CN114330720A

CN114330720A - Knowledge graph construction method and device for cloud computing and storage medium

Info

Publication number: CN114330720A
Application number: CN202111640945.6A
Authority: CN
Inventors: 姜剑
Original assignee: Alibaba China Co Ltd
Current assignee: Alibaba China Co Ltd
Priority date: 2021-12-29
Filing date: 2021-12-29
Publication date: 2022-04-12

Abstract

The embodiment of the application provides a knowledge graph construction method and device for cloud computing and a storage medium. In the embodiment of the application, a knowledge graph relation graph for cloud computing is generated in advance, a plurality of problem type entity types, answer type entity types and edge relations among the entity types in a cloud computing after-sale scene are defined in the knowledge graph relation graph, on the basis of the problem type entity phrases are mined from an original after-sale problem corresponding to the cloud computing service and a standard after-sale problem in an after-sale knowledge base, answer type entity phrases are mined from a standard solution in the after-sale knowledge base, and edges are added between the problem type entity phrases and the answer type entity phrases, so that the knowledge graph for the cloud computing after-sale service is obtained. And then, the knowledge graph is applied to the after-sale service of the cloud computing service, so that various problems brought forward by a user in the process of using the cloud computing service can be solved on line and in time, the cost of human resources is reduced, and the after-sale quality of the cloud computing service is improved.

Description

Knowledge graph construction method and device for cloud computing and storage medium

Technical Field

The present application relates to the field of cloud computing technologies, and in particular, to a method and an apparatus for constructing a knowledge graph for cloud computing, and a storage medium.

Background

Cloud Computing (Cloud Computing) is a kind of distributed Computing, and solves the problem of Computing power through task distribution and combination of Computing results. By means of the cloud computing technology, tens of thousands of data can be processed in a short time, and therefore powerful network services can be achieved. In view of the advantages of cloud computing, more and more enterprise users start to use cloud computing services, and the types of cloud computing services that cloud computing operators can provide are more and more, and the functions of the cloud computing services are also more and more powerful. However, with the problem of after-sales service, more and more users need to consult the related problem of cloud computing service, or after the problem occurs in use, the problem needs to be solved by customer service. However, in the prior art, a manual customer service method is required to greatly reduce the cost of human resources, and an online customer service method is used to reduce the cost of human resources, but the prior online customer service method lacks relevant knowledge storage and cannot well solve various problems that users propose at any time, which is a great technical problem in the field of cloud computing.

Disclosure of Invention

Aspects of the present disclosure provide a method, an apparatus, and a storage medium for constructing a knowledge graph for cloud computing, so as to provide a knowledge graph for cloud computing after-sales services, so as to solve various problems that a user proposes in using a cloud computing service.

The embodiment of the application provides a knowledge graph construction method for cloud computing, which comprises the following steps: mining entity phrases of original after-sale problems corresponding to the cloud computing service and standard after-sale problems in an after-sale knowledge base to obtain a plurality of problem type entity phrases; determining question type entity types to which the question type entity phrases belong based on a plurality of question type entity types defined in a pre-generated knowledge graph relation graph and in combination with attribute information of the question type entity phrases; selecting at least one answer type entity phrase belonging to the answer type entity type from standard solutions in an after-sales knowledge base according to the answer type entity type defined in the knowledge map relational graph; adding edges among the plurality of question type entity phrases and the at least one answer type entity phrase according to edge relations among the question type entity types and between the question type entity types and the answer type entity types defined in the knowledge graph relation graph to obtain a knowledge graph for cloud computing after-sales service.

The embodiment of the application further provides a method for generating a knowledge graph relation graph for cloud computing, which comprises the following steps: generating a plurality of problem type entity types according to the analysis result of the after-sale problems of the cloud computing service, wherein different problem type entity types are used for describing different types of after-sale problems or describing the after-sale problems from different dimensions; generating an answer type entity type, wherein the answer type entity type is used for describing a solution for the after-sales problem; determining edge relations among the multiple problem type entity types and between the multiple problem type entity types and the answer type entity type according to the incidence relation among the after-sales problems described by the multiple problem type entity types and the requirement information of the after-sales problems for the solution; and generating a knowledge graph relation graph according to the multiple question type entity types, the answer type entity types and the edge relation, wherein the knowledge graph relation graph is used for generating a knowledge graph required by cloud computing after-sales service.

An embodiment of the present application further provides an electronic device, including: a memory and a processor; the memory is used for a computer program, and the processor is coupled with the memory and is used for executing the computer program so as to implement the steps in the knowledge graph construction method for cloud computing provided by the embodiment of the application.

An embodiment of the present application further provides an electronic device, including: a memory and a processor; the memory is used for a computer program, and the processor is coupled with the memory and is used for executing the computer program so as to implement the steps in the method for generating the knowledge-graph relationship diagram for cloud computing provided by the embodiment of the application.

Embodiments of the present application further provide a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to implement the steps in the method for constructing a knowledge graph for cloud computing provided by the embodiments of the present application.

Embodiments of the present application further provide a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to implement the steps in the method for generating a knowledge-graph relationship diagram for cloud computing provided by the embodiments of the present application.

In the embodiment of the application, a knowledge graph relation graph for cloud computing is generated in advance, a plurality of problem type entity types, answer type entity types and edge relations among the entity types in a cloud computing after-sale scene are defined in the knowledge graph relation graph, on the basis of the problem type entity phrases are mined from an original after-sale problem corresponding to the cloud computing service and a standard after-sale problem in an after-sale knowledge base, answer type entity phrases are mined from a standard solution in the after-sale knowledge base, and edges are added between the problem type entity phrases and the answer type entity phrases, so that the knowledge graph for the cloud computing after-sale service is obtained. And then, the knowledge graph is applied to the after-sale service of the cloud computing service, so that various problems brought forward by a user in the process of using the cloud computing service can be solved on line and in time, the cost of human resources is reduced, and the after-sale quality of the cloud computing service is improved.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:

fig. 1 is a schematic flowchart of a method for generating a knowledge-graph relationship diagram for cloud computing according to an exemplary embodiment of the present application;

FIG. 2 is an example of a knowledge-graph relationship graph for cloud computing as provided by an exemplary embodiment of the present application;

FIG. 3 is a schematic diagram of a knowledge-graph generated based on the knowledge-graph relationship graph of FIG. 2 provided by an exemplary embodiment of the present application;

FIG. 4 is a flowchart illustrating a method for constructing a knowledge graph for cloud computing according to an exemplary embodiment of the present disclosure;

FIG. 5 is a diagram illustrating a detailed implementation process of step 401 in FIG. 4 according to an exemplary embodiment of the present disclosure;

FIG. 6 is a diagram illustrating a detailed implementation of step 402 in FIG. 4 according to an exemplary embodiment of the present disclosure;

FIG. 7 is a diagram illustrating a detailed implementation of step 403 in FIG. 4 according to an exemplary embodiment of the present disclosure;

fig. 8 is a schematic structural diagram of a knowledge-graph relation graph generating apparatus for cloud computing according to an exemplary embodiment of the present application;

FIG. 9 is a schematic structural diagram of a knowledge graph building apparatus for cloud computing according to an exemplary embodiment of the present application;

fig. 10 is a schematic structural diagram of an electronic device according to an exemplary embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be described in detail and completely with reference to the following specific embodiments of the present application and the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

Aiming at the technical problem that various problems brought forward by a user in the process of using the cloud computing service cannot be solved timely and well, in the embodiment of the application, the knowledge graph concept is applied to the after-sale field of the cloud computing service to obtain the knowledge graph for the cloud computing after-sale service, and then the knowledge graph is applied to the after-sale service of the cloud computing service, so that various problems brought forward by the user in the process of using the cloud computing service can be solved online and timely, the cost of human resources is reduced, and the after-sale quality of the cloud computing service is improved.

The method includes the steps that a precedent for defining and constructing a knowledge graph of an after-sale field of the cloud computing service is not provided in the cloud computing industry, in the embodiment of the application, a brand-new definition mode of the knowledge graph of the after-sale field of the cloud computing service is provided, a knowledge graph relation graph for cloud computing is generated in advance based on the definition, and multiple problem entity types, answer entity types and edge relations among the entity types in the after-sale scene of the cloud computing service are defined through the knowledge graph relation graph; further, various after-sales problems accumulated and precipitated in the after-sales service process of the cloud computing service and some existing solutions are used as data bases to carry out deep analysis and mining to obtain problem type entity phrases under various problem type entity types and answer type entity phrases under the answer type entity types, and edges are added between the problem type entity phrases and the answer type entity phrases based on the edge relationship between the entity types defined in the knowledge graph relation diagram to obtain the knowledge graph for the cloud computing after-sales service. And then, the knowledge graph is applied to the after-sale service of the cloud computing service, so that various problems brought forward by a user in the process of using the cloud computing service can be solved on line and in time, the cost of human resources is reduced, and the after-sale quality of the cloud computing service is improved.

The technical solutions provided by the embodiments of the present application are described in detail below with reference to the accompanying drawings.

Fig. 1 is a flowchart illustrating a method for generating a knowledge-graph relationship diagram for cloud computing according to an exemplary embodiment of the present application. As shown in fig. 1, the method includes:

101. generating a plurality of problem type entity types according to the analysis result of the after-sale problems of the cloud computing service, wherein different problem type entity types are used for describing different types of after-sale problems or describing the after-sale problems from different dimensions;

102. generating an answer type entity type, wherein the answer type entity type is used for describing a solution for the after-sales problem;

103. determining edge relations among the multiple problem type entity types and the answer type entity types according to the incidence relations among the after-sales problems described by the multiple problem type entity types and the requirement information of the after-sales problems on the solution;

104. and generating a knowledge graph relation graph according to the multiple question type entity types, the answer type entity types and the edge relation, wherein the knowledge graph relation graph is used for generating a knowledge graph required by the cloud computing after-sales service.

In the embodiment of the application, after-sale problems corresponding to the cloud computing service are collected, where the after-sale problems are after-sale problems in a broad sense, and include various original after-sale problems that are provided by a user in a process of using the cloud computing service, and also include standard after-sale problems that are formed after being sorted and solved. One of the differences between the original after-sale problem and the standard after-sale problem lies in the expression mode of the problem, the original after-sale problem is directly provided by the user and is more random, flexible or diversified in the expression mode, and the standard after-sale problem is subjected to standardized treatment in the expression mode. For example, one original after-market problem may be: "is the ECS server of my Linux system filled with CPU and available monitoring tools? "correspondingly, a standard after-market question may be" for the ECS server of the Linux system, if there is a monitoring tool for monitoring the usage status of the CPU? ". In this example, the ECS server is a server product that provides an ECS service, which is a cloud computing service. In addition, the original after-market problem differs from the standard after-market problem by two: the original after-market problem does not necessarily have a solution, but the standard after-market problem is a part of the after-market problem having a solution, and for convenience of distinction and description, the solution corresponding to the standard after-market problem is referred to as a standard solution. The after-sale knowledge base corresponding to the cloud computing service can be formed along with the continuous accumulation of the standard after-sale problems and the standard solutions corresponding to the standard after-sale problems, and the standard solutions and the standard problems corresponding to the cloud computing service are stored in the after-sale knowledge base.

In this embodiment, the after-sale problems corresponding to the cloud computing service may be analyzed in multiple dimensions, for example, information of cloud computing products specifically related to each after-sale problem may be analyzed from product dimensions related to the after-sale problems; the problem type dimension can also be used for analyzing the type of the problem to which each after-sale problem belongs; the severity of each after-market problem can also be analyzed from the problem severity dimension; the detail information of each after-sale question can be analyzed from the question detail dimension; the accuracy dimension can also be expressed in questions, whether each after-market question accurately describes the question, and so on.

Based on the analysis result of the after-sales problems corresponding to the cloud computing service, it can be determined from which dimensions the problem types related to the cloud computing service in the after-sales aspect are more reasonable to define, and the definition of the problem types can be given, and the problem types can be used as problem type entity types required for building a knowledge graph. In addition to giving a definition of the type of problem the cloud computing service is involved in after-sales, a definition of a solution is given from the after-sales dimension of the cloud computing service and is taken as the answer-type entity type required for building the knowledge graph. Based on the method, multiple problem type entity types can be generated according to the analysis result of the after-sale problems of the cloud computing service, wherein different problem type entity types are used for describing different types of after-sale problems or describing the after-sale problems from different dimensions; and generating an answer-type entity type for describing a solution to the after-market problem.

Further, after obtaining the plurality of question type entity types and answer type entity types, on one hand, the edge relationships existing between the plurality of question type entity types can be determined according to the incidence relationships between the after-sales questions described by the plurality of question type entity types. For example, there may be edge relationships between different problem entity types that relate to the same cloud computing product for the described after-market problem, edge relationships between problem entity types that describe the same or related after-market problems from different dimensions, and so on. On the other hand, the edge relationship between the multiple question type entity types and the answer type entity types can be determined according to the requirement information of the after-sales questions described by the multiple question type entity types on the solution. The requirement information of the after-sales problem described by the problem type entity type for the solution refers to whether the after-sales problem described by the problem type entity type needs the solution or not, if the solution needs to be solved, an edge relationship exists between the problem type entity type and the answer type entity type, and if the solution does not need to be solved, an edge relationship does not exist between the problem type entity type and the answer type entity type.

After determining the edge relationships that exist between the plurality of question entity types and answer entity types, a knowledge graph relationship graph may be generated from the plurality of question entity types, answer entity types, and the edge relationships, the knowledge graph relationship graph being a basis for generating a knowledge graph required for cloud computing after-market services. A question type entity type, an answer type entity type, and an edge relationship existing between the entity types are defined in the knowledge-graph relationship graph.

Further optionally, in the process of determining the edge relationship existing between the multiple question type entity types and the answer type entity types, the type of the edge relationship may also be determined, and the type of the edge relationship is used to describe the association relationship existing between two entity types corresponding to the edge relationship. In addition, in the embodiment of the present application, names of multiple question type entity types and names of edge relations between multiple question type entity types and answer type entity types may also be determined according to after-sales questions described by the multiple question type entity types. In the embodiment of the present application, names of entity types and names of edge relationships are not limited, and all the naming manners that can distinguish the entity types and the edge relationships are applicable to the embodiment of the present application. Accordingly, a name having a unique identification function may also be determined for the answer-type entity type.

Further optionally, in this embodiment of the application, a cloud computing product related to an after-sale question described by multiple question type entity types may also be obtained, and the cloud computing product is used as attribute information of a corresponding question type entity type and/or answer type entity type. Of course, the name of the question type entity type may also be used as the attribute information of the question type entity type, and similarly, the name of the answer type entity type may also be used as the attribute information.

In the embodiments of the present application, the number and specific implementation of the problem type entity types are not limited, and all entity types that can reflect the after-sales problems of the cloud computing service are suitable for the embodiments of the present application. In an alternative embodiment of the present application, the plurality of problem entity types are divided into two categories, where the first category is a problem-expressing entity type, and the second category is a problem-describing entity type, and each category may further include one or more (i.e., at least one) entity types, i.e., the plurality of problem entity types includes at least one problem-expressing entity type and at least one problem-describing entity type. Wherein each entity type of the problem expression type is used for describing one type of after-sales problems, and each entity type of the problem description type is used for describing the detailed information of the after-sales problems from one dimension.

Further, in an optional embodiment of the present application, 7 entity types are proposed for the after-market domain of cloud computing services, and names are configured for the 7 entity types, specifically, the 7 entity types include: "problem phenomenon", "fuzzy problem", "detail item", "error item", "diagnosis item", "system version", "solution". It should be noted that the names of these 7 entity types are not limited to the examples herein, and other names capable of expressing the same definitions are also applicable to the embodiments of the present application. The definitions for these 7 entity types are shown in table 1 below:

TABLE 1

Of the above 7 entity types, the problem phenomenon, the fuzzy problem, the detail item, the error item, the diagnosis item and the system version belong to the problem type entity type; further, the problem phenomena and the fuzzy problem belong to entity types of a problem expression type, and the detail item, the error report item, the diagnosis item and the system version belong to entity types of a problem description type. The solution belongs to an answer-type entity type.

Further, for the 7 entity types, step 103 is to determine edge relationships between the multiple problem type entity types and the answer type entity type according to the association relationship between the after-sales problems described by the multiple problem type entity types and the requirement information of the after-sales problems for the solution, and may be implemented in the following manner: on one hand, according to the incidence relation among after-sale problems described by various problem type entity types, determining that edge relations exist between problem phenomena and diagnosis items, between error reporting items and problem phenomena, between fuzzy problems and problem phenomena and between fuzzy problems and error reporting items; on the other hand, according to the demand information of the after-sale problems described by the multiple problem type entity types to the solutions, the edge relations between the problem phenomena and the solutions, between the detail items and the solutions, between the system versions and the solutions, between the diagnosis items and the solutions and between the error reporting items and the solutions are determined. Among these, the edge relationships existing between 7 entity types are shown in table 2 below:

TABLE 2

Edge type	Source node	Target node
			Problem solving	Problem phenomena	Solution scheme
Detail drawingThe above-mentioned	Detail item	Solution scheme
			Type of system	System version	Solution scheme
Problem diagnosis	Problem phenomena	Diagnostic items
			Solution to diagnosis	Diagnostic items	Solution scheme
Error reporting phenomenon	Error item	Problem phenomena
			Error reporting solution	Error item	Solution scheme
Included	Problem of blur	Problem phenomenon and error item

In Table 2 above, the edge relationships are directed from the original node to the target node, with either the original node or the target node representing one entity type. In addition, in the above table 2, names of various edge relationships are given, which may indicate types of edge relationships. In detail, the edge relationship directed from problem phenomenon to solution is a problem-solving edge relationship that represents the key phenomenon of the after-market problem described for the problem phenomenon that needs to be addressed; an edge relation pointing to a solution from a detail item is a detail description type edge relation, and the edge relation indicates that detail information of an after-sales problem described by the detail item needs to be provided with a solution; the edge relation pointing from the system version to the solution is an edge relation of the system type, and the edge relation indicates that the solution needs to be given according to the operating system or database information related to the after-sales problem described by the system version; the relationship of the boundary from the problem phenomenon to the diagnosis item is a problem diagnosis type of boundary, which represents that the critical phenomenon of the after-sales problem described for the problem phenomenon needs to be diagnosed; an edge relation pointing from the diagnostic item to the solution is a diagnostic solution type edge relation indicating that a solution needs to be given for the diagnostic information acquired by the diagnostic item; the side relation pointing to the problem phenomenon from the error item is an error image type side relation, which represents the problem phenomenon that the error item needs to be determined according to the error information in the after-sale problem described by the error item; the side relation pointing to the solution from the error item is an error-reporting solution type side relation, and the side relation indicates that the solution needs to be given according to the error-reporting information in the after-sale problem described by the error item; the edge relationship from the fuzzy problem to the problem phenomenon or error item is an inclusive type edge relationship, which indicates that the incomplete or even ambiguous after-sale problem described by the fuzzy problem may belong to the problem phenomenon or error item.

Regarding the edge relationship existing between the 7 entity types given in table 1 and the 7 entity types given in table 2, taking each entity type as a node, and establishing an edge between the nodes according to the edge relationship existing between the entity types, a knowledge graph relationship diagram as shown in fig. 2 can be obtained. In fig. 2, the product nodes appear as attributes of the problem phenomenon node and the fuzzy problem node.

After the knowledge graph relation graph is obtained, entity phrases can be mined from various after-sale problems corresponding to the cloud computing service and existing solutions to obtain entity phrases under various entity types, and the edge relations among the entity phrases are further extracted to construct a specific knowledge graph. Wherein, based on the knowledge-graph relationship shown in fig. 2, the knowledge-graph shown in fig. 3 can be constructed. It should be noted that fig. 3 may be a complete knowledge graph or a part of a complete knowledge graph, depending on the number of entity phrases to be extracted. In fig. 2, the entity types represented by the nodes are distinguished by the gray levels of the nodes, the nodes with different gray levels represent different entity types, and the nodes with the same gray level represent the same entity type; similarly, in fig. 3, the entity types to which the entity phrases represented by the nodes belong are also distinguished by the gray levels of the nodes, the nodes with different gray levels represent the entity phrases belonging to different entity types, and the nodes with the same gray level represent the entity phrases belonging to the same entity type; in addition, in fig. 2 and 3, nodes having the same gray level represent the same entity type.

In the knowledge-graph shown in FIG. 3, the following entity phrases are included: linux, monitoring tools, CPU diagnosis, CPU load over high, server load over high, memory over high, bandwidth run full, ECS, Linux server CPU load over high monitoring program; the "monitoring program with excessive Linux server CPU load" belongs to the entity phrase under the "solution" shown in fig. 2, "Linux" belongs to the entity phrase under the "system version" shown in fig. 2, "monitoring tool" belongs to the entity phrase under the "detail item" shown in fig. 2, "CPU diagnosis" belongs to the entity phrase under the "diagnosis item" shown in fig. 2, "server load is excessive" belongs to the entity phrase under the "fuzzy problem" shown in fig. 2, "CPU load is excessive" belongs to the entity phrase under the "problem phenomenon" shown in fig. 2, "memory is excessive" belongs to the entity phrase under the "problem phenomenon" shown in fig. 2, "bandwidth is full" belongs to the entity phrase under the "problem phenomenon" shown in fig. 2, and "ECS" belongs to the entity phrase under the "product" shown in fig. 2.

An example of the use of the knowledge-graph shown in FIG. 3 is as follows: when the user presents the following after-market problems: "is the ECS server of my Linux system filled with CPU and available monitoring tools? When the problem is solved, various entity phrases such as Linux, ECS server, CPU run-out, monitoring tool and the like can be separated from the after-sale problem, and the solution of the monitoring program with the excessive load of the Linux server CPU can be located by reasoning in the knowledge graph shown in fig. 3 based on the entity phrases, so that an accurate reply is given to the after-sale problem provided by a user, and therefore, the after-sale experience of the cloud computing service can be improved, and the identification accuracy, the solution success rate and the solution efficiency of the after-sale problem can be improved.

In the following embodiments, a process of constructing a knowledge graph based on a knowledge graph relationship diagram provided in the embodiments of the present application will be described in detail.

Fig. 4 is a flowchart illustrating a method for constructing a knowledge graph for cloud computing according to an exemplary embodiment of the present application. The method for constructing the knowledge graph belongs to a cold start construction process of the knowledge graph, and the process for constructing the knowledge graph through cold start is a process for constructing a domain knowledge graph for cloud computing directly based on a knowledge graph relation graph on the basis of no basic knowledge graph. In the embodiment, only a plurality of question type entity types, answer type entity types and edge relations between the question type entity types and between the answer type entity types for the cloud computing after-sales service are defined in the knowledge graph. The cloud computing after-sale service refers to an after-sale service provided for the cloud computing service. Specifically, as shown in fig. 4, the method for constructing the knowledge-graph includes:

401. mining entity phrases of original after-sale problems corresponding to the cloud computing service and standard after-sale problems in an after-sale knowledge base to obtain a plurality of problem type entity phrases;

402. determining problem type entity types to which the problem type entity phrases belong based on a plurality of problem type entity types defined in a pre-generated knowledge graph relation graph and by combining attribute information of the problem type entity phrases;

403. selecting at least one answer type entity phrase belonging to the answer type entity type from standard solutions in an after-sales knowledge base according to the answer type entity type defined in the knowledge map relational graph;

404. adding edges between the plurality of question type entity phrases and the at least one answer type entity phrase according to edge relations between the question type entity types and the answer type entity types defined in the knowledge graph relation graph to obtain a knowledge graph for cloud computing after-sales service.

In this embodiment, on one hand, mining entity phrases is performed on the basis of data, which are original after-sales problems corresponding to cloud computing services and standard after-sales problems in an after-sales knowledge base, to obtain a plurality of problem type entity phrases; classifying the plurality of problem type entity phrases according to a plurality of problem type entity types defined in the knowledge graph relation diagram and combining the attribute information of the plurality of problem type entity phrases, and determining the problem type entity type of each problem type entity phrase. On the other hand, at least one answer-type entity phrase belonging to the answer-type entity type is selected from standard solutions in an after-market knowledge base according to the answer-type entity type defined in the knowledge-graph relationship graph. And then adding edges among the plurality of question type entity phrases and at least one answer type entity phrase according to the edge relations among the question type entity types and the answer type entity types defined in the knowledge graph relation graph to obtain the knowledge graph for the cloud computing after-sales service.

In the embodiment of the present application, the implementation manner of step 401 is not limited, and any implementation manner that entity phrases can be mined from original after-market problems and standard after-market problems is applicable to the embodiment of the present application. In an alternative embodiment of the present application, an implementation of step 401 is given, as shown in fig. 5, and includes: data source selection, phrase mining, phrase vectorization, phrase clustering and entity phrase formation.

Selecting a data source: in the embodiment of the application, a knowledge graph used in the after-sales field of the cloud computing service needs to be constructed, and by combining the knowledge graph relation diagram, it can be known that entity phrases of problem type entity types need to be mined to construct the knowledge graph in the after-sales field, so that the after-sales problems from the cloud computing service are selected as data sources needed for constructing the knowledge graph. In the embodiment of the application, the after-sale problem of the cloud computing service comprises two parts, wherein one part is an original after-sale problem submitted by a user, mainly the original after-sale problem submitted by the user on line or a work order, and the original after-sale problem is relatively random and complex in expression; the other part is that the cloud computing service corresponds to standard after-sales problems which are already precipitated in the after-sales knowledge base, the standard after-sales problems are relatively high-quality problem expressions after being processed in expression, and are similar problems of the original after-sales problems, the standard after-sales problems have corresponding standard solutions, and the standard solutions in the after-sales knowledge base can also be regarded as knowledge points.

Phrase Mining (Phrase Mining): the method aims to automatically extract high-quality entity phrases from a large number of text corpora, mainly solves the problem that a professional dictionary in the professional field is insufficient, and reduces manual sorting cost. In the embodiment of the application, after the data source required for mining the entity phrase is determined, the entity phrase suitable as the problem type entity type is mined from the original after-sale problem and the standard after-sale problem contained in the data source, and the problem of insufficient problem type entity phrases in constructing the knowledge graph used in the after-sale field of the cloud computing service is solved. Alternatively, hot spots or high frequency entity phrases can be mined from these original and standard after-market questions. The mining of the entity phrases specifically comprises the following steps:

firstly, analyzing syntax dependence relation aiming at each original after-sale problem or standard after-sale problem so as to structure the original after-sale problem or standard after-sale problem to obtain words contained in the original after-sale problem or standard after-sale problem; then, counting the frequency of combination occurrence between the words according to whether the relation exists between the two words; then, selecting a part of word pairs with the occurrence frequency higher than a set first frequency threshold or the highest occurrence frequency from all the words according to the occurrence frequency of the combination between the words, extracting the word pairs according to the word order to form phrases, wherein the phrases can be used as entity phrases of the problem entity type, and for convenience of description, the entity phrases of the problem entity type are called problem entity phrases. Further optionally, partial words with the occurrence frequency greater than the second frequency threshold or the highest occurrence frequency may also be extracted, and these words may also be used as partial question type entity phrases. Such as Windows, CentOS, etc., belong to a single word, but are also meaningful entity phrases in the knowledge graph. That is, the question-type entity phrase in the embodiment of the present application is not necessarily a phrase, and may be a word.

It should be noted that, in addition to mining the Entity phrases by using the method provided in the foregoing embodiment, the Entity phrases can be dug by using a text rank (TextRank) algorithm, an AutoPhrase, a Deep Neural Network (DNN), a Named Entity Recognition (NER), and the like.

Phrase vectorization: after phrase mining, original after-sale problems or standard after-sale problems become problem type entity phrases one by one, and the problem type entity phrases are input into a language representation model trained in advance to be vectorized to obtain vector expressions of a plurality of problem type entity phrases. In the embodiment of the present application, the language representation model is not limited, and for example, a neural network model that encodes text using a bidirectional Transformer structure may be used. Specifically, a Pre-training of Deep Bidirectional transforms model, which is a self-coding language model (auto encoder LM), that is a model of Pre-training after Pre-training and fine-tuning (fine-tuning) can be employed, with a bi-directional transform structure that designs two tasks to Pre-train the model. The first task is to train the language model in MaskLM, i.e. when a sentence is entered, randomly select some words to be predicted, then replace them with a special symbol [ MASK ], and then let the model learn the filled words in these places according to the given labels. The second task is additionally provided with a sentence-level continuity prediction task on the basis of a bidirectional Transformer language model, namely whether two pieces of text input into BERT are continuous text is predicted, the model can learn the relation between continuous text segments better by introducing the task, and finally words are represented as mathematical vector expressions through an Embedding (Embedding) model. Of course, in addition to the BERT model, a neural network model such as ELMo or T5(Transfer Text-to-Text Transformer) may be used to perform vectorization processing on the question-type entity phrase.

Phrase clustering: after vectorizing the problem entity phrases extracted above, the problem entity phrases can be subjected to multiple rounds of density clustering according to the vector expression of the problem entity phrases, so as to obtain multiple groups of problem entity phrases. The density clustering algorithm may be, but is not limited to: DBSCAN (sensitivity-Based Spatial Clustering of Applications with Noise). In this embodiment, density clustering refers to calculating distribution density between problem entity phrases through vector expression of the problem entity phrases, analyzing connectivity between the problem entity phrases based on the distribution density, continuously expanding clusters based on the connectivity between the problem entity phrases, and finally obtaining a plurality of clusters, where the problem entity phrases in each cluster are a group of problem entity phrases, and the problem entity phrases in the same group are entity phrases expressing the same or similar meanings, such as "unable to connect", "not connected", and these entity phrases express the same or similar meanings and are clustered into the same group. Further optionally, when the distribution density between the problem entity phrases is calculated through the vector expression of the problem entity phrases, the cosine similarity between the problem entity phrases can be calculated according to the vector expression of the problem entity phrases, and the problem entity phrases are subjected to multiple rounds of density clustering based on the cosine similarity between the problem entity phrases; in addition, except for the first round of clustering, other rounds of clustering are further performed on the basis of the clustering result of the previous round, and the clustering process is performed on each clustering cluster obtained in the previous round. It should be noted that the cosine similarity threshold used in the clustering of different rounds is not the same, and the threshold will gradually increase from low to high as the clustering rounds increase. For example, assuming M problem entity phrases, in a first round of clustering, the M problem entity phrases are clustered into 3 cluster clusters B1 using a first cosine similarity threshold a 1; in the second round of clustering, a second cosine similarity threshold A2 is used, the problem entity phrases in each cluster B1 are further clustered to obtain 2 or more cluster clusters, and the number of new cluster clusters obtained by further clustering different cluster B1 is not necessarily the same; and repeating the steps until a clustering ending condition is met to obtain the final clustering. It should be noted that, in addition to the cosine similarity, the euclidean distance, the explicit distance, and the like may also be used to characterize the distribution density between the problem entity phrases in the clustering process.

In addition, the clustering process can adopt a density clustering mode, and can also adopt clustering algorithms such as K-means and spectral clustering, and the like, and the clustering method is not limited.

And (3) entity formation: after the above steps, a clustering result of the problem entity phrases, i.e., a plurality of groups of problem entity phrases, can be obtained. Further optionally, each group of problem entity phrases may be manually rechecked, and some problem entity phrases that are not suitable for constructing a knowledge graph or are clustered incorrectly are adjusted to obtain a problem entity phrase with better quality.

It should be noted that the above-mentioned phrase vectorization, phrase clustering and entity formation are all optional operations, and may be completed in the entity phrase mining stage, or may be completed in other subsequent stages as needed, and this is not limited. In the present embodiment, the operations described above are described as examples performed in the entity phrase mining phase.

In the above embodiment, a plurality of question type entity phrases can be obtained through mining the entity phrases, and further, the plurality of question type entity phrases can be classified by combining a plurality of question type entity types defined in the knowledge graph relationship diagram, so as to determine the question type entity types to which the plurality of question type entity phrases belong. In the examples of the present application, there is no limitation on the implementation of determining the question entity types to which a plurality of question entity phrases belong. In an alternative embodiment, the detailed implementation process of step 402 includes: dividing problem entity phrases expressing the same or similar meanings into a group according to the meaning expressions of a plurality of problem entity phrases to obtain a plurality of groups of problem entity phrases; determining entity type discrimination information corresponding to each group of problem type entity phrases according to standard after-sales problems of each group of problem type entity phrases in an after-sales knowledge base and the occurrence frequency of standard solutions respectively; and determining the problem type entity type to which each group of problem type entity phrases belongs according to the entity type distinguishing degree information corresponding to each group of problem type entity phrases.

In this specification, in the examples of the present application, the above-mentioned implementation of obtaining multiple sets of problematic entity phrases is not limited. In an optional embodiment, the clustering method mentioned in the phrase clustering section may be adopted, and specifically includes: inputting a plurality of problem type entity phrases into a language representation model for vectorization processing to obtain vector expressions of the problem type entity phrases; and performing multi-turn density clustering on the plurality of problem type entity phrases according to the vector expressions of the plurality of problem type entity phrases to obtain a plurality of groups of problem type entity phrases. It should be noted that, when the clustering method mentioned in the phrase clustering section is used to divide the problem entity phrases expressing the same or similar meanings into a group to obtain multiple groups of problem entity phrases, the operation is performed in the entity phrase mining stage, or the entity phrase classification stage described in step 402, which is not limited herein.

In an optional embodiment, the determining the entity type discrimination information corresponding to each group of problem type entity phrases according to the occurrence frequency of each group of problem type entity phrases in the standard after-sales problem and the standard solution in the after-sales knowledge base respectively includes: for each group of problem type entity phrases, the occurrence frequency of the group of problem type entity phrases in the standard after-sale problem is used as first entity type distinguishing degree information, and the ratio of the occurrence frequency of the group of problem type entity phrases in the standard after-sale problem and the occurrence frequency of the group of problem type entity phrases in the standard solution is calculated and used as second entity type distinguishing degree information.

Further optionally, the plurality of problem entity types includes: an entity type of the problem expression type and an entity type of the problem description type; determining the question type entity type of each group of question type entity phrases according to the entity type distinguishing information corresponding to each group of question type entity phrases, wherein the step of determining the question type entity type of each group of question type entity phrases comprises the following steps: for each group of problem entity phrases, determining that the group of problem entity phrases belong to the entity type of the problem expression type under the condition that the first entity type discrimination information is greater than a set third frequency threshold and the second entity type discrimination information is greater than a set proportion threshold; correspondingly, under the condition that the first entity type discrimination information is less than or equal to the set frequency threshold or the second entity type discrimination information is less than or equal to the set proportion threshold, determining that the group of problem entity phrases belong to the entity types of the problem expression types.

Further optionally, the entity type of the problem expression type at least includes a problem phenomenon, and the definition of the entity type of the problem phenomenon can be referred to the foregoing embodiments, and is not described herein again. Based on this, under the condition that the first entity type discrimination information is greater than the set third frequency threshold and the second entity type discrimination information is greater than the set proportion threshold, determining that the group of problem type entity phrases belong to the entity type of the problem expression type, including: and under the condition that the first entity type discrimination information is greater than a set third frequency threshold and the second entity type discrimination information is greater than a set proportion threshold, determining the problem entity type to which the group of problem entity phrases belongs as a problem phenomenon.

Still further optionally, the entity types of the above problem description type at least include: detail item, error entry and system version. For the definition of the entity type of the centralized question description type, reference may be made to the foregoing embodiments, and details are not repeated here. Based on this, determining that the set of problem entity phrases belong to the entity type of the problem expression type when the first entity type discrimination information is less than or equal to the set third frequency threshold or the second entity type discrimination information is less than or equal to the set proportional threshold includes: determining the problem type entity type to which the group of problem type entity phrases belongs as a system version under the condition that the first entity type discrimination information is greater than a set frequency threshold and the second entity type discrimination information is less than or equal to a set proportion threshold; determining the problem entity type of the group of problem entity phrases as an error item under the condition that the first entity type discrimination information is less than or equal to a set third frequency threshold and the problem entity phrases appear in an error report characteristic rule table; and determining the problem entity type to which the group of problem entity phrases belong as a detail item under the condition that the first entity type discrimination information is less than or equal to a set third frequency threshold and the problem entity phrase does not appear in the error reporting characteristic rule table.

In an alternative embodiment of the present application, the plurality of problem entity types defined in the knowledge-graph relationship graph include: an entity type of at least one problem expressive type and an entity type of at least one problem descriptive type; wherein the entity types of the at least one problem expression type include: problematic phenomena and fuzzy problems; the at least one problem description type entity type includes: detail item, error entry, diagnostic item, and system version. On this basis, the embodiment of the present application provides a detailed implementation manner of the step 402 in determining the problem entity types to which a plurality of problem entity phrases belong, and this detailed implementation manner is also a comprehensive description of the above optional embodiments, specifically, as shown in fig. 6, the detailed implementation process of the step 402 includes: selecting a data source, counting frequency, deciding classification and classifying results.

Selecting a data source: the data source required for this step includes two parts, one part is the multiple problem entity phrases obtained above, and the other part is standard after-sales problems and standard solutions in an after-sales knowledge base.

Frequency statistics: this step requires statistics on the frequency of occurrence of the plurality of problem-type entity phrases in the standard after-market problems in the after-market knowledge base. For example, assuming that "ECS cannot connect" the problem-type entity phrase occurred in 5 standard after-market problems, it can be recorded that the frequency of occurrence of the problem-type entity phrase in standard after-market problems is 5; and the sum of the occurrence frequencies of the problem entity phrases in the same group in the standard after-sales problem is used as the occurrence frequency of the problem entity phrases in the group in the standard after-sales problem. In addition, statistics on the frequency of occurrence of the plurality of problem entity phrases in the standard solutions (i.e., knowledge points) in the after-sales knowledge base are also needed. For example, assuming that "ECS cannot connect" the problem entity phrase occurs in 2 standard solutions as knowledge points, the frequency of occurrence of the problem entity phrase in the standard solution may be recorded as 2, and the sum of the frequency of occurrence of the problem entity phrase in the same group in the standard solution as knowledge points may be recorded as the frequency of occurrence of the problem entity phrase in the group in the standard solution as knowledge points.

And (4) decision classification: this step is a key step of determining which type of problem entity the problem entity phrase belongs to, and in this embodiment, the following classification steps are specifically adopted:

step 1: regarding each group of problem type entity phrases, taking the occurrence frequency of each group of problem type entity phrases in the standard after-sale problem as first entity type distinguishing degree information, judging whether the occurrence frequency of each group of problem type entity phrases in the standard after-sale problem is greater than a preset third frequency threshold value, and if so, executing the

steps

2 and 3; if not, executing the step 4.

Step 2: and calculating second entity type discrimination information d corresponding to each problem type entity phrase according to the occurrence frequency of each problem type entity phrase in the standard after-sale problem and the occurrence frequency of each problem type entity phrase in the standard solution serving as a knowledge point by adopting a formula d f1/f 2. Where f1 represents the frequency of occurrence of each set of problem-type entity phrases in the standard after-market problem, and f2 represents the frequency of occurrence of each set of problem-type entity phrases in the standard solution as a point of knowledge.

And step 3: for each group of problem entity phrases, according to a preset proportion threshold, if second entity type discrimination information d > corresponding to the group of problem entity phrases is equal to the proportion threshold, it indicates that the group of problem entity phrases appears in a plurality of standard after-sale problems corresponding to a standard solution, it determines that the problem entity type to which the group of problem entity phrases belongs is a "problem phenomenon", otherwise, it indicates that the group of problem entity phrases appears more in a plurality of standard solutions, and it determines that the problem entity type to which the group of problem entity phrases belongs is a "system version".

And 4, step 4: for each group of problem entity phrases, if the occurrence frequency of the group of problem entity phrases in the standard after-sale problem is less than or equal to a preset third frequency threshold, firstly, whether the problem entity type of the group of problem entity phrase entities is an error report item is judged according to an error report characteristic rule table. For example, the error reporting feature rule table includes: error reporting words such as error (error), exception (exception), response timeout (time out), failure (failed), etc., error reporting codes such as 404, 505, etc., or various error reporting rule features such as "error reporting is" before a very long continuous english phrase and phrase, etc., if a certain set of problem-type entity phrases contains error reporting words such as error, exception, time out, failed, etc., or contains error reporting codes such as 404, 505, etc., or contains "error reporting is" before a very long continuous english phrase and phrase, etc., it can be determined that the problem-type entity type to which the set of problem-type entity phrases belongs is an error reporting item. For the question entity phrases of the other remaining groups, the question entity type to which they belong can be determined as a detail item.

And (4) classification results: according to the above steps, the present embodiment classifies a plurality of problem type entity phrases into four problem type entity types of problem phenomenon, system version, error item, detail item in units of groups. Further optionally, the classification result may be checked manually. It should be noted that, regarding the two problem entity types, i.e., the diagnostic item and the fuzzy problem, the problem entity phrase required by the two entity types can be constructed manually, but is not limited thereto. For example, in step 4 above, two problem entity types, namely diagnostic terms and fuzzy problems, can be further considered for further subdivision.

In the embodiment of the application, in addition to determining the question type entity type to which the plurality of question type entity phrases belong, the answer type entity type defined in the knowledge graph relationship graph is combined, at least one answer type entity phrase belonging to the answer type entity type is selected from standard solutions in an after-sales knowledge base, and another data condition is provided for constructing the knowledge graph. In the examples of the present application, there is no limitation on the implementation of selecting at least one answer-type entity phrase belonging to an answer-type entity type from standard solutions in an after-market knowledge base. In an alternative embodiment, at least one answer-based entity phrase belonging to the answer-based entity type is selected from the standard solutions in the after-sales knowledge base according to the answer-based entity type defined in the knowledge-graph relationship graph, i.e., one implementation of step 403 above includes:

combining problem type entity phrases belonging to different problem type entity types to obtain a plurality of entity phrase combinations, and counting the co-occurrence frequency of each entity phrase combination in the original after-sale problem and the after-sale knowledge base respectively; and selecting at least one standard solution from the standard solutions in the after-sales knowledge base as at least one answer type entity phrase according to the co-occurrence frequency of each entity phrase combination in the original after-sales problem and the after-sales knowledge base respectively.

Still further optionally, the selecting at least one standard solution from the standard solutions in the after-market knowledge base according to the co-occurrence frequency of each entity phrase combination in the original after-market question and the after-market knowledge base respectively as an embodiment of the at least one answer-type entity phrase comprises: for each entity phrase combination, if the co-occurrence frequency of the entity phrase combination in the original after-sales problem is greater than the first co-occurrence frequency threshold and the co-occurrence frequency of the entity phrase combination in the after-sales knowledge base is greater than the second co-occurrence frequency threshold, the standard solution co-occurring by the entity phrase combination in the after-sales knowledge base is taken as the answer type entity phrase corresponding to the entity phrase combination.

Further, in an alternative embodiment of the present application, a detailed implementation of the step 403 when selecting at least one answer-type entity phrase is given, and this detailed implementation is also a comprehensive description of the above alternative embodiments, specifically, as shown in fig. 7, a detailed implementation of the step 403 includes: selecting a data source, a first co-occurrence statistic, a second co-occurrence statistic and selecting an entity phrase.

Selecting a data source: in this step, the data source has three parts, one part is the plurality of problem entity phrases obtained above, another part is the original after-sales problem, and the third part is the standard after-sales problem and standard solution in the after-sales knowledge base.

First co-occurrence statistics: after obtaining the data source, the step first combines the question entity phrases classified under the several question entity types of "question phenomenon" and "detail item", "error item", "system version", etc., to obtain a plurality of entity phrase combinations. Alternatively, the combination of the question entity phrases classified under the several question entity types of "question phenomenon" and "detail item", "error item", "system version", etc. may be permutation combinations, each resulting in an entity phrase combination. For example, the combination of "ECS cannot be connected", the combination of "reset password" and "Linux" can obtain an entity phrase combination. For each entity phrase combination, the frequency of the simultaneous occurrence of the entity phrase combination in the original after-sales problem, i.e., the co-occurrence frequency, can be counted and recorded.

Second co-occurrence statistics: this step is mainly used to combine the above mentioned entity phrases, count the frequency of their simultaneous occurrence in the standard after-sales problems and corresponding standard solutions in the after-sales knowledge base, i.e. co-occurrence frequency, and record it. And for the same entity phrase combination, summing the co-occurrence frequency of the entity phrase combination in each standard after-sale problem and the co-occurrence frequency of the entity phrase combination in each standard solution to obtain the co-occurrence frequency of the entity phrase combination in an after-sale knowledge base.

Selection of entity phrases: the step can also be called as relation extraction, and is mainly used for identifying answer type entity phrases from standard solutions in an after-sales knowledge base based on the co-occurrence frequency of each entity phrase combination obtained in the step in the original after-sales problem and the after-sales knowledge base respectively. For each entity phrase combination, one of three situations may occur depending on its co-occurrence frequency in the original after-market question and after-market knowledge base, respectively:

the first condition is as follows: for any entity phrase combination, the co-occurrence frequency in the original after-sales problem is higher, for example, greater than the first co-occurrence frequency threshold, and the co-occurrence frequency in the after-sales knowledge base is also higher, for example, greater than the second co-occurrence frequency threshold, which means that the standard solution co-occurring by the entity phrase combination can be used as the solution for the original after-sales problem co-occurring by the entity phrase combination, and therefore, the standard solution co-occurring by the entity phrase combination can be used as the answer type entity phrase corresponding to the entity phrase combination. It should be noted that, in the embodiment of the present application, values of the first co-occurrence frequency threshold and the second co-occurrence frequency threshold are not limited, and may be flexibly set according to application requirements, and the magnitude relationship between the two co-occurrence frequency thresholds is not limited.

Case two: for any entity phrase combination, its co-occurrence frequency in the original after-market problem is high, e.g., greater than the first co-occurrence frequency threshold, but its co-occurrence frequency in the after-market knowledge base is also low, e.g., less than or equal to the second co-occurrence frequency threshold, even though it is not co-occurring in the after-market knowledge base, which indicates that standard solutions corresponding to which original after-market problems co-occurring with that entity phrase combination may be missing from the after-market knowledge base. Further optionally, standard solutions corresponding to the original after-market questions which were co-occurring with the entity phrase combination may be supplemented in the after-market knowledge base, and the newly supplemented standard solutions serve as answer-type entity phrases corresponding to the entity phrase combination.

Case three: for any entity phrase combination, the co-occurrence frequency in the original after-sales problem is low, for example, less than or equal to the first co-occurrence frequency threshold, and the co-occurrence frequency in the after-sales knowledge base is also low, for example, less than or equal to the second co-occurrence frequency threshold, even if the entity phrase combination does not co-occur in the after-sales knowledge base, which indicates that the entity phrase combination may not be too common to form a valuable problem-type entity phrase, and there is no corresponding solution. Such entity phrase combinations may be retained without concern for the answer-type entity phrase corresponding thereto. It should be noted that not all entity phrase combinations may correspond to answer-type entity phrases.

After obtaining the plurality of question-type entity phrases and the question-type entity types to which the question-type entity phrases belong and the at least one answer-type entity phrase, adding edges between the plurality of question-type entity phrases and the at least one answer-type entity phrase according to the edge relations between the question-type entity types and between the answer-type entity types defined in the knowledge graph relationship diagram to obtain the knowledge graph for the cloud computing after-sales service. In an alternative embodiment, an implementation of obtaining a knowledge-graph for cloud computing after-market services includes: adding edges between question type entity phrases in each entity phrase combination corresponding to answer type entity phrases according to an edge relation between question type entity types defined in the knowledge graph relation graph; further, according to an edge relation between the question type entity type and the answer type entity type defined in the knowledge graph relation graph, edges are added between each question type entity phrase in the entity phrase combination and the answer type entity phrase corresponding to the entity phrase combination, and the knowledge graph for cloud computing after-sales service is obtained.

For example, the following steps are carried out: in conjunction with the knowledge-graph relationship diagram shown in FIG. 2, assume that an entity phrase combination is: "Linux, CPU diagnosis, CPU overload", and the corresponding answer type entity phrase is a monitoring program "Linux server CPU overload", the "Linux" belongs to an entity phrase of system version type, the "CPU diagnosis" belongs to an entity phrase of diagnosis item type, and the "CPU overload" belongs to an entity phrase of problem phenomenon type, and it can be known from the relationship between these problem type entity types as shown in fig. 2 that an edge needs to be added between the problem type entity phrase "CPU overload" and the problem type entity phrase "CPU diagnosis"; further, according to the edge relationship between the question type entity type and the answer type entity type shown in fig. 2, an edge is added between the question type entity phrase "CPU is overloaded" and the answer type entity phrase "monitor program with overloaded Linux server CPU", an edge is added between the question type entity phrase "CPU diagnosis" and the answer type entity phrase "monitor program with overloaded Linux server CPU", and an edge is added between the question type entity phrase "Linux" and the answer type entity phrase "monitor program with overloaded Linux server CPU". After all the entity phrase combinations corresponding to the answer-type entity phrases are processed in the same manner, a knowledge graph as shown in fig. 3 can be obtained.

It should be noted that after the knowledge graph is obtained, the knowledge graph may be rechecked and corrected manually, and various entity phrases and their relationships are corrected, so as to obtain a more complete knowledge graph.

In the embodiment of the application, the architecture definition of the knowledge graph in the cloud computing field, namely the knowledge graph relation graph, is innovatively provided for the first time, the construction method of the knowledge graph in the cloud computing field at the cold start stage is innovatively provided for the first time, the data source only needs the original after-sale problems and the existing after-sale knowledge base, which are provided by a user, a large amount of sample data does not need to be collected, the requirement of the knowledge graph construction on the data source can be reduced, and the knowledge graph construction efficiency is improved.

Further, after the knowledge graph in the cloud computing field is obtained, the constructed knowledge graph can be applied to an after-sale client robot of the cloud computing service, so that the after-sale problems brought forward by a user in the process of using the cloud computing service can be solved timely, accurately and efficiently, the use experience of the user on the cloud computing service is improved, and the use and popularization of the cloud computing service can be further promoted.

It should be noted that the execution subjects of the steps of the methods provided in the above embodiments may be the same device, or different devices may be used as the execution subjects of the methods. For example, the execution subjects of steps 401 to 404 may be device a; for another example, the execution subject of

steps

401 and 402 may be device a, and the execution subject of

steps

403 and 404 may be device B; and so on.

In addition, in some of the flows described in the above embodiments and the drawings, a plurality of operations are included in a specific order, but it should be clearly understood that the operations may be executed out of the order presented herein or in parallel, and the sequence numbers of the operations, such as 401, 402, etc., are merely used to distinguish various operations, and the sequence numbers themselves do not represent any execution order. Additionally, the flows may include more or fewer operations, and the operations may be performed sequentially or in parallel. It should be noted that, the descriptions of "first", "second", etc. in this document are used for distinguishing different messages, devices, modules, etc., and do not represent a sequential order, nor limit the types of "first" and "second" to be different.

Fig. 8 is a schematic structural diagram of a knowledge-graph relation graph generating apparatus for cloud computing according to an exemplary embodiment of the present application. As shown in fig. 8, the apparatus includes: a first generation module 81, a determination module 82 and a second generation module 83.

The first generation module 81 is configured to generate multiple problem type entity types according to an analysis result of the after-sales problems of the cloud computing service, where different problem type entity types are used to describe different types of the after-sales problems or describe the after-sales problems from different dimensions; and generating an answer type entity type, wherein the answer type entity type is used for describing a solution for the after-sales problem.

The determining module 82 is configured to determine, according to the association relationship between the after-sales problems described by the multiple problem type entity types and the requirement information of the after-sales problems for the solution, the edge relationship existing between the multiple problem type entity types and the answer type entity types.

And a second generating module 83, configured to generate a knowledge graph relationship diagram according to the multiple question type entity types, the answer type entity types, and the edge relationship, where the knowledge graph relationship diagram is used to generate a knowledge graph required by the cloud computing after-sales service.

In an optional embodiment, the second generating module 83 is further configured to: and acquiring a cloud computing product related to the after-sale questions described by the plurality of question type entity types, and using the cloud computing product as attribute information corresponding to the question type entity types and/or answer type entity types.

In an optional embodiment, the second generating module 83 is specifically configured to: determining edge relations between problem phenomena and diagnosis items, between error reporting items and problem phenomena, between fuzzy problems and problem phenomena and between fuzzy problems and error reporting items according to the incidence relations between after-sale problems described by the multiple problem type entity types; determining that edge relations exist between problem phenomena and solutions, between detail items and solutions, between system versions and solutions, between diagnosis items and solutions, and between error reporting items and solutions according to requirement information of after-sale problems described by the multiple problem type entity types on the solutions; wherein the solution is a name of the answer-type entity type.

Fig. 9 is a schematic structural diagram of a knowledge graph constructing apparatus for cloud computing according to an exemplary embodiment of the present application. As shown in fig. 9, the apparatus includes: a mining module 91, a determination module 92, a selection module 93 and a generation module 94.

The mining module 91 is configured to mine entity phrases for the original after-market problems corresponding to the cloud computing service and the standard after-market problems in the after-market knowledge base, so as to obtain a plurality of problem-type entity phrases.

The determining module 92 is configured to determine, based on multiple question type entity types defined in a pre-generated knowledge-graph relationship diagram, a question type entity type to which a plurality of question type entity phrases belong, in combination with attribute information of the plurality of question type entity phrases.

A selecting module 93, configured to select at least one answer type entity phrase belonging to the answer type entity type from standard solutions in an after-sales knowledge base according to the answer type entity type defined in the knowledge-graph relationship diagram.

A generating module 94, configured to add edges between the plurality of question-type entity phrases and the at least one answer-type entity phrase according to the edge relationships between the question-type entity types and the answer-type entity types defined in the knowledge graph relationship diagram, to obtain a knowledge graph for cloud computing after-sales service.

In an alternative embodiment, the determining module 92, when determining the question entity types to which the question entity phrases belong, is specifically configured to: dividing problem entity phrases expressing the same or similar meanings into a group according to the meaning expressions of a plurality of problem entity phrases to obtain a plurality of groups of problem entity phrases; determining entity type discrimination information corresponding to each group of problem type entity phrases according to standard after-sales problems of each group of problem type entity phrases in an after-sales knowledge base and the occurrence frequency of standard solutions respectively; and determining the problem type entity type to which each group of problem type entity phrases belongs according to the entity type distinguishing degree information corresponding to each group of problem type entity phrases.

Further optionally, when obtaining multiple sets of problem entity phrases, the determining module 92 is specifically configured to: inputting a plurality of problem type entity phrases into a language representation model for vectorization processing to obtain vector expressions of the problem type entity phrases; and performing multi-turn density clustering on the plurality of problem type entity phrases according to the vector expressions of the plurality of problem type entity phrases to obtain a plurality of groups of problem type entity phrases.

Further optionally, when determining the entity type discrimination information corresponding to each group of problem-type entity phrases, the determining module 92 is specifically configured to: for each group of problem type entity phrases, the occurrence frequency of the group of problem type entity phrases in the standard after-sale problem is used as first entity type distinguishing degree information, and the ratio of the occurrence frequency of the group of problem type entity phrases in the standard after-sale problem and the occurrence frequency of the group of problem type entity phrases in the standard solution is calculated and used as second entity type distinguishing degree information.

Further optionally, the plurality of problem entity types include: an entity type of the problem expressive type and an entity type of the problem descriptive type. Based on this, when determining the entity type discrimination information corresponding to each group of problem-type entity phrases, the determining module 92 is specifically configured to: for each group of problem entity phrases, determining that the group of problem entity phrases belong to the entity type of the problem expression type under the condition that the first entity type discrimination information is greater than a set frequency threshold and the second entity type discrimination information is greater than a set proportion threshold; and determining the entity type of the question entity phrase group belonging to the question expression type under the condition that the first entity type discrimination information is less than or equal to a set frequency threshold or the second entity type discrimination information is less than or equal to a set proportion threshold.

Further, the entity type of the problem expression type includes at least a problem phenomenon, which is an entity type for describing a key phenomenon exhibited by an after-market problem. Based on this, when determining that the set of problem entity phrases belongs to the entity types of the problem expression type, the determining module 92 is specifically configured to: and under the condition that the first entity type discrimination information is greater than a set frequency threshold and the second entity type discrimination information is greater than a set proportion threshold, determining the problem entity type to which the group of problem entity phrases belongs as a problem phenomenon.

Further, the entity types of the problem description type at least include: detail item, error entry and system version; the detail item is an entity type of detail information for describing the after-sales problem, the error item is an entity type of error information occurring in the after-sales problem, and the system version is an entity type of operating system information for describing the after-sales problem. Based on this, when determining that the set of problem entity phrases belongs to the entity types of the problem expression type, the determining module 92 is specifically configured to: determining the problem type entity type to which the group of problem type entity phrases belongs as a system version under the condition that the first entity type discrimination information is greater than a set frequency threshold and the second entity type discrimination information is less than or equal to a set proportion threshold; determining the problem type entity type of the group of problem type entity phrases as an error report item under the condition that the first entity type discrimination information is less than or equal to a set frequency threshold and the problem type entity phrases appear in an error report characteristic rule table; and determining the problem type entity type to which the group of problem type entity phrases belong as a detail item under the condition that the first entity type discrimination information is less than or equal to a set frequency threshold and the problem type entity phrases do not appear in the error reporting characteristic rule table.

In an alternative embodiment, the selecting module 83, when selecting at least one answer-type entity phrase belonging to the answer-type entity type, is specifically configured to: combining problem type entity phrases belonging to different problem type entity types to obtain a plurality of entity phrase combinations, and counting the co-occurrence frequency of each entity phrase combination in the original after-sale problem and the after-sale knowledge base respectively; and selecting at least one standard solution from the standard solutions in the after-sales knowledge base as at least one answer type entity phrase according to the co-occurrence frequency of each entity phrase combination in the original after-sales problem and the after-sales knowledge base respectively.

Further optionally, the selecting module 83, when selecting at least one standard solution from the standard solutions in the after-sales knowledge base as at least one answer-type entity phrase, is specifically configured to: for each entity phrase combination, if the co-occurrence frequency of the entity phrase combination in the original after-sales problem is greater than the first co-occurrence frequency threshold and the co-occurrence frequency of the entity phrase combination in the after-sales knowledge base is greater than the second co-occurrence frequency threshold, the standard solution co-occurring by the entity phrase combination in the after-sales knowledge base is taken as the answer type entity phrase corresponding to the entity phrase combination.

In an optional embodiment, when obtaining the knowledge-graph for the cloud computing after-sales service, the generating module 84 is specifically configured to: adding edges among all question type entity phrases in each entity phrase combination corresponding to the answer type entity phrases according to an edge relation between the question type entity types defined in the knowledge graph relation graph; and adding edges between each question type entity phrase in the entity phrase combination and the answer type entity phrase corresponding to the entity phrase combination according to an edge relation between the question type entity type and the answer type entity type defined in the relation graph of the knowledge graph to obtain the knowledge graph for the cloud computing after-sales service.

Fig. 10 is a schematic structural diagram of an electronic device according to an exemplary embodiment of the present application. As shown in fig. 10, the electronic apparatus includes: storage 1001, processor 1002, and communications component 1003.

The memory 1001 is used for storing a computer program and may be configured to store other various data to support operations on the electronic device. Examples of such data include instructions, messages, pictures, videos, etc. for any application or method operating on the electronic device.

The memory 1001 may be implemented by any type of volatile or non-volatile memory device or combination thereof, such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk.

A processor 1002, coupled to the memory 1001, for executing the computer programs in the memory 1001 to: mining entity phrases of original after-sale problems corresponding to the cloud computing service and standard after-sale problems in an after-sale knowledge base to obtain a plurality of problem type entity phrases; determining problem type entity types to which the problem type entity phrases belong based on a plurality of problem type entity types defined in a pre-generated knowledge graph relation graph and by combining attribute information of the problem type entity phrases; selecting at least one answer type entity phrase belonging to the answer type entity type from standard solutions in an after-sales knowledge base according to the answer type entity type defined in the knowledge map relational graph; adding edges between the plurality of question type entity phrases and the at least one answer type entity phrase according to edge relations between the question type entity types and the answer type entity types defined in the knowledge graph relation graph to obtain a knowledge graph for cloud computing after-sales service.

In an alternative embodiment, the processor 1002, when determining the question entity types to which the question entity phrases belong, is specifically configured to: dividing problem entity phrases expressing the same or similar meanings into a group according to the meaning expressions of a plurality of problem entity phrases to obtain a plurality of groups of problem entity phrases; determining entity type discrimination information corresponding to each group of problem type entity phrases according to standard after-sales problems of each group of problem type entity phrases in an after-sales knowledge base and the occurrence frequency of standard solutions respectively; and determining the problem type entity type to which each group of problem type entity phrases belongs according to the entity type distinguishing degree information corresponding to each group of problem type entity phrases.

Further optionally, when the processor 1002 obtains a plurality of sets of problem entity phrases, it is specifically configured to: inputting a plurality of problem type entity phrases into a language representation model for vectorization processing to obtain vector expressions of the problem type entity phrases; and performing multi-turn density clustering on the plurality of problem type entity phrases according to the vector expressions of the plurality of problem type entity phrases to obtain a plurality of groups of problem type entity phrases.

Further optionally, when determining the entity type discrimination information corresponding to each group of problem entity phrases, the processor 1002 is specifically configured to: for each group of problem type entity phrases, the occurrence frequency of the group of problem type entity phrases in the standard after-sale problem is used as first entity type distinguishing degree information, and the ratio of the occurrence frequency of the group of problem type entity phrases in the standard after-sale problem and the occurrence frequency of the group of problem type entity phrases in the standard solution is calculated and used as second entity type distinguishing degree information.

Further optionally, the plurality of problem entity types include: an entity type of the problem expressive type and an entity type of the problem descriptive type. Based on this, when determining the entity type distinction degree information corresponding to each group of problem type entity phrases, the processor 1002 is specifically configured to: for each group of problem entity phrases, determining that the group of problem entity phrases belong to the entity type of the problem expression type under the condition that the first entity type discrimination information is greater than a set frequency threshold and the second entity type discrimination information is greater than a set proportion threshold; and determining the entity type of the question entity phrase group belonging to the question expression type under the condition that the first entity type discrimination information is less than or equal to a set frequency threshold or the second entity type discrimination information is less than or equal to a set proportion threshold.

Further, the entity type of the problem expression type includes at least a problem phenomenon, which is an entity type for describing a key phenomenon exhibited by an after-market problem. Based on this, the processor 1002, when determining that the set of problem entity phrases belongs to the entity types of the problem expression, is specifically configured to: and under the condition that the first entity type discrimination information is greater than a set frequency threshold and the second entity type discrimination information is greater than a set proportion threshold, determining the problem entity type to which the group of problem entity phrases belongs as a problem phenomenon.

Further, the entity types of the problem description type at least include: detail item, error entry and system version; the detail item is an entity type of detail information for describing the after-sales problem, the error item is an entity type of error information occurring in the after-sales problem, and the system version is an entity type of operating system information for describing the after-sales problem. Based on this, the processor 1002, when determining that the set of problem entity phrases belongs to the entity types of the problem expression, is specifically configured to: determining the problem type entity type to which the group of problem type entity phrases belongs as a system version under the condition that the first entity type discrimination information is greater than a set frequency threshold and the second entity type discrimination information is less than or equal to a set proportion threshold; determining the problem type entity type of the group of problem type entity phrases as an error report item under the condition that the first entity type discrimination information is less than or equal to a set frequency threshold and the problem type entity phrases appear in an error report characteristic rule table; and determining the problem type entity type to which the group of problem type entity phrases belong as a detail item under the condition that the first entity type discrimination information is less than or equal to a set frequency threshold and the problem type entity phrases do not appear in the error reporting characteristic rule table.

In an alternative embodiment, the processor 1002, when selecting at least one answer-type entity phrase belonging to an answer-type entity type, is specifically configured to: combining problem type entity phrases belonging to different problem type entity types to obtain a plurality of entity phrase combinations, and counting the co-occurrence frequency of each entity phrase combination in the original after-sale problem and the after-sale knowledge base respectively; and selecting at least one standard solution from the standard solutions in the after-sales knowledge base as at least one answer type entity phrase according to the co-occurrence frequency of each entity phrase combination in the original after-sales problem and the after-sales knowledge base respectively.

Further optionally, the processor 1002, when selecting at least one standard solution from the standard solutions in the after-sales knowledge base as the at least one answer-type entity phrase, is specifically configured to: for each entity phrase combination, if the co-occurrence frequency of the entity phrase combination in the original after-sales problem is greater than the first co-occurrence frequency threshold and the co-occurrence frequency of the entity phrase combination in the after-sales knowledge base is greater than the second co-occurrence frequency threshold, the standard solution co-occurring by the entity phrase combination in the after-sales knowledge base is taken as the answer type entity phrase corresponding to the entity phrase combination.

In an alternative embodiment, the processor 1002, when obtaining the knowledge-graph for the cloud computing after-market service, is specifically configured to: adding edges among all question type entity phrases in each entity phrase combination corresponding to the answer type entity phrases according to an edge relation between the question type entity types defined in the knowledge graph relation graph; and adding edges between each question type entity phrase in the entity phrase combination and the answer type entity phrase corresponding to the entity phrase combination according to an edge relation between the question type entity type and the answer type entity type defined in the relation graph of the knowledge graph to obtain the knowledge graph for the cloud computing after-sales service.

Further, as shown in fig. 10, the electronic device further includes: a display 1004, a power component 1005, and audio components 1006. Only some of the components are schematically shown in fig. 10, and the electronic device is not meant to include only the components shown in fig. 10. In addition, the components within the dashed line frame in fig. 10 are optional components, not necessary components, and may be determined according to the product form of the electronic device. The electronic device of this embodiment may be implemented as a terminal device such as a desktop computer, a notebook computer, a smart phone, or an IOT device, or may be a server device such as a conventional server, a cloud server, or a server array. If the electronic device of this embodiment is implemented as a terminal device such as a desktop computer, a notebook computer, a smart phone, etc., the electronic device may include components within a dashed line frame in fig. 10; if the electronic device of the present embodiment is implemented as a server device such as a conventional server, a cloud server, or a server array, the components in the dashed box in fig. 10 may not be included.

It should be noted that, the embodiment of the present application also provides another electronic device, which has the same or similar structure as the electronic device shown in fig. 10, and therefore is not shown in the drawings. The electronic device of the present embodiment comprises at least a memory and a processor, the processor executing a computer program stored in the processor for: generating a plurality of problem type entity types according to the analysis result of the after-sale problems of the cloud computing service, wherein different problem type entity types are used for describing different types of after-sale problems or describing the after-sale problems from different dimensions; generating an answer type entity type, wherein the answer type entity type is used for describing a solution for the after-sales problem; determining edge relations among the multiple problem type entity types and the answer type entity types according to the incidence relations among the after-sales problems described by the multiple problem type entity types and the requirement information of the after-sales problems on the solution; and generating a knowledge graph relation graph according to the multiple question type entity types, the answer type entity types and the edge relations, wherein the knowledge graph relation graph is used for generating a knowledge graph required by the cloud computing after-sales service.

In an alternative embodiment, the processor is further configured to: and acquiring a cloud computing product related to the after-sale questions described by the plurality of question type entity types, and using the cloud computing product as attribute information corresponding to the question type entity types and/or answer type entity types.

In an optional embodiment, the processor is specifically configured to: determining edge relations between problem phenomena and diagnosis items, between error reporting items and problem phenomena, between fuzzy problems and problem phenomena and between fuzzy problems and error reporting items according to the incidence relations between after-sale problems described by the multiple problem type entity types; determining that edge relations exist between problem phenomena and solutions, between detail items and solutions, between system versions and solutions, between diagnosis items and solutions, and between error reporting items and solutions according to requirement information of after-sale problems described by the multiple problem type entity types on the solutions; wherein the solution is a name of the answer-type entity type.

Accordingly, the present application further provides a computer-readable storage medium storing a computer program, where the computer program can implement the steps in the method embodiment shown in fig. 1 when executed.

Accordingly, the present application also provides a computer readable storage medium storing a computer program, where the computer program can implement the steps in the method embodiments shown in fig. 4 to 7 when executed.

The communications component of fig. 10 described above is configured to facilitate communications between the device in which the communications component is located and other devices in a wired or wireless manner. The device where the communication component is located can access a wireless network based on a communication standard, such as a WiFi, a 2G, 3G, 4G/LTE, 5G and other mobile communication networks, or a combination thereof. In an exemplary embodiment, the communication component receives a broadcast signal or broadcast related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communication component further includes a Near Field Communication (NFC) module to facilitate short-range communications. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, Ultra Wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

The display in fig. 10 described above includes a screen, which may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive an input signal from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation.

The power supply assembly of fig. 10 described above provides power to the various components of the device in which the power supply assembly is located. The power components may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device in which the power component is located.

The audio component of fig. 10 described above may be configured to output and/or input an audio signal. For example, the audio component includes a Microphone (MIC) configured to receive an external audio signal when the device in which the audio component is located is in an operational mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signal may further be stored in a memory or transmitted via a communication component. In some embodiments, the audio assembly further comprises a speaker for outputting audio signals.

As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention 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, and the like) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may 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 means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.

Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A knowledge graph construction method for cloud computing is characterized by comprising the following steps:

mining entity phrases of original after-sale problems corresponding to the cloud computing service and standard after-sale problems in an after-sale knowledge base to obtain a plurality of problem type entity phrases;

determining question type entity types to which the question type entity phrases belong based on a plurality of question type entity types defined in a pre-generated knowledge graph relation graph and in combination with attribute information of the question type entity phrases;

selecting at least one answer type entity phrase belonging to the answer type entity type from standard solutions in an after-sales knowledge base according to the answer type entity type defined in the knowledge map relational graph;

adding edges among the plurality of question type entity phrases and the at least one answer type entity phrase according to edge relations among the question type entity types and between the question type entity types and the answer type entity types defined in the knowledge graph relation graph to obtain a knowledge graph for cloud computing after-sales service.

2. The method of claim 1, wherein determining the question entity types to which the question entity phrases belong based on a plurality of question entity types defined in a pre-generated knowledge-graph relationship graph in combination with attribute information of the question entity phrases comprises:

dividing problem entity phrases expressing the same or similar meanings into a group according to the meaning expressions of the problem entity phrases to obtain a plurality of groups of problem entity phrases;

determining entity type discrimination information corresponding to each group of problem type entity phrases according to standard after-sales problems of each group of problem type entity phrases in an after-sales knowledge base and the occurrence frequency of standard solutions respectively;

and determining the problem type entity type to which each group of problem type entity phrases belongs according to the entity type distinguishing degree information corresponding to each group of problem type entity phrases.

3. The method of claim 2, wherein dividing the question entity phrases expressing the same or similar meanings into a group according to the meaning expressions of the question entity phrases to obtain a plurality of groups of question entity phrases, comprises:

inputting the question entity phrases into a language representation model for vectorization to obtain vector expressions of the question entity phrases;

and performing multi-turn density clustering on the plurality of question type entity phrases according to the vector expressions of the plurality of question type entity phrases to obtain a plurality of groups of question type entity phrases.

4. The method of claim 2, wherein determining entity type discrimination information corresponding to each group of problem type entity phrases according to the frequency of occurrence of each group of problem type entity phrases in standard after-market problems and standard solutions in an after-market knowledge base, respectively, comprises:

for each group of problem type entity phrases, the occurrence frequency of the group of problem type entity phrases in the standard after-sale problem is used as first entity type distinguishing degree information, and the ratio of the occurrence frequency of the group of problem type entity phrases in the standard after-sale problem and the occurrence frequency of the group of problem type entity phrases in the standard solution is calculated and used as second entity type distinguishing degree information.

5. The method of claim 4, wherein the plurality of problem entity types comprises: an entity type of the problem expression type and an entity type of the problem description type; determining the question type entity type of each group of question type entity phrases according to the entity type distinguishing information corresponding to each group of question type entity phrases, wherein the step of determining the question type entity type of each group of question type entity phrases comprises the following steps:

for each group of problem entity phrases, determining that the group of problem entity phrases belong to the entity type of the problem expression type under the condition that the first entity type discrimination information is greater than a set frequency threshold and the second entity type discrimination information is greater than a set proportion threshold;

and determining the entity type of the question entity phrase group belonging to the question expression type under the condition that the first entity type discrimination information is less than or equal to a set frequency threshold or the second entity type discrimination information is less than or equal to a set proportion threshold.

6. The method of claim 2, wherein selecting at least one answer-based entity phrase belonging to the answer-based entity type from standard solutions in an after-market knowledge base based on answer-based entity types defined in the knowledge-graph relationship graph comprises:

combining problem type entity phrases belonging to different problem type entity types to obtain a plurality of entity phrase combinations, and counting the co-occurrence frequency of each entity phrase combination in the original after-sale problem and the after-sale knowledge base respectively;

selecting at least one standard solution from standard solutions in an after-sales knowledge base as the at least one answer-type entity phrase according to the co-occurrence frequency of each entity phrase combination in the original after-sales problem and the after-sales knowledge base respectively.

7. The method of claim 6, wherein selecting at least one standard solution from standard solutions in an after-market knowledge base as the at least one answer-type entity phrase according to a co-occurrence frequency of each entity phrase combination in the original after-market question and the after-market knowledge base, respectively, comprises:

for each entity phrase combination, if the co-occurrence frequency of the entity phrase combination in the original after-sales problem is greater than the first co-occurrence frequency threshold and the co-occurrence frequency of the entity phrase combination in the after-sales knowledge base is greater than the second co-occurrence frequency threshold, the standard solution co-occurring by the entity phrase combination in the after-sales knowledge base is taken as the answer type entity phrase corresponding to the entity phrase combination.

8. The method of claim 7, wherein adding edges between the plurality of question entity phrases and the at least one answer entity phrase according to edge relationships between question entity types and answer entity types defined in the knowledge-graph relationship graph to obtain a knowledge-graph for cloud computing after-market services, comprises:

adding edges among all question type entity phrases in each entity phrase combination corresponding to the answer type entity phrases according to the edge relation between the question type entity types defined in the knowledge graph relation graph; and adding edges between the question type entity phrases in the entity phrase combination and answer type entity phrases corresponding to the entity phrase combination according to the edge relation between the question type entity type and the answer type entity type defined in the knowledge graph relation graph to obtain the knowledge graph for the cloud computing after-sales service.

9. A method for generating a knowledge-graph relationship diagram for cloud computing is characterized by comprising the following steps:

generating a plurality of problem type entity types according to the analysis result of the after-sale problems of the cloud computing service, wherein different problem type entity types are used for describing different types of after-sale problems or describing the after-sale problems from different dimensions;

generating an answer type entity type, wherein the answer type entity type is used for describing a solution for the after-sales problem;

determining edge relations among the multiple problem type entity types and between the multiple problem type entity types and the answer type entity type according to the incidence relation among the after-sales problems described by the multiple problem type entity types and the requirement information of the after-sales problems for the solution;

and generating a knowledge graph relation graph according to the multiple question type entity types, the answer type entity types and the edge relation, wherein the knowledge graph relation graph is used for generating a knowledge graph required by cloud computing after-sales service.

10. The method of claim 9, when generating the knowledge-graph relationship graph, further comprising:

and acquiring a cloud computing product related to the after-sale questions described by the plurality of question type entity types, and using the cloud computing product as attribute information corresponding to the question type entity types and/or answer type entity types.

11. The method of claim 9 or 10, wherein the plurality of problem entity types comprises: an entity type of at least one problem expressive type and an entity type of at least one problem descriptive type; the entity type of each problem expression type is used for describing one type of after-sales problems, and the entity type of each problem expression type is used for describing the detailed information of the after-sales problems from one dimension;

the entity types of the at least one question expression include: problematic phenomena and fuzzy problems; the problem phenomenon is an entity type used for describing a key phenomenon shown by an after-sale problem, and the fuzzy problem is an entity type used for describing an incomplete or ambiguous after-sale problem;

the at least one problem description type entity type includes: detail item, error report item, diagnostic item, and system version; the detail item is an entity type of detail information for describing the after-sales problem, the error item is an entity type of error information occurring in the after-sales problem, the diagnosis item is an entity type for describing diagnosis of the after-sales problem and obtaining diagnosis information, and the system version is an entity type for describing information of an operating system or a database related to the after-sales problem.

12. The method of claim 11, wherein determining the relationships between the plurality of problem type entity types and the answer type entity type according to the association between the after-sales questions described by the plurality of problem type entity types and the requirement information of the after-sales questions for the solution comprises:

determining edge relations between problem phenomena and diagnosis items, between error reporting items and problem phenomena, between fuzzy problems and problem phenomena and between fuzzy problems and error reporting items according to the incidence relations between after-sale problems described by the multiple problem type entity types;

determining that edge relations exist between problem phenomena and solutions, between detail items and solutions, between system versions and solutions, between diagnosis items and solutions, and between error reporting items and solutions according to requirement information of after-sale problems described by the multiple problem type entity types on the solutions; wherein the solution is a name of the answer-type entity type.

13. An electronic device, comprising: a memory and a processor; the memory is for a computer program, and the processor is coupled to the memory for executing the computer program for implementing the steps in the method of any one of claims 1-12.

14. A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, causes the processor to carry out the steps of the method of any one of claims 1-12.