mentioned technical problems, the specific technical solution in this specification is as follows:

[0006] On the one hand, embodiments of this specification provide a knowledge graph-based question-and-answer method for power grid information, including:

[0007] The received problem information is processed to extract problem elements, resulting in address region information, problem type, and problem representation information.

[0008] Based on the address region information, the target knowledge graph is determined;

[0009] Based on the problem representation information, sub-answer information is determined from the target knowledge graph; and

[0010] Using the target information processing model corresponding to the question type, the sub-answer information is processed to obtain the answer information corresponding to the question information.

[0011] Furthermore, determining the target knowledge graph based on the address region information further includes:

[0012] From a preset knowledge graph, determine the target entities that match the address region information and the target relationships between the target entities; and

[0013] The target knowledge graph is determined based on the target entity and the target relationship between the target entities.

[0014] Furthermore, determining the target knowledge graph based on the address region information further includes,

[0015] Based on the address area information, obtain the regional power grid topology information;

[0016] The power grid topology information of the region is processed by extracting triplet information to obtain the first pre-sub triplet information;

[0017] Retrieve webpage information and rule information;

[0018] The webpage information and rule information are processed by extracting triplet information respectively to construct a second pre-triplet information corresponding to the webpage information and a third pre-triplet information corresponding to the rule information;

[0019] Consistency processing is performed on the first preform triplet information, the second preform triplet information, and the third preform triplet information to obtain a consistent triplet information set; and

[0020] Based on the consistent triplet information set, the target knowledge graph is constructed.

[0021] Furthermore, the acquisition of regional power grid topology information based on the address area information further includes:

[0022] Based on the address region information, the power grid entity nodes are determined from the database;

[0023] The entity node is indexed in the database to determine the parameter information corresponding to the entity node; and

[0024] Based on the parameter information and the entity node, the power grid topology information of the region is determined.

[0025] Furthermore, the consistency processing performed on the first pre-sub triplet information, the second pre-sub triplet information, and the third pre-sub triplet information to obtain a consistent triplet information set further includes,

[0026] For each first entity node in the first pre-sub triplet information, each second entity node in the second pre-sub triplet information, and each third entity node in the third pre-sub triplet information, similarity processing is performed to obtain multiple similarities, and the similarity indicates the degree of similarity between the first entity node, the corresponding second entity node, and the corresponding third entity node.

[0027] From the plurality of similarities, at least one target similarity is determined that has a similarity greater than a preset threshold;

[0028] For each of the at least one target similarity, determine whether the first entity node, the second entity node, and the third entity node of the first target corresponding to the target similarity are consistent;

[0029] If it is determined that the first target first entity node, the first target second entity node, and the first target third entity node are inconsistent, the first target first entity node, the first target second entity node, and the first target third entity node are updated so that the updated first target first entity node, the updated first target second entity node, and the updated first target third entity node are all consistent with the first target first entity node; and

[0030] Based on the second target first entity node, the corresponding second target second entity node, and the corresponding second target third entity node corresponding to similarity less than or equal to the preset threshold, and the updated first target first entity node, the updated first target second entity node, and the updated first target third entity node, the consistent triplet information set is constructed.

[0031] Furthermore, the sub-answer information is processed using a target information processing model corresponding to the question type to obtain answer information corresponding to the question information, which further includes:

[0032] Determine the source information corresponding to each of the sub-answer information;

[0033] Based on the problem type, the target information processing model is determined from multiple preset information processing models;

[0034] Using the target information processing model, the source information is processed to obtain a weight value corresponding to each source information; and

[0035] The answer information is obtained by processing the corresponding sub-answer information using the weight values.

[0036] Furthermore, after processing the sub-answer information using the target information processing model corresponding to the question type to obtain the answer information corresponding to the question information, the process further includes:

[0037] Compare the answer information with the answer threshold information corresponding to the question representation information;

[0038] If the relationship between the answer information and the answer threshold information does not satisfy a preset relationship, obtain the answer knowledge graph corresponding to the answer information; and

[0039] Based on the knowledge graph of the answer, the predicted faulty entity nodes are identified.

[0040] On the other hand, embodiments of this specification also provide a knowledge graph-based power grid information question-and-answer device, including,

[0041] The extraction unit is used to extract problem elements from the received problem information to obtain address area information, problem type, and problem representation information.

[0042] The first determining unit is used to determine the target knowledge graph based on the address region information;

[0043] The second determining unit is configured to determine sub-answer information from the target knowledge graph based on the question representation information; and

[0044] The processing unit is used to process the sub-answer information using a target information processing model corresponding to the question type, so as to obtain answer information corresponding to the question information.

[0045] On the other hand, embodiments of this specification also provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method.

[0046] On the other hand, embodiments of this specification also provide a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the above-described method.

[0047] On the other hand, embodiments of this specification also provide a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0048] Using the embodiments of this specification, upon receiving question information, question elements are extracted to obtain address region information, question type, and question representation information; a target knowledge graph corresponding only to the address region information is determined; based on the target knowledge graph, sub-answer information corresponding to the question representation information is determined; then, using a target information processing model corresponding to the question type, the determined sub-answer information is processed to obtain the answer information corresponding to the question information. Because the sub-answer information is determined only using the target knowledge graph corresponding to the address region information, rather than determining the answer information based on the entire knowledge graph corresponding to the power grid, the waste of computing resources is reduced. Simultaneously, because the answer information is obtained by processing the determined sub-answer information based on the target information processing model corresponding to the question type, rather than arbitrarily determining a sub-answer information as the answer information, the accuracy of the answer information is improved. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 The figure shown is a schematic diagram of an implementation system for a knowledge graph-based power grid information question-answering method according to an embodiment of this specification;

[0051] Figure 2 The diagram shown is a flowchart of a knowledge graph-based power grid information question-and-answer method according to an embodiment of this specification.

[0052] Figure 3 The diagram shown is a flowchart of a method for determining predicted faulty entity nodes according to an embodiment of this specification.

[0053] Figure 4A The diagram shown is a flowchart of a method for determining a target knowledge graph according to an embodiment of this specification;

[0054] Figure 4B The diagram shown is a flowchart of a method for determining a set of consistent triplet information according to an embodiment of this specification;

[0055] Figure 5 The diagram shown is a flowchart of a method for determining answer information according to an embodiment of this specification;

[0056] Figure 6A The diagram shown is a structural schematic of a knowledge graph-based power grid information question-and-answer device according to an embodiment of this specification.

[0057] Figure 6BThe diagram shown is a structural schematic of a knowledge graph-based power grid information question-and-answer device according to an embodiment of this specification.

[0058] Figure 7 This is a schematic diagram of the structure of a computer device according to an embodiment of this specification.

[0059] [Explanation of Labels in the Attached Image]

[0060] 101. User terminal;

[0061] 102. Server;

[0062] 610. Extraction Unit;

[0063] 620. First Determined Unit;

[0064] 630. Second Determined Unit;

[0065] 640. Processing Unit;

[0066] 650. Contrast Unit;

[0067] 660. Acquisition Unit;

[0068] 670. The third unit to be determined;

[0069] 702. Computer equipment;

[0070] 704. Processing equipment;

[0071] 706. Storage resources;

[0072] 708. Drive mechanism;

[0073] 710. Input / Output Module;

[0074] 712. Input devices;

[0075] 714. Output devices;

[0076] 716. Presentation equipment;

[0077] 718. Graphical User Interface;

[0078] 720. Network interface;

[0079] 722. Communication link;

[0080] 724. Communication bus. Detailed Implementation

[0081] The technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this specification.

[0082] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, apparatus, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0083] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0084] In the technical solutions of this specification, the collection, storage, use, processing, transmission, provision, disclosure, and application of user personal information and comment information all comply with the provisions of relevant laws and regulations, necessary confidentiality measures have been taken, and they do not violate public order and good morals.

[0085] Figure 1The diagram illustrates an implementation system for a knowledge graph-based power grid information question-and-answer method according to an embodiment of this specification. The system may include a user terminal 101 and a server 102, which communicate via a network. This network may include a Local Area Network (LAN), a Wide Area Network (WAN), the Internet, or a combination thereof, and is connected to a website, user equipment (e.g., computing devices), and a backend system. An administrator can send a power grid information question-and-answer request from the user terminal 101 to the server 102. This request includes address region information and specific question content information. Upon receiving the request, the server 102 determines the address region information, question type, and question representation information based on the request. It then determines a target knowledge graph based on the address region information; determines sub-answer information from the target knowledge graph based on the question representation information; and processes the sub-answer information using a target information processing model corresponding to the question type to obtain the answer information corresponding to the question information, and sends the answer information to the user terminal 101.

[0086] In addition, after determining the answer information, the server 102 can also compare the answer information with the answer threshold information corresponding to the question representation information; if the relationship between the answer information and the answer threshold information does not meet the preset relationship, it can obtain the answer knowledge graph corresponding to the answer information; and based on the answer knowledge graph, it can determine the predicted fault entity node and send the predicted fault entity node to the user terminal 101 to assist the administrator in making fault decisions.

[0087] Alternatively, server 102 may be a node of a cloud computing system (not shown in the figure), or each server 102 may be a separate cloud computing system comprising multiple computers interconnected by a network and operating as a distributed processing system.

[0088] In an optional embodiment, the user terminal 101 may include electronic devices, including but not limited to smartphones, data acquisition devices, desktop computers, tablets, laptops, smart speakers, digital assistants, augmented reality (AR) / virtual reality (VR) devices, smart wearable devices, and other similar electronic devices. Optionally, the operating system running on the electronic device may include, but is not limited to, Android, iOS, Linux, Windows, etc.

[0089] In addition, it should be noted that, Figure 1 The example shown is merely one application environment provided in this manual. In actual applications, it may include multiple user terminals 101, and this manual does not impose any restrictions.

[0090] like Figure 2 The diagram shows a flowchart of a knowledge graph-based power grid information question-and-answer method according to an embodiment of this specification. This diagram describes the knowledge graph-based power grid information question-and-answer process, but it may include more or fewer operational steps based on conventional or non-creative labor. The order of steps listed in the embodiment is merely one possible execution order among many and does not represent the only possible execution order. In actual system or device products, the methods shown in the embodiment or the accompanying drawings can be executed sequentially or in parallel. Specifically, as shown... Figure 2 As shown, the method may include:

[0091] S210, extract problem elements from the received problem information to obtain address area information, problem type and problem representation information;

[0092] S220, determine the target knowledge graph based on address region information;

[0093] S230, Based on the problem representation information, determine the sub-answer information from the target knowledge graph;

[0094] S240, using the target information processing model corresponding to the question type, processes the sub-answer information.

[0095] The process yields the answer information corresponding to the question information.

[0096] Using the embodiments of this specification, upon receiving question information, question elements are extracted to obtain address region information, question type, and question representation information; a target knowledge graph corresponding only to the address region information is determined; based on the target knowledge graph, sub-answer information corresponding to the question representation information is determined; then, using a target information processing model corresponding to the question type, the determined sub-answer information is processed to obtain the answer information corresponding to the question information. Because the sub-answer information is determined only using the target knowledge graph corresponding to the address region information, rather than determining the answer information based on the entire knowledge graph corresponding to the power grid, the waste of computing resources is reduced. Simultaneously, because the answer information is obtained by processing the determined sub-answer information based on the target information processing model corresponding to the question type, rather than arbitrarily determining a sub-answer information as the answer information, the accuracy of the answer information is improved.

[0097] According to one embodiment of this specification, the question information is a question received by the server from an administrator or user based on a user terminal, such as "What is the current local voltage?". The address region information is the address region to which the question points. For example, if the question information is "What is the current local voltage?", and the user terminal's address (IP address) indicates province A, city B, and district C, the address region information could be one of province A, city B, or district C. The question type can be, for example, numeric or character. Numeric type indicates that the answer to the question is a single value, while character type indicates that the answer to the question is one or more characters. In the case of the question information "What is the current local voltage?", the question type is numeric. The question representation information is information that can index the answer information. In the case of the question information "What is the current local voltage?", the question representation information could be, for example, the current voltage and a numerical value.

[0098] For example, a pre-defined knowledge graph can be constructed based on power grid information. The data in this pre-defined knowledge graph can be updated periodically. For instance, the pre-defined knowledge graph can be constructed using triple information obtained from a power grid database to determine sub-answer information.

[0099] After determining the address region information, a target knowledge graph corresponding to that address region information is selected from a preset knowledge graph. For example, if the question is "What is the current voltage value here?", and the user terminal's address information (IP address) indicates province A, and the address region information is province A, a sub-knowledge graph related to province A is selected from the preset knowledge graph, and this sub-knowledge graph is used as the target knowledge graph.

[0100] After determining the problem representation information and the target knowledge graph, sub-answer information is indexed from the target knowledge graph based on the problem representation information. Specifically, indexing sub-answer information from the target knowledge graph can be, for example, by determining the answer entity node from multiple power grid entity nodes in the target knowledge graph based on the problem representation information; and determining the sub-answer information based on the connection relationships between the answer entity node and the answer entity node.

[0101] For each question type, a corresponding preset information processing model is configured in advance. This preset information processing model may include multiple preset accuracy values, each corresponding to the data source of the power grid entity node and the connection relationship between the power grid entity nodes. For example, in the construction of the preset knowledge graph, if the data of the three power grid entity nodes G, N, and U and the connection relationship between the three power grid entity nodes comes from the power grid database, then the preset accuracy of the sub-answer information determined based on the connection relationship between the three power grid entity nodes G, N, and U can be X%. If the data of the three power grid entity nodes G, N, and U and the connection relationship between the three power grid entity nodes comes from a webpage, then the preset accuracy of the sub-answer information determined based on the connection relationship between the three power grid entity nodes G, N, and U can be Y%. X can be greater than Y. X and Y can be the accuracy of historical answer information determined based on the corresponding data source. Specifically, for numerical question types, the preset information processing model may include multiple preset accuracy values; for character question types, the preset information processing model may include multiple preset weight values.

[0102] After determining the sub-answer information, determine the number of sub-answer information items. If the number of sub-answer information items is 1, determine that sub-answer information item as answer information.

[0103] If the number of sub-answer information entries is greater than or equal to one, a target processing model is determined from the preset processing models based on the question type. Based on the source information of each sub-answer information entry, the target accuracy information corresponding to that sub-answer information entry is determined from the target processing model. The sub-answer information corresponding to the highest target accuracy information among multiple target accuracy information entries is taken as the answer information. The source information represents the data source information that determines the power grid entity node and the connection relationship between power grid entity nodes. For example, the source information can be a power grid database, a webpage, or a rule database, etc.

[0104] According to another embodiment of this specification, determining a target knowledge graph based on address region information includes: determining target entities that match the address region information and target relationships between target entities from a preset knowledge graph; and determining the target knowledge graph based on the target entities and target relationships between target entities.

[0105] When constructing the pre-defined knowledge graph, corresponding address information is configured for each power grid entity node. After determining the address region information, a consistency match is performed between this address region information and the address information associated with each power grid entity node in the pre-defined knowledge graph. The power grid entity nodes that match the address region information are identified as target entities. Based on the target entities and the pre-defined knowledge graph, the target relationships between these target entity nodes are determined. It should be noted that these target relationships are the connection relationships between the power grid entity nodes that match the address region information. After determining the target nodes and target relationships, a target knowledge graph is constructed based on these target nodes and target relationships.

[0106] Figure 3 The diagram shows a flowchart of a method for determining predicted faulty entity nodes according to an embodiment of this specification. This diagram describes a process for determining predicted faulty entity nodes, but based on conventional or non-inventive labor, it may include more or fewer operational steps. Specifically, as shown... Figure 3 As shown, the method may include:

[0107] S350, compare the answer information with the answer threshold information corresponding to the question representation information;

[0108] S360: If the relationship between the answer information and the answer threshold information does not meet the preset relationship, obtain the answer knowledge graph corresponding to the answer information;

[0109] S370, based on the answer knowledge graph, determines the predicted fault entity nodes.

[0110] Using the embodiments of this specification, after determining the answer information, the administrator needs to identify which power grid entity node among the corresponding power grid entity nodes is faulty based on the answer information. This process requires a high level of knowledge from the administrator. Therefore, the embodiments of this specification compare the obtained answer information with the answer threshold information corresponding to the characterization information to identify the predicted faulty entity node, thereby assisting the administrator in decision-making and improving the automation level of the fault decision-making process.

[0111] According to another embodiment of this specification, for each connection relationship between two power grid entity nodes, based on each problem representation information, a corresponding preset threshold information is configured. This preset threshold information is a reasonable value range or reasonable expression information corresponding to the target answer information shown by the connection relationship. If the current target answer information exceeds or falls below the extreme value of this reasonable value range, the two power grid entity nodes may be in a fault state. The target answer information can be, for example, the voltage value between the two current power grid entity nodes. It should be noted that the reasonable expression information is, for example, the range of reasonable expressions for the target answer information when the problem type is character-based. For example, if the reasonable expression information is (OK, Correct, No problem, Reasonable, Right), and the target answer information is "No problem," since the target answer information is one of the reasonable expression information, the two power grid entity nodes may not be in a fault state.

[0112] For example, after determining the answer information, if a fault prediction request is received from the user terminal, the answer information is compared with the answer threshold information corresponding to the question representation information. This answer threshold information corresponding to the question representation information is a preset threshold information associated with the connection relationship between the answer entity nodes corresponding to the answer information.

[0113] After determining the answer threshold information, the relationship between the answer information and the answer threshold information is determined, and it is judged whether the relationship meets preset conditions. If the relationship does not meet the preset conditions, the answer knowledge graph corresponding to the answer information is obtained. If the relationship meets the preset conditions, a fault-free state is determined. The information corresponding to the fault-free state is then sent to the user terminal for display to the administrator.

[0114] The preset relationship between answer information and answer threshold information includes inclusion and non-inclusion. If answer information is included by answer threshold information, the relationship between the answer information and answer threshold information is determined to be inclusion, and this relationship satisfies the preset condition. If answer information is not included by answer threshold information, the relationship between the answer information and answer threshold information is determined to be non-inclusion, and this relationship does not satisfy the preset condition.

[0115] The answer knowledge graph consists of answer entity nodes and the connections between them. Based on this answer knowledge graph, the predicted faulty entity nodes can be identified, for example, by determining the answer entity nodes included in the answer knowledge graph as predicted faulty entity nodes and sending the corresponding identifiers of these predicted faulty entity nodes to the user terminal for display to the administrator.

[0116] Figure 4AThe diagram shows a flowchart of a method for determining a target knowledge graph according to an embodiment of this specification. While the process of determining the target knowledge graph is described in this diagram, it may include more or fewer operational steps based on conventional or non-creative work. Specifically, as shown... Figure 4A As shown, the method may include:

[0117] S421, based on address area information, obtain regional power grid topology information;

[0118] S422, perform triplet information extraction processing on the regional power grid topology information to obtain the first pre-sub triplet information;

[0119] S423, retrieve webpage information and rule information;

[0120] S424, Perform triplet information extraction processing on web page information and rule information respectively to construct the second pre-triplet information corresponding to the web page information and the third pre-triplet information corresponding to the rule information;

[0121] S425, perform consistency processing on the information of the first pre-sub triplet, the information of the second pre-sub triplet, and the information of the third pre-sub triplet to obtain a consistent triplet information set;

[0122] S426, constructing a target knowledge graph based on a consistent triplet information set.

[0123] According to another embodiment of this specification, for each physical region, a corresponding preset power grid topology is constructed based on the power grid database. The physical region can be a province, a city, or a district, etc. The preset power grid topology result information can be, for example, a topology map constructed from the nodes corresponding to all power supply equipment in that region.

[0124] The address area information is matched with each physical area to determine the target physical area that matches the address area information, and the preset power grid topology information corresponding to the target physical area is used as the regional power grid topology result information.

[0125] The process of extracting triplet information from the regional power grid topology information can be used to obtain the first pre-triplet information. For example, it can include determining the connection relationship between every two power grid entity nodes in the regional power grid topology information from the power grid database, thereby obtaining the triplet information of power grid entity node-connection relationship-power grid entity node.

[0126] Obtaining webpage information may include, for example,: determining the target website identifier; determining the corresponding crawler based on the target website identifier; obtaining link information corresponding to the target website based on the crawler; and downloading webpage information based on the link information. Specifically, the target website identifier may be any identifier corresponding to a website that includes target information, and the target information may be information related to address area information.

[0127] Unstructured data, including power grid operation and management regulations, power grid assessment guidelines, and historical power grid evaluation cases, is pre-set in the rule database. This unstructured data is then used as rule information when determining the target knowledge graph.

[0128] After determining the webpage information and rule information, data cleaning, extraction, and construction steps are performed on the webpage information to obtain the second pre-triple information corresponding to the webpage information. It should be noted that extraction can be, for example, entity node extraction and connection relationship extraction, and construction can be, for example, combining the extracted entity nodes and connection relationships into triple information.

[0129] Using data extraction methods, triplet information is extracted from rule information to obtain the third pre-triple information corresponding to the rule information. This data extraction method may include at least one of supervised entity relation extraction, weakly supervised entity relation extraction, and unsupervised entity relation extraction, and at least one of entity relation extraction based on syntactic and semantic features and entity relation extraction based on dictionary construction and rule learning.

[0130] After determining the first pre-triple information (structured data) obtained from the power grid database, the second pre-triple information (semi-structured data) obtained from web page information, and the third pre-triple information (unstructured data) obtained from the rule database, consistency processing is performed on the first, second, and third pre-triple information to obtain a consistent triple information set. Specifically, the node identifiers representing the same power supply equipment in the first, second, and third pre-triple information are updated to consistent node identifiers, resulting in a consistent triple information set.

[0131] Using a knowledge graph construction script, the target knowledge graph is constructed by processing each consistent triplet in the consistent triplet information set.

[0132] Because the embodiments in this specification construct a target knowledge graph based on three types of data: structured data, semi-structured data, and unstructured data, and perform consistency processing on the power grid entity nodes in the first pre-sub triplet information, the second pre-sub triplet information, and the third pre-sub triplet information determined based on these three types of data, the target knowledge graph includes richer data, can support more problems to be solved, and the answers to the determined problems are more accurate.

[0133] According to another embodiment of this specification, obtaining regional power grid topology information based on address area information includes: determining power grid entity nodes from a database based on address area information; indexing the entity nodes in the database to determine parameter information corresponding to the entity nodes; and determining regional power grid topology information based on the parameter information and entity nodes.

[0134] Determining power grid entity nodes from a database based on address region information may include, for example: connecting to the power grid database; constructing and performing a query statement (SQL statement) to retrieve power grid entity nodes; converting the data type, structure, and attributes of the retrieved target data to obtain readable data; determining whether the readable data indicates that the power grid entity node information exists in the database; if it indicates that it exists, constructing a subquery statement to extract the power grid entity node information, continuing this process until the power grid database query is complete; if it indicates that it does not exist, constructing and performing another query statement (SQL statement) to retrieve power grid entity nodes, continuing this process until the power grid database query is complete. The obtained power grid entity node information is then converted into a power grid entity node.

[0135] After identifying the power grid physical nodes, a relational query is constructed based on the storage location corresponding to each physical node to extract the parameter information associated with that node. Then, based on the obtained parameter information and the physical nodes, the regional power grid topology information is determined. For example, if the power grid physical nodes include node Q and node N, and the current flowing from node Q to node N is I, then the power grid topology information is QIN. It is important to note that the parameter information here is the same as the information indicated by the connection relationships described above.

[0136] Figure 4B The diagram shows a flowchart of a method for determining a set of consistent triplet information according to an embodiment of this specification. While the process of determining the set of consistent triplet information is described in this diagram, it can include more or fewer steps based on conventional or non-creative methods. Specifically, as shown... Figure 4B As shown, the method may include:

[0137] S4251, perform similarity processing on each first entity node in the first pre-triple information, each second entity node in the second pre-triple information, and each third entity node in the third pre-triple information to obtain multiple similarities. The similarity indicates the degree of similarity between the first entity node, the corresponding second entity node, and the corresponding third entity node.

[0138] S4252, determine at least one target similarity from multiple similarities that has a similarity greater than a preset threshold;

[0139] S4253, for each of the at least one target similarity, determine whether the first entity node, the second entity node and the third entity node of the first target corresponding to the target similarity are consistent;

[0140] S4254, If it is determined that the first entity node of the first target, the second entity node of the first target, and the third entity node of the first target are inconsistent, the first entity node of the first target, the second entity node of the first target, and the third entity node of the first target are updated so that the updated first entity node of the first target, the updated second entity node of the first target, and the updated third entity node of the first target are all consistent with the first entity node of the first target.

[0141] S4255, based on the second target first entity node and the corresponding second target second entity node and the corresponding second target third entity node corresponding to similarity less than or equal to a preset threshold, and the updated first target first entity node, the updated first target second entity node and the updated first target third entity node, construct a consistent triplet information set.

[0142] According to another embodiment of this specification, similarity processing can, for example, employ a text similarity determination model to process a first entity node, its corresponding second entity node, and its corresponding third entity node, determining the similarity between the three entities. This similarity characterizes the similarity between the three characters; for example, the similarity between "electric grid," "electric grid," and "electric grid" is 1. The text similarity determination model can be any model capable of determining the similarity between characters.

[0143] After performing similarity processing on each first entity node, second entity node, and third entity node, multiple similarity scores are obtained.

[0144] The preset threshold can be, for example, 0.98. Each of the multiple similarities is compared with the preset threshold, and the similarities greater than the preset threshold are determined as the target similarities, thus obtaining at least one target similarity.

[0145] For each target similarity, determine whether the corresponding first entity node, second entity node, and third entity node of the first target are consistent, that is, whether the three characters are the same.

[0146] If the first entity node, the second entity node, and the third entity node of the first target are found to be inconsistent, the second entity node and the third entity node of the first target are modified to be the same as the first entity node of the first target, resulting in the updated first entity node, the updated second entity node, and the updated third entity node of the first target.

[0147] Furthermore, based on the second target first entity node and the corresponding second target second entity node and the corresponding second target third entity node corresponding to similarity less than or equal to a preset threshold, and the updated first target first entity node, the updated first target second entity node and the updated first target third entity node, multiple consistent entity nodes are determined. Then, based on these multiple entity nodes, consistent connection relationships are determined from the corresponding first pre-triple information, the corresponding second pre-triple information and the corresponding third pre-triple information. Finally, based on these multiple consistent entity nodes and consistent connection relationships, a consistent triple information set is determined.

[0148] Figure 5 The diagram shows a flowchart of a method for determining answer information according to an embodiment of this specification. While the process of determining answer information is described in this figure, it may include more or fewer steps based on conventional or non-creative labor. Specifically, as shown... Figure 5 As shown, the method may include:

[0149] S541, determine the source information corresponding to each sub-answer information;

[0150] S542, Based on the problem type, determine the target information processing model from multiple preset information processing models;

[0151] S543, using the target information processing model, processes the source information to obtain the weight value corresponding to each source information;

[0152] S544 uses weight values ​​to process the corresponding sub-answer information to obtain the answer information.

[0153] According to another embodiment of this specification, the source information represents the data source information that determines the power grid entity node and the connection relationship between the power grid entity nodes for the sub-answer information. For each sub-answer information, the corresponding source information is determined separately.

[0154] For each question type, a corresponding preset information processing model is pre-configured. This preset information processing model may include multiple preset weight values, each of which corresponds to a data source for the power grid entity node and the connection relationships between power grid entity nodes. Each preset weight value may be, for example, the accuracy rate of historical answer information determined based on the corresponding data source. For instance, if the accuracy rate of the answer information determined based on the power grid database is 0.97, then the preset weight data corresponding to the power grid database is 0.97. From the multiple preset information processing models, the preset information processing model associated with the question type determined by the question information is determined as the target information processing model.

[0155] After determining the source information and target information processing model, the preset weight value corresponding to each source information is taken from the target information processing model as the weight value.

[0156] Using determined weight values, the sub-answer information corresponding to the source information is weighted to obtain the answer information.

[0157] The weight value corresponding to the power grid database is 0.97 and the sub-answer information is 15; the weight value corresponding to the webpage is 0.01 and the sub-answer information is 19; and the weight value corresponding to the rule database is 0.02 and the sub-answer information is 14. Therefore, the sub-answer information is 15.2 (0.97×15+0.01×19+0.02×14).

[0158] Figure 6A The diagram shown is a structural schematic of a knowledge graph-based power grid information question-and-answer device according to an embodiment of this specification. Figure 6A As shown, including,

[0159] The extraction unit 610 is used to extract problem elements from the received problem information to obtain address area information, problem type and problem characterization information;

[0160] The first determining unit 620 is used to determine the target knowledge graph based on address region information;

[0161] The second determining unit 630 is used to determine sub-answer information from the target knowledge graph based on the problem representation information; and

[0162] The processing unit 640 is used to process the sub-answer information using a target information processing model corresponding to the question type, so as to obtain the answer information corresponding to the question information.

[0163] Since the principle of the above-mentioned device in solving the problem is similar to that of the above-mentioned method, the implementation of the above-mentioned device can refer to the implementation of the above-mentioned method, and the repeated parts will not be described again.

[0164] Figure 6B The diagram shown is a structural schematic of a knowledge graph-based power grid information question-and-answer device according to an embodiment of this specification. Figure 6B As shown, including,

[0165] The comparison unit 650 is used to compare the answer information with the answer threshold information corresponding to the question representation information;

[0166] Acquisition unit 660 is used to acquire an answer knowledge graph corresponding to the answer information when the relationship between the answer information and the answer threshold information does not satisfy a preset relationship; and

[0167] The third determining unit 670 is used to determine the predicted fault entity nodes based on the answer knowledge graph.

[0168] Since the principle of the above-mentioned device in solving the problem is similar to that of the above-mentioned method, the implementation of the above-mentioned device can refer to the implementation of the above-mentioned method, and the repeated parts will not be described again.

[0169] like Figure 7 The diagram illustrates the structure of a computer device according to an embodiment of this specification. The apparatus described in this specification can be the computer device in this embodiment, executing the methods described above. The computer device 702 may include one or more processing devices 704, such as one or more central processing units (CPUs), each of which can implement one or more hardware threads. The computer device 702 may also include any storage resource 706 for storing information of any kind, such as code, settings, data, etc. Without limitation, for example, the storage resource 706 may include any one or more combinations of: any type of RAM, any type of ROM, flash memory, hard disk, optical disk, etc. More generally, any storage resource can use any technology to store information. Furthermore, any storage resource can provide volatile or non-volatile retention of information. Further, any storage resource may represent a fixed or removable component of the computer device 702. In one case, when the processing device 704 executes associated instructions stored in any storage resource or combination of storage resources, the computer device 702 can perform any operation of the associated instructions. The computer device 702 also includes one or more drive mechanisms 708 for interacting with any storage resource, such as a hard disk drive mechanism, an optical disk drive mechanism, etc.

[0170] Computer device 702 may also include an input / output module 710 (I / O) for receiving various inputs (via input device 712) and providing various outputs (via output device 714). A specific output mechanism may include a presentation device 716 and an associated graphical user interface (GUI) 718. In other embodiments, the input / output module 710 (I / O), input device 712, and output device 714 may be omitted, and the device may function solely as a computer device within a network. Computer device 702 may also include one or more network interfaces 720 for exchanging data with other devices via one or more communication links 722. One or more communication buses 724 couple the components described above together.

[0171] Communication link 722 can be implemented in any way, such as via a local area network, a wide area network (e.g., the Internet), a point-to-point connection, or any combination thereof. Communication link 722 may include any combination of hardwired links, wireless links, routers, gateway functions, name servers, etc., governed by any protocol or combination of protocols.

[0172] This specification also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0173] This specification also provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described method.

[0174] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, this specification may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this specification may take the form of a computer program product embodied 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.

[0175] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0176] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0177] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0178] The above specific embodiments further illustrate the purpose, technical solutions, and beneficial effects of this specification. It should be understood that the above are merely specific embodiments of this specification and are not intended to limit the scope of protection of this specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this specification should be included within the scope of protection of this specification.

Claims

1. A knowledge graph-based question-answering method for power grid information, characterized in that, include: The received problem information is processed to extract problem elements, resulting in address region information, problem type, and problem representation information. Based on the address region information, the target knowledge graph is determined; Based on the problem representation information, sub-answer information is determined from the target knowledge graph; as well as Using the target information processing model corresponding to the question type, the sub-answer information is processed to obtain the answer information corresponding to the question information; The determination of the target knowledge graph based on the address region information includes: Based on the address area information, obtain the regional power grid topology information; The power grid topology information of the region is processed by extracting triplet information to obtain the first pre-sub triplet information; Retrieve webpage information and rule information; The webpage information and rule information are processed by extracting triplet information respectively to construct a second pre-triplet information corresponding to the webpage information and a third pre-triplet information corresponding to the rule information; Consistency processing is performed on the first preform triplet information, the second preform triplet information, and the third preform triplet information to obtain a consistent triplet information set; and Based on the consistent triplet information set, the target knowledge graph is constructed.

2. The method according to claim 1, characterized in that, The step of determining the target knowledge graph based on the address region information further includes: From a preset knowledge graph, determine the target entities that match the address region information and the target relationships between the target entities; and The target knowledge graph is determined based on the target entity and the target relationship between the target entities.

3. The method according to claim 1, characterized in that, The process of obtaining regional power grid topology information based on the address area information includes: Based on the address region information, the power grid entity nodes are determined from the database; The entity node is indexed in the database to determine the parameter information corresponding to the entity node; and Based on the parameter information and the entity node, the power grid topology information of the region is determined.

4. The method according to claim 1, characterized in that, The consistency processing of the first pre-sub triplet information, the second pre-sub triplet information, and the third pre-sub triplet information to obtain a consistent triplet information set includes: For each first entity node in the first pre-sub triplet information, each second entity node in the second pre-sub triplet information, and each third entity node in the third pre-sub triplet information, similarity processing is performed to obtain multiple similarities, and the similarity indicates the degree of similarity between the first entity node, the corresponding second entity node, and the corresponding third entity node. From the plurality of similarities, at least one target similarity is determined that has a similarity greater than a preset threshold; For each of the at least one target similarity, determine whether the first entity node, the second entity node, and the third entity node of the first target corresponding to the target similarity are consistent; If it is determined that the first target first entity node, the first target second entity node, and the first target third entity node are inconsistent, the first target first entity node, the first target second entity node, and the first target third entity node are updated so that the updated first target first entity node, the updated first target second entity node, and the updated first target third entity node are all consistent with the first target first entity node; and Based on the second target first entity node, the corresponding second target second entity node, and the corresponding second target third entity node corresponding to similarity less than or equal to the preset threshold, and the updated first target first entity node, the updated first target second entity node, and the updated first target third entity node, the consistent triplet information set is constructed.

5. The method according to claim 1, characterized in that, The step of processing the sub-answer information using a target information processing model corresponding to the question type to obtain answer information corresponding to the question information includes: Determine the source information corresponding to each of the sub-answer information; Based on the problem type, the target information processing model is determined from multiple preset information processing models; Using the target information processing model, the source information is processed to obtain a weight value corresponding to each source information; and The answer information is obtained by processing the corresponding sub-answer information using the weight values.

6. The method according to claim 1, characterized in that, After processing the sub-answer information using the target information processing model corresponding to the question type to obtain the answer information corresponding to the question information, the method further includes: Compare the answer information with the answer threshold information corresponding to the question representation information; If the relationship between the answer information and the answer threshold information does not satisfy a preset relationship, obtain the answer knowledge graph corresponding to the answer information; and Based on the knowledge graph of the answer, the predicted faulty entity nodes are identified.

7. A knowledge graph-based power grid information question-and-answer device, characterized in that, include: The extraction unit is used to extract problem elements from the received problem information to obtain address area information, problem type, and problem representation information. The first determining unit is used to determine the target knowledge graph based on the address region information; The second determining unit is used to determine sub-answer information from the target knowledge graph based on the problem representation information; as well as The processing unit is used to process the sub-answer information using a target information processing model corresponding to the question type to obtain answer information corresponding to the question information; The determination of the target knowledge graph based on the address region information includes: Based on the address area information, obtain the regional power grid topology information; The power grid topology information of the region is processed by extracting triplet information to obtain the first pre-sub triplet information; Retrieve webpage information and rule information; The webpage information and rule information are processed by extracting triplet information respectively to construct a second pre-triplet information corresponding to the webpage information and a third pre-triplet information corresponding to the rule information; Consistency processing is performed on the first preform triplet information, the second preform triplet information, and the third preform triplet information to obtain a consistent triplet information set; and Based on the consistent triplet information set, the target knowledge graph is constructed.

8. A computer 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 computer program, it implements the method of any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the method of any one of claims 1-6.

Images