Question and answer credibility determination method and device based on knowledge graph, equipment and medium

Abstract

The application relates to the technical field of Internet, and particularly discloses a knowledge graph-based question and answer credibility determination method and device, equipment and medium. The method comprises the following steps: obtaining consultation information of a user; extracting a consultation triple in the consultation information; obtaining answer information corresponding to the consultation information; extracting an answer triple from the answer information according to the consultation triple; obtaining the number of answer triples that are consistent with a preset knowledge graph in terms of an answer entity, attribute information of the answer entity and relationship information between the answer entities, to obtain a first answer triple number; obtaining the number of any two of the answer entity, the attribute information of the answer entity and the relationship between the answer entities that are consistent with the knowledge graph, to obtain a second answer triple number; and determining the credibility of the answer triple according to the first answer triple number and the second answer triple number. The application facilitates the user to judge whether the current question and answer result meets the user's demand according to the credibility.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of Internet, and particularly relates to a knowledge graph-based question and answer credibility determination method and device, equipment and medium. BACKGROUND

[0002] The insurance industry is a highly professional industry. In the insurance industry, users often do not understand or know the specific insurance clauses involved in the insurance products they purchase, and these insurance clauses are highly professional, and their semantic interpretation is often complex. Therefore, some users often consult insurance companies before, during or after purchasing insurance products.

[0003] In recent years, insurance companies have also provided good services in product consulting services. In particular, many insurance companies provide professional insurance question and answer systems, and users can consult related insurance questions using the insurance question and answer system. In addition, based on the insurance question and answer system, a manual service port has also been developed to facilitate user consultation.

[0004] However, users themselves do not have professional insurance knowledge and cannot evaluate the credibility of the question and answer results, that is, they cannot know whether the question and answer results are good or bad, whether they are professional, etc. Therefore, how to evaluate the credibility of the professional question and answer results in the insurance industry to facilitate users to know the credibility of the question and answer results at the first time has become a problem to be solved. SUMMARY

[0005] The present application provides a knowledge graph-based question and answer credibility determination method, device, equipment and medium to facilitate users to know the credibility of the question and answer results at the first time.

[0006] In a first aspect, the present application provides a knowledge graph-based question and answer credibility determination method, which comprises:

[0007] Obtaining the consultation information of a user;

[0008] Extracting the consultation triple in the consultation information; wherein the consultation triple comprises a consultation entity, attribute information of the consultation entity, and relationship information between the consultation entities;

[0009] Obtaining answer information corresponding to the consultation information;

[0010] According to the consultation triple, extracting an answer triple from the answer information; the answer triple comprises an answer entity, attribute information of the answer entity, and relationship information between the answer entities;

[0011] The first answer triple quantity is obtained by acquiring the quantity of answer triples that are consistent with the preset knowledge graph in terms of the answer entity, attribute information of the answer entity, and relationship information between the answer entities.

[0012] The credibility of the answer triples is determined according to the first answer triple quantity and the second answer triple quantity.

[0013] In the above scheme, the second answer triple quantity is obtained by acquiring the quantity of triples that are consistent with the knowledge graph in terms of any two of the entity, attribute information of the entity, and relationship between the entities; and the credibility of the answer triples is obtained by using the first answer triple quantity and the second answer triple quantity, so that the user can judge whether the current question and answer result meets the own demand according to the credibility.

[0014] In a second aspect, the present application further provides a question and answer credibility determination device based on a knowledge graph, which comprises:

[0015] The consultation information acquisition module is configured to acquire consultation information of a user.

[0016] The consultation triple extraction module is configured to extract a consultation triple in the consultation information, wherein the consultation triple comprises a consultation entity, attribute information of the consultation entity, and relationship information between the consultation entities.

[0017] The answer information acquisition module is configured to acquire answer information corresponding to the consultation information.

[0018] The answer triple extraction module is configured to extract an answer triple from the answer information according to the consultation triple, wherein the answer triple comprises an answer entity, attribute information of the answer entity, and relationship information between the answer entities.

[0019] The first quantity module is configured to acquire the quantity of answer triples that are consistent with a preset knowledge graph in terms of the answer entity, attribute information of the answer entity, and relationship information between the answer entities, and obtain a first answer triple quantity.

[0020] The second quantity module is configured to acquire the quantity of any two of the answer entity, attribute information of the answer entity, and relationship between the answer entities that are consistent with the knowledge graph, and obtain a second answer triple quantity.

[0021] The credibility evaluation module is configured to determine the credibility of the answer triples according to the first answer triple quantity and the second answer triple quantity.

[0022] In a third aspect, the present application also provides a computer device, comprising a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program and realize the knowledge graph-based question and answer credibility determination method as described above when executing the computer program.

[0023] In a fourth aspect, the present application also provides a computer readable storage medium, which stores a computer program, and the computer program makes the processor realize the knowledge graph-based question and answer credibility determination method as described above when executed by the processor.

[0024] The present application discloses a knowledge graph-based question and answer credibility determination method and device, computer device and storage medium, comprising: obtaining the consultation information of a user; extracting the consultation triple in the consultation information; wherein the consultation triple comprises consultation entities, attribute information of the consultation entities and relationship information between the consultation entities; obtaining the answer information corresponding to the consultation information; extracting the answer triple from the answer information according to the consultation triple; the answer triple comprises answer entities, attribute information of the answer entities and relationship information between the answer entities; obtaining the number of answer triples in which the answer entities, the attribute information of the answer entities and the relationship information between the answer entities are consistent with the preset knowledge graph, to obtain the first number of answer triples; obtaining the number of any two of the answer entities, the attribute information of the answer entities and the relationship information between the answer entities that are consistent with the knowledge graph, to obtain the second number of answer triples; and determining the credibility of the answer triple according to the first number of answer triples and the second number of answer triples. The present application can provide the credibility of the answer information to the user, and the user can determine whether the current answer information is accurate and reliable according to the credibility. BRIEF DESCRIPTION OF DRAWINGS

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0026] Figure 1 The knowledge graph-based question and answer credibility determination system provided for the embodiments of the present application;

[0027] Figure 2 The schematic flow chart of the knowledge graph-based question and answer credibility determination method provided for the embodiments of the present application;

[0028] Figure 3 is Figure 2 the sub-step schematic flow chart of the knowledge graph-based question and answer credibility determination method in

[0029] Figure 4 A schematic block diagram of a knowledge graph-based question and answer credibility determination device provided for an embodiment of the present application is provided.

[0030] Figure 5 A schematic block diagram of another knowledge graph-based question and answer credibility determination device provided for an embodiment of the present application is provided.

[0031] Figure 6 A structural schematic block diagram of a computer device provided for an embodiment of the present application is provided. DETAILED DESCRIPTION

[0032] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0033] The flowchart shown in the drawings is only an example and does not necessarily include all the contents and operations / steps, nor does it necessarily need to be executed in the order described. For example, some operations / steps can be further divided, combined or partially combined, so the actual execution order may be changed according to the actual situation.

[0034] It should be understood that the terms used in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification and the appended claims of the present application, unless otherwise clear from the context, the singular forms "a", "an" and "the" are intended to include the plural forms.

[0035] It should also be understood that the term "and / or" used in the specification and the appended claims of the present application means any combination of one or more of the associated listed items and all possible combinations, and includes these combinations.

[0036] Embodiments of the present application provide a knowledge graph-based question and answer credibility determination method, device, computer device and storage medium. The knowledge graph-based question and answer credibility determination method can be applied to a server to evaluate the credibility of answers and provide a reference standard for users. The server can be a standalone server or a server cluster.

[0037] Some embodiments of the present application will be described in detail below with reference to the accompanying drawings. The following embodiments and features in the embodiments can be combined with each other without conflict.

[0038] As shown in Figure 1 Figure 1 A knowledge graph-based question and answer credibility determination system is provided for the embodiments of the present application, which comprises a terminal, a background end and a server, and the terminal, the background end and the server are in communication connection.

[0039] The terminal comprises mobile phones, tablet computers, notebook computers, desktop computers, personal digital assistants, wearable devices and other electronic devices.

[0040] The server comprises a single server or a server cluster.

[0041] In the following, the knowledge graph-based question and answer credibility determination method provided by the embodiments of the present application will be described in detail based on the knowledge graph-based question and answer credibility determination system.

[0042] As shown in Figure 2 Figure 2 A schematic flow chart of the knowledge graph-based question and answer credibility determination method provided for the embodiments of the present application; the knowledge graph-based question and answer credibility determination method specifically comprises steps S101 to S106.

[0043] S101, obtaining the consultation information of a user.

[0044] In the insurance field, because the insurance problem is too professional, there are many related insurance types and clauses. Although the staff will introduce the insurance types and corresponding clauses in detail when signing the relevant insurance agreement. However, afterwards, there are still many users who consult related problems again.

[0045] Most of the time, the user will consult online through the relevant APP on the terminal. For example, the user can consult the clause meaning, insurance amount, etc. through the "Ping'an Jin Guan Jia" APP on the mobile phone, and these consultation information will be recorded by the background end. The server can directly obtain the consultation information of the user.

[0046] In summary, in the embodiments of the present application, the server can obtain the consultation information sent by the user to the background end through the terminal.

[0047] S102, extracting a consultation triple in the consultation information; wherein the consultation triple comprises a consultation entity, attribute information of the consultation entity, and relationship information between the consultation entities.

[0048] The consultation entity generally refers to the specific transaction in the consultation information, including insurance types, concept entities and their attributes, insured persons, beneficiary entities, category entities, etc. For example, when the user consults: "A wants to know how much the annual premium of accidental insurance is?", the consultation entity can include "A" and "accidental insurance".

[0049] ​​For example, if the user rides a bicycle to the destination and finds that the bicycle cannot be locked, the user can feed back to the background through the corresponding APP on the mobile phone: "the bicycle ridden by B cannot be locked". At this time, the consulting entity can be "B" and "bicycle".

[0050] The attribute data of the consultation refers to the nature, relationship, description, explanation, and specific numerical value of the consulting entity. The attribute data points from an entity to its attribute value. For example, when the user consults: "does A want to know that the annual premium of accident insurance is 3000?", the attribute data of the consulting entity is the premium and 3000. Or, when the user consults: "A is 28 years old this year, how much is the annual premium of accident insurance?", the attribute data of the consulting entity "A" is "28 years old"; and the attribute data of the consulting entity "accident insurance" is "premium".

[0051] The relationship information between the consulting entities refers to the association between different consulting entities in the consultation information. For example, the user consults: "A is 28 years old this year, and has bought accident insurance, how much is the premium of the accident insurance?", the consulting entities include "A" and "accident insurance", and the relationship information between the two consulting entities is "A has bought accident insurance".

[0052] The consulting triple, i.e., the consulting entity, the attribute information of the consulting entity, and the relationship information between the consulting entities, represents knowledge in a certain form. By extracting the consulting triple from the consultation information, the key information of the user's consultation can be clearly grasped.

[0053] Specifically, as shown in Figure 3 , the sub-step schematic flowchart of the knowledge graph-based question and answer credibility determination method in Figure 3 is Figure 2 . The extraction of the consulting triple in the consultation information can include the following steps:

[0054] S1021, extracting the consulting entity of the consultation information and the attribute information of the consulting entity.

[0055] S1022, determining the relationship information between the consulting entities according to the consultation information, and obtaining a plurality of answer triples related to the consultation information.

[0056] In an embodiment of the present application, the relationship information between the consulting entities can be obtained according to the consultation information. For example, when the user consults: "does A want to buy accident insurance with a premium of 3000?", the consulting entities include A and accident insurance; and the relationship information between the two consulting entities is "A wants to buy the accident insurance".

[0057] In an embodiment of the present application, the relationship information of the consulting entity can also be determined according to the consulting entity and the attribute data. For example, when the user consults, "How much is the premium of the accident insurance purchased by A?", two consulting entities, "A" and "accident insurance", are included, and the attribute data includes "premium". At this time, whether A wants to purchase accident insurance or has already purchased accident insurance, it can be inferred that "A wants to know the premium price of the accident insurance", that is, the relationship information between the two consulting entities is "A wants to know the premium price of the accident insurance".

[0058] The method for constructing the knowledge graph can comprise: using a web crawler technology to crawl professional knowledge in a field required by a user to obtain the question and answer data sample; and / or using a pre-stored text database to obtain the question and answer data sample, the question and answer data sample at least comprising a consulting data sample and an answer data sample corresponding to the consulting data sample.

[0059] Further, the question and answer data sample can be entity-labeled to obtain an entity-labeled training set. Then, a deep network learning model is used to extract entities of the question and answer data sample, attribute information of the entities, and relationship information between the entities from the entity-labeled training set; and a knowledge graph represented by a triple is constructed according to the entities, the attribute information of the entities, and the relationship information between the entities.

[0060] S103, obtaining answer information corresponding to the consulting information.

[0061] In most cases, the user will consult online through a related APP, and whether the intelligent robot answers online or the staff in the background answers online, the server can directly obtain the answer content to obtain the answer information.

[0062] For example, the user can consult the clause meaning, insurance amount, etc. through the "Ping'an Jin Guan Jia" APP of the terminal to the background, and then the intelligent robot built in the "Ping'an Jin Guan Jia" can automatically reply to the related question, or the staff in the background can help answer online. In this case, the server in the background can directly obtain the answer content to obtain the answer information.

[0063] It should be noted that when providing the answer information, the attributes of the consulting entity can not be limited, and more important attributes can also be considered. For example, for the premium, social security and gender are more important attributes, and when the user consults, "How much is the premium of the accident insurance purchased by Xiaoming?", the corresponding answer information can include: Xiaoming, 35-40 years old, male, and the premium of the accident insurance is 4000 yuan.

[0064] S104, extracting an answer triple from the answer information according to the consultation triple; the answer triple includes an answer entity, attribute information of the answer entity, and relationship information between the answer entities.

[0065] It can be understood that the answer triple is obtained according to the consultation triple.

[0066] For example, taking the consultation information "Xiaoming is 38 years old this year, and how much does he need to pay for the insurance premium of accidental insurance?" As an example; at this time, the consultation information triple includes consultation entities: Xiaoming, accidental insurance; entity attributes include: 38 years old, insurance premium; the relationship between the consultation entities includes: Xiaoming wants to buy accidental insurance.

[0067] Correspondingly, the answer information can include the following:

[0068] 1. With social security, 35-40 years old, Xiaoming, male, the insurance premium for purchasing accidental insurance is 2500 yuan.

[0069] 2. Without social security, 35-40 years old, Xiaoming, male, the insurance premium for purchasing accidental insurance is 4000 yuan.

[0070] 3. With social security, 35-40 years old, Xiaoming, female, the insurance premium for purchasing accidental insurance is 2500 yuan.

[0071] 4. Without social security, 35-40 years old, Xiaoming, female, the insurance premium for purchasing accidental insurance is 4000 yuan.

[0072] Of course, the answer information can also expand the age stage. That is, the answer information can also include:

[0073] 5. With social security, 30-35 years old, Xiaoming, male, the insurance premium for purchasing accidental insurance is 2200 yuan.

[0074] 6. Without social security, 30-35 years old, Xiaoming, male, the insurance premium for purchasing accidental insurance is 3500 yuan.

[0075] 7. With social security, 30-35 years old, Xiaoming, female, the insurance premium for purchasing accidental insurance is 2400 yuan.

[0076] 8. Without social security, 30-35 years old, Xiaoming, female, the insurance premium for purchasing accidental insurance is 3600 yuan.

[0077] ...

[0078] Of course, it can be understood that the amounts in the above content are exemplary and not actual amounts.

[0079] From the answer information, in the answer triple, the answer entity includes: Xiaoming, accidental insurance; the attribute information of the answer entity includes: age, gender, whether there is social security, and premium.

[0080] S105, obtain the number of answer triples that the answer entity, the attribute information of the answer entity, and the relationship information between the answer entities conform to the preset knowledge graph, to obtain the first number of answer triples; obtain the number of answer triples that any two of the answer entity, the attribute information of the answer entity, and the relationship between the answer entities conform to the knowledge graph, to obtain the second number of answer triples, and the remaining is the third number of answer triples.

[0081] S106, determine the credibility of the answer triple according to the first number of answer triples and the second number of answer triples.

[0082] The user can determine whether the current obtained answer information is accurate according to the credibility. Of course, the credibility evaluation method can include: setting the first weight, the second weight, and the third weight, multiplying the first weight by the first number of answer triples, multiplying the second weight by the second number of answer triples, multiplying the third weight by the third number of answer triples, and summing the three multiplication results, and then dividing by the total number of triples.

[0083] Specifically, in the embodiment of the application, the credibility is obtained by the following method:

[0084] Obtain the product of the first number of answer triples and the first preset proportion coefficient to obtain the first answer triple weighted number; obtain the product of the second number of answer triples and the second preset proportion coefficient to obtain the second answer triple weighted number; multiply the sum of the first answer triple weighted number and the second answer triple weighted number by the third preset proportion coefficient and the third number of answer triples to obtain the sum of the third triple weighted number, to obtain the total weighted number of answer triples; obtain the proportion of the total weighted number of answer triples and the total number of answer triples to obtain the credibility of the answer triple.

[0085] Considering the accuracy of the answer, in the embodiment, the first preset proportion coefficient is preferably 1, the second preset proportion coefficient is preferably 2 / 3, and the third preset proportion coefficient is preferably 0.

[0086] Specifically, assuming that N answer triples can be extracted from an answer information, N is a positive integer. Assuming that the number of answer triples that the answer entity, the attribute information of the answer entity, and the relationship information between the answer entities can all conform to the knowledge graph is N1, the number of answer triples that any two of the answer entity, the attribute information of the answer entity, and the relationship information between the answer entities conform to the knowledge graph is N2, and the remaining is N3; then the credibility calculation formula of a question and answer text is as follows:

[0087]

[0088] Wherein, Credibility is the credibility, the greater the value of the credibility, the higher the credibility.

[0089] Through the above scheme, when the APP on the terminal feeds back the answer information to the user, the result with the credibility evaluation is automatically brought. The user can judge whether the current answer information is reliable and accurate according to the value of the credibility result.

[0090] Based on the field knowledge graph, the credibility of the insurance question and answer is calculated, and the credibility is dependent on the constructed knowledge graph. The more professional and real the constructed knowledge graph is, the higher the text credibility calculated based on it is. Therefore, the knowledge graph needs a positive feedback method for continuous updating. Therefore, in an embodiment of the present application, the question and answer credibility determination method based on the knowledge graph further comprises:

[0091] Receiving user scoring information on the answer information; according to the scoring information, optimizing the knowledge graph.

[0092] The user can evaluate the answer information of the customer through the APP of the terminal.

[0093] Further, according to the scoring information, the knowledge graph is optimized, comprising:

[0094] According to the scoring information, an evaluation score of the answer information is obtained; when the evaluation score of the answer information is less than a preset threshold, the answer information is discarded.

[0095] Wherein, the preset threshold can be determined according to the actual situation.

[0096] The method of calculating the text credibility based on the knowledge graph can be used when the artificial customer service or intelligent customer service answers the insurance professional knowledge for the customer. When the credibility of the customer service answer content is calculated and pushed to the customer, the user can evaluate its credibility. Based on the user's evaluation of the text credibility, the knowledge graph is corrected.

[0097] Suppose that in the time interval of the knowledge graph update node, the system receives N text credibility evaluations, and the evaluation full score is 10, then the credibility evaluation value of each text is n. Wherein the average evaluation score of each answer information in the credibility evaluation feedback is as follows:

[0098]

[0099] Wherein, n i represents the text credibility customer evaluation value containing a certain answer information; Z irepresents the text credibility of the answer information. When the text credibility of a certain answer information is very high, it means that the answer information is reliable. i When the text credibility of a certain answer information is very high, it means that the answer information is reliable.

[0100] When the text credibility of a certain answer information is very high, it means that the answer information is reliable.

[0101] Please refer to Figure 4 、 Figure 5 , Figure 4 is a schematic block diagram of a question and answer credibility determination device based on a knowledge graph provided by an embodiment of the present application, Figure 5 is a schematic block diagram of another question and answer credibility determination device based on a knowledge graph provided by an embodiment of the present application; the question and answer credibility determination device based on a knowledge graph is used to execute the question and answer credibility determination method based on a knowledge graph described above. Wherein, the question and answer credibility determination device based on a knowledge graph can be configured in a server.

[0102] The question and answer credibility determination device based on a knowledge graph 300 comprises: a consultation information acquisition module 301, a consultation triple extraction module 302, an answer information acquisition module 303, an answer triple extraction module 304, a first quantity module 305, a second quantity module 306, and a credibility evaluation module 307.

[0103] The consultation information acquisition module 301 is configured to acquire the consultation information of a user.

[0104] The consultation triple extraction module 302 is configured to extract consultation triples from the consultation information; wherein, the consultation triples comprise consultation entities, attribute information of the consultation entities, and relationship information between the consultation entities.

[0105] The answer information acquisition module 303 is configured to acquire answer information corresponding to the consultation information.

[0106] The answer triple extraction module 304 is configured to extract answer triples from the answer information according to the consultation triples; wherein, the answer triples comprise answer entities, attribute information of the answer entities, and relationship information between the answer entities.

[0107] The first quantity module 305 is configured to acquire the number of answer triples in which the answer entities, the attribute information of the answer entities, and the relationship information between the answer entities are consistent with a preset knowledge graph, to obtain a first answer triple quantity.

[0108] The second quantity module 306 is configured to obtain a quantity of any two of the answer entity, the attribute information of the answer entity, and the relationship between the answer entities that are consistent with the knowledge graph, to obtain a second answer triple quantity.

[0109] The credibility evaluation module 307 is configured to determine the credibility of the answer triple according to the first answer triple quantity and the second answer triple quantity.

[0110] In an embodiment, the knowledge graph question and answer credibility determination apparatus 300 is further configured to:

[0111] obtain a question and answer data sample, the question and answer data sample at least including a consultation data sample and an answer data sample corresponding to the consultation data sample; extract an entity, attribute information of the entity, and relationship information between the entities of the question and answer data sample by using a deep network learning model; and construct a knowledge graph represented by a triple according to the entity, the attribute information of the entity, and the relationship information between the entities.

[0112] In an embodiment, the knowledge graph question and answer credibility determination apparatus 300 is further configured to: obtain the question and answer data sample by using a web crawler technology to crawl professional knowledge in a field required by a user; and / or obtain the question and answer data sample by using a pre-stored text database.

[0113] In an embodiment, the consultation triple extraction module 302 is further configured to: extract a consultation entity of the consultation information and attribute information of the consultation entity; and determine relationship information between the consultation entities according to the consultation information, to obtain a plurality of answer triples related to the consultation information.

[0114] In an embodiment, the credibility evaluation module 307 is further configured to: obtain a first answer triple weighted quantity by multiplying the first answer triple quantity by a first preset proportionality coefficient; obtain a second answer triple weighted quantity by multiplying the second answer triple quantity by a second preset proportionality coefficient; obtain a total answer triple weighted quantity according to a sum of the first answer triple weighted quantity and the second answer triple weighted quantity; and obtain the credibility of the answer triple by obtaining a proportion of the total answer triple weighted quantity to the total number of answer triples.

[0115] In an embodiment, the knowledge graph question and answer credibility determination apparatus 300 further includes a correction module 308, which is configured to: receive scoring information of the answer triple from a user; and optimize the knowledge graph according to the scoring information.

[0116] In an embodiment, the correction module 308 is further configured to: obtain an evaluation score of the answer triple according to the score information; and discard the answer triple when the evaluation score of the answer triple is less than a preset threshold.

[0117] It should be noted that, for the convenience and brevity of description, the specific working processes of the apparatus and the modules described above can be clearly understood by those skilled in the art, and the corresponding processes in the foregoing method embodiments can be referred to, which will not be described herein.

[0118] The apparatus described above can be implemented in the form of a computer program, which can run on a computer device such as the computer device shown in Figure 6 .

[0119] Please refer to Figure 6 , Figure 6 is a structural schematic block diagram of a computer device provided by an embodiment of the present application. The computer device can be a server.

[0120] Please refer to Figure 6 , the computer device includes a processor, a memory and a network interface connected through a system bus, wherein the memory can include a non-volatile storage medium and an internal memory.

[0121] The non-volatile storage medium can store an operating system and a computer program. The computer program includes program instructions which, when executed, can cause the processor to perform any kind of knowledge graph-based question and answer credibility determination method.

[0122] The processor is configured to provide computing and control capabilities to support the operation of the entire computer device.

[0123] The internal memory provides an environment for the running of the computer program in the non-volatile storage medium, and the computer program, when executed by the processor, can cause the processor to perform any kind of knowledge graph-based question and answer credibility determination method.

[0124] The network interface is configured to perform network communication, such as sending assigned tasks. Those skilled in the art can understand that the structure shown in Figure 6 , is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0125] It should be appreciated that the processor can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0126] In one embodiment, the processor is configured to run a computer program stored in the memory to implement the following steps:

[0127] Obtain the consultation information of the user;

[0128] Extract the consultation triple in the consultation information; the consultation triple includes consultation entities, attribute information of the consultation entities, and relationship information between the consultation entities;

[0129] Obtain the answer information corresponding to the consultation information;

[0130] According to the consultation triple, extract the answer triple from the answer information; the answer triple includes answer entities, attribute information of the answer entities, and relationship information between the answer entities;

[0131] Obtain the number of answer triples that are consistent with the preset knowledge graph in terms of the answer entities, the attribute information of the answer entities, and the relationship between the answer entities, to obtain a first number of answer triples;

[0132] Obtain the number of answer triples that are consistent with the knowledge graph in terms of any two of the answer entities, the attribute information of the answer entities, and the relationship between the answer entities, to obtain a second number of answer triples;

[0133] According to the first number of answer triples and the second number of answer triples, determine the credibility of the answer triple.

[0134] In one embodiment, when implementing the method for determining the credibility of the question and answer of the knowledge graph, the processor is configured to implement:

[0135] In one embodiment, when implementing the method for determining the credibility of the question and answer of the knowledge graph, the processor is configured to implement:

[0136] obtaining a question and answer data sample, the question and answer data sample comprising at least a consultation data sample and an answer data sample corresponding to the consultation data sample;

[0137] extracting, by using a deep network learning model, an entity, attribute information of the entity, and relationship information between the entities of the question and answer data sample;

[0138] constructing a knowledge graph represented by a triple according to the entity, the attribute information of the entity, and the relationship information between the entities.

[0139] In an embodiment, when implementing the method for determining the credibility of a question and answer based on a knowledge graph, the processor is configured to implement: obtaining the question and answer data sample by using a web crawler to crawl professional knowledge in a field required by a user; and / or obtaining the question and answer data sample by using a pre-stored text database.

[0140] In an embodiment, when implementing the method for determining the credibility of a question and answer based on a knowledge graph, the processor is configured to implement: extracting a consultation entity and attribute information of the consultation entity of the consultation information; and determining relationship information between the consultation entities according to the consultation information to obtain a plurality of answer triples related to the consultation information.

[0141] In an embodiment, when implementing the method for determining the credibility of a question and answer based on a knowledge graph, the processor is configured to implement: obtaining the credibility of the answer triple according to the first answer triple quantity and the second answer triple quantity, comprising:

[0142] obtaining a first answer triple weighted quantity by multiplying the first answer triple quantity by a first preset proportionality coefficient; obtaining a second answer triple weighted quantity by multiplying the second answer triple quantity by a second preset proportionality coefficient; obtaining an answer triple total weighted quantity according to a sum of the first answer triple weighted quantity and the second answer triple weighted quantity; and obtaining the credibility of the answer triple by obtaining a proportion of the answer triple total weighted quantity to the total number of answer triples.

[0143] In an embodiment, when implementing the method for determining the credibility of a question and answer based on a knowledge graph, the processor is configured to implement: receiving score information of the answer triple by a user; and optimizing the knowledge graph according to the score information.

[0144] In an embodiment, when implementing the method for determining the credibility of a question and answer based on a knowledge graph, the processor is configured to implement: obtaining an evaluation score of the answer triple according to the score information; and discarding the answer triple when the evaluation score of the answer triple is less than a preset threshold.

[0145] The embodiment of the present application also provides a medium, specifically a computer readable storage medium, which stores a computer program, the computer program comprising program instructions, and the processor executes the program instructions to realize any one of the knowledge graph based question and answer credibility determination methods provided by the embodiments of the present application.

[0146] The computer readable storage medium can be an internal storage unit of the computer device, for example, a hard disk or a memory of the computer device. The computer readable storage medium can also be an external storage device of the computer device, for example, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card and the like.

[0147] The above merely provides a specific implementation of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of various equivalent modifications or replacements within the technical range disclosed by the present application, and these modifications or replacements should be encompassed in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A knowledge graph-based question answer credibility determination method, characterized in that, The method comprises the following steps: obtaining consultation information of a user; extracting a consultation triple from the consultation information; wherein the consultation triple comprises a consultation entity, attribute information of the consultation entity, and relationship information between the consultation entities; obtaining answer information corresponding to the consultation information; extracting an answer triple from the answer information according to the consultation triple; wherein the answer triple comprises an answer entity, attribute information of the answer entity, and relationship information between the answer entities; obtaining a first answer triple quantity of the answer triple that is consistent with a preset knowledge graph in terms of the answer entity, the attribute information of the answer entity, and the relationship information between the answer entities; and obtaining a second answer triple quantity of the answer triple that is consistent with the knowledge graph in terms of any two of the answer entity, the attribute information of the answer entity, and the relationship information between the answer entities; determining a credibility of the answer triple according to the first answer triple quantity and the second answer triple quantity. 2.The knowledge graph based question answering credibility determination method according to claim 1, characterized in that, The method further comprises the following steps: obtaining a question and answer data sample, wherein the question and answer data sample comprises at least a consultation data sample and an answer data sample corresponding to the consultation data sample; extracting an entity, attribute information of the entity, and relationship information between the entities of the question and answer data sample by using a deep network learning model; constructing a knowledge graph represented by triples according to the entity, the attribute information of the entity, and the relationship information between the entities. 3.The knowledge graph based question answering credibility determination method according to claim 2, characterized in that, The step of obtaining the question and answer data sample comprises the following steps: obtaining professional knowledge in a field required by a user by using a web crawler technology to obtain the question and answer data sample; and / or obtaining the question and answer data sample by using a pre-stored text database. 4.The knowledge graph based question answering credibility determination method according to claim 1, characterized in that, The step of extracting the consultation triple from the consultation information comprises the following steps: extracting a consultation entity and attribute information of the consultation entity from the consultation information; determining relationship information between the consultation entities according to the consultation information to obtain a plurality of answer triples related to the consultation information. 5.The knowledge graph based question answering credibility determination method according to claim 1, characterized in that, The step of determining the credibility of the answer triple according to the first answer triple quantity and the second answer triple quantity comprises the following steps: obtaining a first answer triple weighted quantity by multiplying the first answer triple quantity by a first preset proportionality coefficient; obtaining a second answer triple weighted quantity by multiplying the second answer triple quantity by a second preset proportionality coefficient; obtaining an answer triple total weighted quantity according to a sum of the first answer triple weighted quantity and the second answer triple weighted quantity; obtaining the credibility of the answer triple by obtaining a proportion of the answer triple total weighted quantity to a total number of the answer triples. 6.The knowledge graph based question answer credibility determination method according to any one of claims 1 to 5, characterized in that, The method further comprises the following steps: receiving score information of the answer information provided by a user; optimizing the knowledge graph according to the score information. 7.The knowledge graph based question answering credibility determination method according to claim 6, characterized in that, The step of optimizing the knowledge graph according to the score information comprises the following steps: obtaining an evaluation score of the answer information according to the score information; discarding the answer information when the evaluation score of the answer information is less than a preset threshold.

8. A knowledge graph based question answer credibility determination apparatus, characterized in that, The method comprises the following steps: a consultation information obtaining module configured to obtain consultation information of a user; The consultation triple extraction module is configured to extract a consultation triple from the consultation information, wherein the consultation triple comprises a consultation entity, attribute information of the consultation entity, and relationship information between the consultation entities. The answer information acquisition module is configured to acquire answer information corresponding to the consultation information. The answer triple extraction module is configured to extract an answer triple from the answer information according to the consultation triple, wherein the answer triple comprises an answer entity, attribute information of the answer entity, and relationship information between the answer entities. The first quantity module is configured to acquire a quantity of answer triples that are consistent with a preset knowledge graph in terms of the answer entity, the attribute information of the answer entity, and the relationship information between the answer entities, to obtain a first answer triple quantity. The second quantity module is configured to acquire a quantity of any two of the answer entity, the attribute information of the answer entity, and the relationship information between the answer entities that are consistent with the knowledge graph, to obtain a second answer triple quantity. The credibility evaluation module is configured to determine the credibility of the answer triple according to the first answer triple quantity and the second answer triple quantity.

9. A computer device, comprising: The computer device comprises a memory and a processor. The memory is configured to store a computer program. The processor is configured to execute the computer program and implement the knowledge graph-based question and answer credibility determination method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program causes the processor to implement the knowledge graph-based question and answer credibility determination method according to any one of claims 1 to 7 when executed by the processor.

Images