Schema model construction method and device in knowledge graph, equipment and storage medium

Abstract

The invention provides a schema model construction method and device in a knowledge graph, equipment and a storage medium, and belongs to the technical field of communication. The method comprises the steps of obtaining a text knowledge base of operation and maintenance fault cases; obtaining each schema modeling domain of the operation and maintenance fault knowledge graph from the text knowledge base, and determining a representative ontology name corresponding to each schema modeling domain based on the text knowledge base; the schema modeling domain comprises a fault domain, a fault representation domain, a troubleshooting item domain, a troubleshooting content domain, a troubleshooting conclusion domain, a fault root cause domain and a fault solution domain; acquiring an extended ontology name corresponding to each schema modeling domain and an entity name corresponding to the extended ontology name from a text knowledge base; and constructing the schema model of the operation and maintenance fault knowledge graph by adopting the representative ontology name, the extended ontology name and the entity name corresponding to the extended ontology name of each schema modeling domain. The method is used for solving the problem that a schema model of an operation and maintenance fault knowledge graph is difficult to cover all valuable knowledge through manual design.

Description

Technical Field

[0001] The present invention relates to the field of communication technologies, and in particular to a method, device, equipment and storage medium for constructing a schema model in a knowledge graph. Background Art

[0002] There are many faults in the current communication network operation and maintenance work that need to be handled by field business experts. Although there are fault cases recorded in the historical database, a relatively simple search and recommendation matching method is usually used to meet the work needs of field operation and maintenance personnel. The fault cases retrieved in this way are mostly isolated case texts, and the operation and maintenance personnel are required to analyze, identify and filter the content in them later, which is a difficult task for the operation and maintenance personnel. At the same time, the content obtained based on search and recommendation often cannot meet the actual field needs in terms of accuracy.

[0003] With the development of artificial intelligence, especially the rise of knowledge graph technology, knowledge graph technology can be used to model massive fault cases and construct knowledge graphs. The use of knowledge graph technology can help operation and maintenance personnel quickly and accurately locate fault cases, quickly match potential fault solutions or root causes, and perform statistical analysis on all fault cases with similar fault phenomena. However, the knowledge content of network operation and maintenance fault cases is diverse, covers a wide range, and represents different levels of content richness. It is difficult for manually designed knowledge graph structured (schema) models to cover all valuable knowledge. Summary of the invention

[0004] The present invention provides a method, device, equipment and storage medium for constructing a schema model in a knowledge graph.

[0005] The first aspect of the present disclosure provides a method for constructing a schema model in a knowledge graph, comprising:

[0006] Obtain a text knowledge base of operation and maintenance failure cases;

[0007] Acquire each schema modeling domain of the operation and maintenance fault knowledge graph from the text knowledge base, and determine the representative ontology name corresponding to each schema modeling domain based on the text knowledge base; wherein the schema modeling domain includes a fault domain, a fault manifestation domain, a troubleshooting item domain, a troubleshooting content domain, a troubleshooting conclusion domain, a fault root cause domain, and a fault solution domain;

[0008] Acquire, from the text knowledge base, an extended ontology name corresponding to a representative ontology name of each schema modeling domain, and an entity name corresponding to the extended ontology name;

[0009] The schema model of the operation and maintenance fault knowledge graph is constructed by using the representative ontology names, extended ontology names and entity names corresponding to the extended ontology names corresponding to the schema modeling domains.

[0010] A second aspect of the present disclosure provides a schema model construction device in a knowledge graph, comprising:

[0011] The first acquisition module is used to acquire a text knowledge base of operation and maintenance failure cases;

[0012] The second acquisition module is used to obtain each schema modeling domain of the operation and maintenance fault knowledge graph from the text knowledge base, and determine the representative ontology name corresponding to each schema modeling domain based on the text knowledge base; wherein the schema modeling domain includes a fault domain, a fault manifestation domain, a troubleshooting item domain, a troubleshooting content domain, a troubleshooting conclusion domain, a fault root cause domain and a fault solution domain;

[0013] A third acquisition module is used to acquire, from the text knowledge base, the extended ontology name corresponding to the representative ontology name of each schema modeling domain, and the entity name corresponding to the extended ontology name;

[0014] A construction module is used to construct a schema model of the operation and maintenance fault knowledge graph by using representative ontology names, extended ontology names and entity names corresponding to the extended ontology names corresponding to each of the schema modeling domains.

[0015] A third aspect of the present disclosure provides an electronic device, including:

[0016] at least one processor;

[0017] A memory having one or more programs stored thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method according to the first aspect;

[0018] At least one I / O interface is connected between the processor and the memory and is configured to implement information interaction between the processor and the memory.

[0019] A fourth aspect of the present disclosure provides a computer-readable medium having a computer program stored thereon, wherein when the program is executed by a processor, the method according to the first aspect is implemented.

[0020] The present disclosure has the following advantages:

[0021] Each schema modeling domain of the operation and maintenance fault knowledge graph, the representative ontology name corresponding to each schema modeling domain, the extended ontology name corresponding to the representative ontology name, and the entity name corresponding to the extended ontology name are obtained from the text knowledge base of the operation and maintenance fault cases, and the schema model of the operation and maintenance fault knowledge graph is constructed by using the representative ontology name, the extended ontology name and the entity name corresponding to the extended ontology name of each schema modeling domain, thereby realizing the automatic construction of the schema model of the operation and maintenance fault knowledge graph. Moreover, by identifying the fault domain, fault symptom domain, troubleshooting item domain, troubleshooting content domain, troubleshooting conclusion domain, fault root cause domain and fault solution domain, a complete operation and maintenance fault handling process is covered. Even if the fault knowledge content is extensive and covers a wide range, all valuable content covering the entire fault handling process can be extracted, which greatly simplifies the abstract process of operation and maintenance fault knowledge and lays the foundation for realizing an operation and maintenance system based on a knowledge graph. Moreover, the schema model is constructed based on actual operation and maintenance fault cases, which can cover important and valuable information in operation and maintenance fault cases, and can maximize the refinement of operation and maintenance fault knowledge, thereby enhancing the use value of the knowledge graph constructed based on the schema model in operation and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A schematic diagram of a method for constructing a schema model in a knowledge graph provided in an embodiment of the present disclosure;

[0023] Figure 2 A schematic diagram of the main process of performing representative ontology name discovery and screening according to a schema modeling domain provided in an embodiment of the present disclosure;

[0024] Figure 3 A schematic diagram of a process of extracting an ontology from a schema modeling domain provided in an embodiment of the present disclosure;

[0025] Figure 4 A schematic diagram of a structure of a schema model building device in a knowledge graph provided in an embodiment of the present disclosure;

[0026] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0027] The specific implementation of the present disclosure is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described herein is only used to illustrate and explain the present disclosure, and is not used to limit the present disclosure.

[0028] As used in this disclosure, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0029] The terms used in the present disclosure are only used to describe specific embodiments and are not intended to limit the present disclosure.As used in the present disclosure, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

[0030] When the terms “comprising” and / or “made of…” are used in the present disclosure, it specifies the existence of the stated features, integers, steps, operations, elements and / or components, but does not exclude the existence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0031] Unless otherwise defined, all terms (including technical and scientific terms) used in this disclosure have the same meaning as those commonly understood by those of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and this disclosure, and will not be interpreted as having an idealized or overly formal meaning unless explicitly defined in this disclosure.

[0032] The knowledge graph is a structured semantic knowledge base that describes entities in the physical world and their relationships through a schema model. The schema model defines the format of the data to be added to the knowledge graph. The data in the knowledge graph needs to satisfy the ontology model and relationship model pre-defined by the schema.

[0033] The construction of the schema model in the operation and maintenance fault knowledge graph refers to abstracting a set of business modeling logic from the knowledge of operation and maintenance fault cases (usually operation and maintenance fault case texts), so that the business model can strive to cover the important and valuable information in most operation and maintenance fault knowledge texts. The operation and maintenance fault case can be an operation and maintenance fault case in the field of communication networks. The constructed schema model can be used for the use and modeling of fault knowledge in network management and network maintenance work under communication network operation and maintenance.

[0034] The disclosed embodiment provides a method for constructing a schema model in a knowledge graph, so as to solve the problem that the schema model of an operation and maintenance fault knowledge graph designed manually is difficult to cover all valuable knowledge. Figure 1 The figure shows a schematic diagram of the method for constructing a schema model in a knowledge graph, which can be executed by any electronic device. The method mainly includes:

[0035] Step 101, obtaining a text knowledge base of operation and maintenance failure cases.

[0036] In some embodiments, the knowledge text related to the operation and maintenance failure cases of the communication network is obtained in a variety of ways to construct a text knowledge base of the operation and maintenance failure cases, including but not limited to obtaining the knowledge text related to the operation and maintenance failure cases from company websites, professional books, historical archives and document records. The operation and maintenance failure cases of the communication network can be operation and maintenance failure cases of wireless communication networks or wired communication networks, which are not limited here.

[0037] Step 102, obtaining each schema modeling domain of the operation and maintenance fault knowledge graph from the text knowledge base, and determining the representative ontology name corresponding to each schema modeling domain based on the text knowledge base; wherein the schema modeling domain (also called domain or schema domain) includes a fault domain, a fault manifestation domain, a troubleshooting item domain, a troubleshooting content domain, a troubleshooting conclusion domain, a fault root cause domain and a fault solution domain.

[0038] In some embodiments, each schema modeling domain of the operation and maintenance fault knowledge graph is obtained from a text knowledge base, including: extracting a first semantic layer from the text knowledge base, the first semantic layer including fault discovery semantic information, fault analysis semantic information and fault resolution semantic information corresponding to the complete fault handling process; extracting a second semantic layer associated with the first semantic layer from the text knowledge base, the second semantic layer including: a fault domain and a fault manifestation domain associated with the fault discovery semantic information, a troubleshooting project domain, a troubleshooting content domain, a troubleshooting conclusion domain and a fault root cause domain associated with the fault analysis semantic information, and a fault resolution method domain associated with the fault resolution semantic information.

[0039] Exemplarily, the complete fault handling process includes three processing steps: fault discovery, fault analysis, and fault resolution. According to the complete fault handling process, the first semantic layer is extracted from the text knowledge base, and the first semantic layer includes fault discovery semantic information, fault analysis semantic information, and fault resolution semantic information; based on the fault case text in the text knowledge base, the first semantic layer is expanded to obtain the second semantic layer, and the second semantic layer includes: the fault domain and the fault manifestation domain expanded based on the fault discovery semantic information, and the fault solution domain expanded based on the fault analysis semantic information; the domains in the second semantic layer are summarized to obtain the schema modeling domains of the operation and maintenance fault knowledge graph.

[0040] The text of the operation and maintenance fault case is recorded by different field personnel, and different personnel have different recording preferences. The language organization form of the text is also different, and the richness of the knowledge content is also different. In the embodiment of the present disclosure, the semantic extraction of the text knowledge base of the operation and maintenance fault case is used to identify each schema modeling domain, which solves the problem of difficulty in extracting core knowledge caused by the diverse narrative structure and large differences in content expression of the operation and maintenance fault case text. In addition, in the embodiment of the present disclosure, semantic extraction is performed through the complete fault handling process, so that the obtained schema modeling domains can cover the core value points related to the entire operation and maintenance fault processing process. Semantic extraction is more efficient than manual design, and each schema modeling domain can be obtained based on the text of more operation and maintenance fault cases. Even if the expression content of some operation and maintenance fault cases is not rich enough, a complete schema modeling domain can be obtained.

[0041] In some embodiments, determining the representative ontology name corresponding to each schema modeling domain based on a text knowledge base includes: enumerating at least one ontology expression word corresponding to each schema modeling domain from some operation and maintenance failure cases in the text knowledge base; determining the number of occurrences of each ontology expression word in the text knowledge base; obtaining the similarity between each ontology expression word and the text corresponding to the corresponding schema modeling domain; for each schema modeling domain, determining the representative ontology name of the schema modeling domain according to the number of occurrences and similarity of each ontology expression word corresponding to the schema modeling domain. Specifically, determining the number of occurrences of each ontology expression word in the text knowledge base may be determining the number of occurrences of each ontology expression word in the part of operation and maintenance failure cases.

[0042] Among them, the representative ontology name is used to express the ontology and / or attributes of the schema modeling domain. Figure 2 The following is a schematic diagram of the main process of performing representative ontology name discovery and screening based on the schema modeling domain, including:

[0043] Step 201, performing schema modeling domain expansion based on the text of some operation and maintenance fault cases in the text knowledge base;

[0044] Step 202 , the best ontology expression words are screened in combination with the text of the part of operation and maintenance fault cases, and the best expression words are used as representative ontology names.

[0045] The method of obtaining some operation and maintenance failure cases in the text knowledge base includes: sampling the text knowledge base of operation and maintenance failure cases, with a sampling ratio of not less than 10%, and using the sampled part of the operation and maintenance failure cases for extracting ontology expressions. The best ontology expression word is the representative ontology name obtained by screening according to the frequency of occurrence and similarity.

[0046] Among them, the ontology expressions of the schema modeling domain include ontology words and / or attribute words. For example, from the text knowledge associated with the [fault domain] in the text knowledge base, enumerate the attribute words such as [fault number], [fault location], [fault time], and [fault name]; from the text knowledge associated with the [fault manifestation domain] in the text knowledge base, enumerate the ontology words such as [fault phenomenon], [fault appearance], [fault impact], and [fault manifestation]; from the text knowledge associated with the [fault root cause domain] in the text knowledge base, enumerate the ontology words such as [fault root cause], [fault cause], and [fault source]; from the text knowledge associated with the [fault solution domain] in the text knowledge base, enumerate the ontology words such as [fault phenomenon], [fault appearance], [fault impact], and [fault manifestation]; From the text knowledge associated with the [troubleshooting project domain] in the text knowledge base, ontology words such as [troubleshooting project], [troubleshooting means], [troubleshooting method], and [troubleshooting method] are enumerated; from the text knowledge associated with the [troubleshooting content domain] in the text knowledge base, ontology words such as [troubleshooting content] are enumerated; from the text knowledge associated with the [troubleshooting conclusion domain] in the text knowledge base, ontology words such as [troubleshooting conclusion], [troubleshooting result], [verification conclusion], and [verification result] are enumerated.

[0047] In some embodiments, the similarity between each ontology expression word and the corresponding text of the corresponding schema modeling domain is determined, including: extracting the title text in the text corresponding to each schema modeling domain from some operation and maintenance failure cases, and determining the first semantic vector corresponding to the title text; determining the second semantic vector corresponding to each ontology expression word; obtaining the average value of the first semantic vectors of each title text corresponding to the same schema modeling domain, and obtaining the third semantic vector of the text corresponding to the schema modeling domain; for each ontology expression word, obtaining the cosine similarity between the corresponding second semantic vector and the third semantic vector of the corresponding schema modeling domain as the similarity between the ontology expression word and the corresponding schema modeling domain.

[0048] Among them, the text of the operation and maintenance fault case has a certain format, which specifies the recording location of the text related to each step in the fault operation and maintenance process. Although different people have different expression habits, the basic format of the text is the same. In the embodiment of the present disclosure, the semantic information reflected in the title text of the operation and maintenance fault case is used to express the content of the corresponding schema modeling domain, ensuring the accuracy of the semantics. Specifically, the first line of each paragraph in the text corresponding to the schema modeling domain can be extracted as the title text, or the paragraph with a specific format in the text corresponding to the schema modeling domain can be extracted as the title text. The specific format can be a paragraph without indentation format, a paragraph marked with bold or underline, a paragraph starting with Arabic numerals, etc.

[0049] Exemplarily, extracting title texts from the texts corresponding to each of the schema modeling domains from the partial operation and maintenance fault cases, and determining a first semantic vector corresponding to the title texts, includes: performing the following process for each schema modeling domain: extracting the text corresponding to the schema modeling domain from the partial operation and maintenance fault cases, extracting title texts from the texts corresponding to the extracted schema modeling domains, inputting each extracted title text into a pre-trained language model (such as a bidirectional encoder representation from a transformer (BERT, Bidirectional Encoder Representation from Transformers) model) to encode the text corresponding to the schema modeling domain, and obtaining a first semantic vector corresponding to the schema modeling domain.

[0050] Assuming that the number of title texts corresponding to a certain schema modeling domain is M, then the schema modeling domain corresponds to M first semantic vectors. The vector average value is calculated for each first semantic vector corresponding to the schema modeling domain as the third semantic vector corresponding to the schema modeling domain.

[0051] It should be noted that the semantic vector is obtained here by taking the BERT model as an example. The specific process of obtaining the semantic vector includes: inputting the title text (specifically, a sentence) into the BERT model, and using the vector starting with [CLS] token output by the BERT model as the vector of the sentence. Other methods can also be used to obtain the semantic vector, such as using the word2vec model to obtain the vector of each word in the sentence, and then adding and averaging all the word vectors of the sentence as the vector of the sentence.

[0052] Exemplarily, determining the second semantic vector corresponding to each ontology expression word includes: inputting each ontology expression word into a pre-trained language model (such as a BERT model) for encoding, and obtaining the second semantic vector of each ontology expression word, wherein the number of ontology expression words corresponding to each schema modeling domain is not necessarily the same. Assuming that the number of ontology expression words corresponding to a certain schema modeling domain is N, then the schema modeling domain corresponds to N second semantic vectors.

[0053] It should be noted that the semantic vector is obtained here by taking the BERT model as an example. The specific process of obtaining the semantic vector includes: inputting the title text (specifically, the sentence) into the BERT model, and using the vector starting with [CLS] token output by the BERT model as the vector of the sentence. Other methods can also be used to obtain the semantic vector, such as using the word2vec model to obtain the vector of each word in the sentence, and then adding and averaging all the word vectors of the sentence as the vector of the sentence.

[0054] Exemplarily, after determining the similarity between each ontology expression word and the corresponding text of the corresponding schema modeling domain, the method further includes: filtering out ontology expression words whose similarity is lower than a preset similarity threshold, and if all ontology expression words of the target schema modeling domain are filtered out, restoring the ontology expression words of the target schema modeling domain before filtering out, so that at least one ontology expression word can be retained for each schema modeling domain.

[0055] For example, the similarity between [fault manifestation domain] and [fault phenomenon] is 90%, and the ontology expressions with a cosine similarity threshold greater than 0.8 are retained, then [fault phenomenon] is retained as the ontology expression word of [fault manifestation domain].

[0056] In some embodiments, for each schema modeling domain, determining a representative ontology name corresponding to the schema modeling domain according to the number of occurrences and similarities of each ontology expression word corresponding to the schema modeling domain includes:

[0057] For each schema modeling domain, the following process is performed:

[0058] Obtain the sum of the number of occurrences of each ontology expression word corresponding to the schema modeling domain, and obtain the ratio of the number of occurrences of each ontology expression word corresponding to the schema modeling domain to the sum; obtain the score value of each ontology expression word in the schema modeling domain, and the score value of any ontology expression word is the larger value of the ratio corresponding to the ontology expression word and the corresponding similarity; determine the ontology expression word with the highest score value as the representative ontology name corresponding to the schema modeling domain.

[0059] In this embodiment, the ratio is used to represent the frequency of occurrence of the ontology expression words, and the similarity is used to represent the degree of representation of the extracted content. The ontology expression words are screened from these two dimensions, so that the screened ontology expression words have wide applicability and representativeness.

[0060] In the exemplary embodiment, taking [investigation conclusion domain] as an example, assuming that [investigation conclusion domain] corresponds to 4 ontology expressions, the number of each of the 4 ontology expressions is obtained, the sum of the number of the 4 ontology expressions is calculated, and the ratio of the number of each ontology expression word to the sum is calculated as the frequency of occurrence of the ontology expression word, assuming that the frequency of occurrence of [investigation conclusion] in [investigation conclusion domain] accounts for 10%, the frequency of occurrence of [investigation result] in [investigation conclusion domain] accounts for 80%, and the frequency of occurrence of [verification result] in [investigation conclusion domain] accounts for 5%. Taking [fault manifestation domain] as an example, the frequency of occurrence of [fault phenomenon] accounts for 99%.

[0061] In an exemplary embodiment, the larger value is taken from the ratio of the frequency of occurrence corresponding to the ontology expression word and the corresponding similarity as the score value of the ontology expression word. For example, the frequency of occurrence of [fault phenomenon] in [fault manifestation domain] accounts for 99%; the frequency of occurrence of [fault manifestation] in [fault manifestation domain] accounts for 1%; the similarity between [fault manifestation domain] and [fault phenomenon] is 90%; the similarity between [fault manifestation domain] and [fault manifestation] is 80%; then: the score of [fault manifestation] is max(0.1,0.8)=0.8; the score of [fault phenomenon] is: max(0.99,0.9)=0.99.

[0062] Step 103: Acquire, from the text knowledge base, the extended ontology name corresponding to the representative ontology name of each schema modeling domain, and the entity name corresponding to the extended ontology name.

[0063] In some embodiments, an extended ontology name corresponding to a representative ontology name of each schema modeling domain and an entity name corresponding to the extended ontology name are obtained from a text knowledge base, including: using a pre-trained entity extraction model to extract entity description information from texts corresponding to each schema modeling domain in the text knowledge base; extracting proper nouns from the entity description information based on a pre-configured proper noun library, and using the extracted proper nouns as the extended ontology name corresponding to the corresponding schema modeling domain and as the extended ontology name corresponding to the corresponding entity description information.

[0064] In an exemplary embodiment, Figure 3 The figure shows a schematic diagram of the process of extracting the extended ontology of the schema modeling domain, including:

[0065] Step 301, using predefined entities to pre-extract ontology expressions of the schema modeling domain, the pre-extracted entity description information in the text corresponding to the modeling domain is obtained;

[0066] Step 302 , the ontology is expanded in combination with the knowledge in the field of communication operation and maintenance, that is, the special terms in the field of communication operation and maintenance are extracted from the entity description information.

[0067] For example, during the ontology pre-extraction process, the following situations occur in the pre-extraction results of the schema modeling domain [Troubleshooting Content] in the text of some operation and maintenance failure cases:

[0068] ① Check whether the Radio Resource Control (RRC) connection success rate exceeds the threshold;

[0069] ② Check whether the remote radio unit (RRU) alarm of the cell has a link break alarm;

[0070] ③ Check the basic configuration of the network element baseband unit (BBU);

[0071] ④ Check the TraceStart signaling process;

[0072] ⑤Analyze the success rate of Evolved Radio Access Bearer (ERAB) connection establishment;

[0073] In the process of ontology expansion, the extraction results of [Troubleshooting content] mostly contain special terms in the field of communication operation and maintenance, such as performance indicators, alarms, configurations, etc. These more specific special terms in the field are used to concretize the modeling domain, so that the representative ontology name can be expanded. The descriptive information of the ontology such as performance indicators, alarms, and configurations in the pre-extraction results of the troubleshooting content can be used to expand the representative ontology name of the troubleshooting content. Therefore, the representative ontology name [Troubleshooting content] is concretized by 5 extended ontology names: [Performance indicators], [Alarms], [Configuration], [Parameters], and [Signaling].

[0074] Step 104, using the representative ontology names, extended ontology names, and entity names corresponding to the extended ontology names corresponding to each schema modeling domain, to construct a schema model of the operation and maintenance fault knowledge graph.

[0075] In some embodiments, representative ontology names, extended ontology names, and entity description information corresponding to the extended ontology names of each schema modeling domain are used to construct a schema model of an operation and maintenance fault knowledge graph, including: using representative ontology names, extended ontology names, and entity description information corresponding to the extended ontology names of each schema modeling domain to construct an ontology model of the relationship among the representative ontology names, extended ontology names, and entity description information; determining a head entity based on the extended ontology name corresponding to the schema fault domain, determining a relationship type and a tail entity based on the extended ontology name and entity description information corresponding to the fault manifestation domain, troubleshooting item domain, troubleshooting content domain, troubleshooting conclusion domain, fault root cause domain, and fault solution domain, and constructing a relationship model among the head entity, relationship type, and tail entity; constructing a schema model of an operation and maintenance fault knowledge graph based on the ontology model and the relationship model.

[0076] Table 1 shows an example of an ontology model. In Table 1, the ontology expression words are used as the ontology in the ontology model, and the entity names in the entity description information corresponding to the ontology expression words are used as the entities.

[0077] Table 1 Ontology model

[0078]

[0079] The ontology of the fault domain in the ontology model is taken as the head entity, the ontologies of the fault manifestation domain, troubleshooting project domain, troubleshooting content domain, troubleshooting conclusion domain, fault root cause domain and fault solution domain in the ontology model are taken as the relationship type, and the corresponding entities of the fault manifestation domain, troubleshooting project domain, troubleshooting content domain, troubleshooting conclusion domain, fault root cause domain and fault solution domain in the ontology model are taken as the tail entity, and a relationship model in which the relationship type, the head entity and the tail entity are associated is constructed.

[0080] Table 2 shows an example of a relational model, where relations (i.e., relation types) refer to the ontology of other schema modeling domains except the fault domain.

[0081] Table 2 Relationship model

[0082]

[0083] After constructing the schema model, you can build a knowledge graph based on specific knowledge graph design requirements, that is, the construction of the {ontology (attribute)-relationship-ontology (attribute, here refers to the entity corresponding to the ontology)} triple schema. The knowledge graph example is as follows:

[0084] Fault 34245 -> The fault phenomenon is -> The call is stuck;

[0085] Fault 34245 --> The root cause of the fault is --> BBU failure;

[0086] Fault 34245 --> Solution is --> Replace the BBU;

[0087] Fault 34245 ——> Troubleshooting method is ——> Alarm verification;

[0088] Fault 34245 ——> Troubleshooting method is ——> Signaling verification;

[0089] Fault 34245 ——> Troubleshooting method is ——> Parameter verification;

[0090] Fault 34245 -> Alarm troubleshooting item -> BBU disconnection alarm;

[0091] Fault 34245 -> Signaling troubleshooting item -> DeactiveTrace;

[0092] Fault 34245 -> Parameter troubleshooting item -> BBU activation parameters;

[0093] Fault 34245 -> Troubleshooting conclusion is -> Alarm is normal;

[0094] Fault 34245 -> Troubleshooting conclusion is -> Signaling is normal;

[0095] Fault 34245——>The troubleshooting conclusion is——>Parameter error.

[0096] In the disclosed embodiment, each schema modeling domain of the operation and maintenance fault knowledge graph, the representative ontology name corresponding to each of the schema modeling domains, the extended ontology name corresponding to the representative ontology name, and the entity name corresponding to the extended ontology name are obtained from the text knowledge base of the operation and maintenance fault cases, and the schema model of the operation and maintenance fault knowledge graph is constructed by using the representative ontology name, the extended ontology name and the entity name corresponding to the extended ontology name of each schema modeling domain, thereby realizing the automated construction of the schema model of the operation and maintenance fault knowledge graph. Moreover, by identifying the fault domain, fault symptom domain, troubleshooting item domain, troubleshooting content domain, troubleshooting conclusion domain, fault root cause domain and fault solution domain, a complete operation and maintenance fault handling process is covered. Even if the fault knowledge content is extensive and covers a wide range, it can fully cover the valuable content in the entire fault handling process, greatly simplifying the abstract process of operation and maintenance fault knowledge, and laying the foundation for realizing an operation and maintenance system based on a knowledge graph. Moreover, the schema model is constructed based on actual operation and maintenance fault cases, which can cover the important and valuable information in the operation and maintenance fault cases, and can maximize the refinement of operation and maintenance fault knowledge, thereby enhancing the use value of the knowledge graph constructed based on the schema model in operation and maintenance.

[0097] The step division of the above various methods is only for clear description. When implemented, they can be combined into one step or some steps can be split and decomposed into multiple steps. As long as they include the same logical relationship, they are all within the protection scope of this disclosure. Adding insignificant modifications to the algorithm or process or introducing insignificant designs without changing the core design of the algorithm and process are all within the protection scope of this disclosure.

[0098] The presently disclosed embodiment provides a device for constructing a schema model in a knowledge graph. The specific implementation of the device can be found in the relevant description of the method embodiment, which will not be repeated here. Figure 4 The schematic diagram of the structure of the device is shown, which mainly includes:

[0099] The first acquisition module 401 is used to acquire a text knowledge base of operation and maintenance fault cases;

[0100] The second acquisition module 402 is used to obtain each schema modeling domain of the operation and maintenance fault knowledge graph from the text knowledge base, and determine the representative ontology name corresponding to each schema modeling domain based on the text knowledge base; wherein the schema modeling domain includes a fault domain, a fault manifestation domain, a troubleshooting item domain, a troubleshooting content domain, a troubleshooting conclusion domain, a fault root cause domain and a fault solution domain;

[0101] A third acquisition module 403 is used to acquire, from the text knowledge base, the extended ontology name corresponding to the representative ontology name of each schema modeling domain, and the entity name corresponding to the extended ontology name;

[0102] The construction module 404 is used to construct a schema model of the operation and maintenance fault knowledge graph by using the representative ontology names, extended ontology names and entity names corresponding to the extended ontology names corresponding to the schema modeling domains.

[0103] The functions or modules included in the device provided in the embodiments of the present disclosure can be used to execute the method described in the method embodiments. The specific implementation and technical effects thereof can be referred to the description of the method embodiments above, and will not be described again here for the sake of brevity.

[0104] It should be noted that all modules involved in this embodiment are logic modules. In practical applications, a logic unit may be a physical unit, or a part of a physical unit, or may be implemented as a combination of multiple physical units. In addition, in order to highlight the innovative part of the present disclosure, this embodiment does not introduce units that are not closely related to solving the technical problems proposed by the present disclosure, but this does not mean that there are no other units in this embodiment.

[0105] Reference Figure 5 , an embodiment of the present disclosure provides an electronic device, comprising:

[0106] at least one processor 501;

[0107] A memory 502 having at least one program stored thereon, and when the at least one program is executed by the at least one processor, the at least one processor implements the above method;

[0108] At least one I / O interface 503 is connected between the processor and the memory and is configured to implement information exchange between the processor and the memory.

[0109] Among them, the processor 501 is a device with data processing capabilities, including but not limited to a central processing unit (CPU); the memory 502 is a device with data storage capabilities, including but not limited to random access memory (RAM, more specifically SDRAM, DDR, etc.), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory (FLASH); the I / O interface (read-write interface) 503 is connected between the processor 501 and the memory 502, and can realize information interaction between the processor 501 and the memory 502, including but not limited to a data bus (Bus), etc.

[0110] In some embodiments, the processor 501, the memory 502, and the I / O interface 503 are connected to each other through a bus, and further connected to other components of the computing device.

[0111] This embodiment further provides a computer-readable medium on which a computer program is stored. When the program is executed by a processor, the method provided in this embodiment is implemented. To avoid repeated description, the specific steps of the method are not repeated here.

[0112] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods applied for above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In hardware implementations, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0113] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0114] Those skilled in the art will understand that although some embodiments described herein include certain features included in other embodiments but not other features, the combination of features from different embodiments is meant to be within the scope of the present embodiment and to form different embodiments.

[0115] It is to be understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of the present disclosure, but the present disclosure is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and substance of the present disclosure, and these modifications and improvements are also considered to be within the scope of protection of the present disclosure.

Claims

1. A method for constructing a schema model in a knowledge graph, characterized in that: include: Obtain a text knowledge base of operation and maintenance failure cases; Acquire each schema modeling domain of the operation and maintenance fault knowledge graph from the text knowledge base, and determine the representative ontology name corresponding to each schema modeling domain based on the text knowledge base; wherein each schema modeling domain includes a fault domain, a fault manifestation domain, a troubleshooting item domain, a troubleshooting content domain, a troubleshooting conclusion domain, a fault root cause domain, and a fault solution domain; Acquire, from the text knowledge base, an extended ontology name corresponding to a representative ontology name of each schema modeling domain, and an entity name corresponding to the extended ontology name; The schema model of the operation and maintenance fault knowledge graph is constructed by using the representative ontology names, extended ontology names and entity names corresponding to the extended ontology names corresponding to the schema modeling domains.

2. The method for constructing a schema model in a knowledge graph according to claim 1, characterized in that: The schema modeling domains of the operation and maintenance fault knowledge graph obtained from the text knowledge base include: Extracting a first semantic layer from the text knowledge base, wherein the first semantic layer includes semantic information of fault discovery, semantic information of fault analysis, and semantic information of fault resolution corresponding to a complete fault handling process; The second semantic layer associated with the first semantic layer is extracted from the text knowledge base, and the second semantic layer includes: the fault domain and the fault manifestation domain associated with the fault discovery semantic information, the troubleshooting project domain, the troubleshooting content domain, the troubleshooting conclusion domain and the fault root cause domain associated with the fault analysis semantic information, and the fault solution domain associated with the fault resolution semantic information.

3. The method for constructing a schema model in a knowledge graph according to claim 1, characterized in that: The determining, based on the text knowledge base, a representative ontology name corresponding to each schema modeling domain comprises: Enumerate at least one ontology expression word corresponding to each of the schema modeling domains from some operation and maintenance failure cases in the text knowledge base; Obtaining the number of occurrences of each ontology expression word in the text knowledge base; Determining the similarity between each ontology expression word and the corresponding text of the schema modeling domain; For each of the schema modeling domains, a representative ontology name corresponding to the schema modeling domain is determined according to the number of occurrences and similarities of each ontology expression word corresponding to the schema modeling domain.

4. The method for constructing a schema model in a knowledge graph according to claim 3, characterized in that: Determining the similarity between each ontology expression word and the corresponding text of the schema modeling domain includes: Extracting title texts in texts corresponding to each of the schema modeling domains from the partial operation and maintenance fault cases, and determining a first semantic vector corresponding to the title text; Determine the second semantic vector corresponding to each of the ontology expression words; Obtaining an average value of the first semantic vectors of the title texts corresponding to the same schema modeling domain, and obtaining a third semantic vector of the text corresponding to the schema modeling domain; For each ontology expression word, the cosine similarity between the corresponding second semantic vector and the third semantic vector of the corresponding schema modeling domain is obtained as the similarity between the ontology expression word and the corresponding schema modeling domain.

5. The method for constructing a schema model in a knowledge graph according to claim 3, characterized in that: For each of the schema modeling domains, determining a representative ontology name corresponding to the schema modeling domain according to the number of occurrences and similarities of each ontology expression word corresponding to the schema modeling domain includes: For each of the schema modeling domains, the following process is performed: Obtaining a sum of the number of occurrences of each ontology expression word corresponding to the schema modeling domain, and obtaining a ratio of the number of occurrences of each ontology expression word corresponding to the schema modeling domain to the sum; Obtaining the score value of each ontology expression word of the schema modeling domain, wherein the score value of any ontology expression word is the larger value of the ratio corresponding to the ontology expression word and the corresponding similarity; The ontology expression word with the highest score is determined as the representative ontology name corresponding to the schema modeling domain.

6. The method for constructing a schema model in a knowledge graph according to claim 1, characterized in that: The step of acquiring, from the text knowledge base, an extended ontology name corresponding to a representative ontology name of each schema modeling domain and an entity name corresponding to the extended ontology name comprises: Using a pre-trained entity extraction model, extracting entity description information from the text corresponding to each schema modeling domain in the text knowledge base; Based on a pre-configured proper noun library, proper nouns are extracted from the entity description information, and the extracted proper nouns are used as the extended ontology names corresponding to the schema modeling domain and as the extended ontology names corresponding to the entity description information.

7. The method for constructing a schema model in a knowledge graph according to claim 1, characterized in that: The use of representative ontology names, extended ontology names, and entity description information corresponding to the extended ontology names corresponding to the schema modeling domains to construct a schema model of the operation and maintenance fault knowledge graph includes: Using the representative ontology name corresponding to each of the schema modeling domains, the extended ontology name, and the entity description information corresponding to the extended ontology name, an ontology model of the relationship among the representative ontology name, the extended ontology name, and the entity description information is constructed; Determine the head entity based on the extended ontology name corresponding to the schema fault domain, determine the relationship type and the tail entity based on the extended ontology names and entity description information corresponding to the fault manifestation domain, troubleshooting item domain, troubleshooting content domain, troubleshooting conclusion domain, fault root cause domain and fault solution domain, and construct a relationship model among the head entity, the relationship type and the tail entity; A schema model of the operation and maintenance fault knowledge graph is constructed based on the ontology model and the relationship model.

8. A device for constructing a schema model in a knowledge graph, characterized in that: include: The first acquisition module is used to acquire a text knowledge base of operation and maintenance failure cases; The second acquisition module is used to obtain each schema modeling domain of the operation and maintenance fault knowledge graph from the text knowledge base, and determine the representative ontology name corresponding to each schema modeling domain based on the text knowledge base; wherein the schema modeling domain includes a fault domain, a fault manifestation domain, a troubleshooting item domain, a troubleshooting content domain, a troubleshooting conclusion domain, a fault root cause domain and a fault solution domain; A third acquisition module is used to acquire, from the text knowledge base, the extended ontology name corresponding to the representative ontology name of each schema modeling domain, and the entity name corresponding to the extended ontology name; A construction module is used to construct a schema model of the operation and maintenance fault knowledge graph by using representative ontology names, extended ontology names and entity names corresponding to the extended ontology names corresponding to each of the schema modeling domains.

9. An electronic device, characterized in that: include: at least one processor; A memory having one or more programs stored thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7; At least one I / O interface is connected between the processor and the memory and is configured to implement information interaction between the processor and the memory.

10. A computer readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method according to any one of claims 1 to 7.

Images