CN111651567A - Service question and answer data processing method and device - Google Patents

Service question and answer data processing method and device Download PDF

Info

Publication number
CN111651567A
CN111651567A CN202010301636.5A CN202010301636A CN111651567A CN 111651567 A CN111651567 A CN 111651567A CN 202010301636 A CN202010301636 A CN 202010301636A CN 111651567 A CN111651567 A CN 111651567A
Authority
CN
China
Prior art keywords
data
weight ratio
service
service type
question
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010301636.5A
Other languages
Chinese (zh)
Other versions
CN111651567B (en
Inventor
林青云
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing QIYI Century Science and Technology Co Ltd
Original Assignee
Beijing QIYI Century Science and Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing QIYI Century Science and Technology Co Ltd filed Critical Beijing QIYI Century Science and Technology Co Ltd
Priority to CN202010301636.5A priority Critical patent/CN111651567B/en
Publication of CN111651567A publication Critical patent/CN111651567A/en
Application granted granted Critical
Publication of CN111651567B publication Critical patent/CN111651567B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/332—Query formulation
    • G06F16/3329—Natural language query formulation or dialogue systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35—Clustering; Classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0281—Customer communication at a business location, e.g. providing product or service information, consulting

Abstract

The embodiment of the invention provides a service question and answer data processing method, which comprises the steps of obtaining a first weight ratio according to the correlation degree between question data and service labels after receiving the question data, and determining a target service type according to the first weight ratio and a second weight ratio among service types. The answer data corresponding to the target service type is used as candidate answer data, the target answer data is determined according to a third weight ratio between the question data and the candidate answer data, and the target service type related to the question data can be determined according to the relation between the service types while the relation between the question data and the service types is considered. Because the target answer data is determined from the candidate answer data corresponding to the target service type, the dimensionality of the target answer data corresponding to the question data is expanded, the depth and the breadth of the answer data can be improved, the accuracy in obtaining the answer data corresponding to the question data is improved, the problem can be solved in a more professional manner, and the requirement of a user for obtaining the answer data is met.

Description

Service question and answer data processing method and device
Technical Field
The present invention relates to the field of data processing technologies, and in particular, to a method and an apparatus for processing service question and answer data.
Background
With the continuous development of cloud computing, big data, AI (Artificial intelligence) and other related technologies, users can solve the puzzles more and more conveniently through retrieval, questioning and other modes on the network. Therefore, in order to save labor cost and improve operation and maintenance efficiency, in the process of communication between merchants, enterprises and the like and users, intelligent customer service is more and more set to deal with some fundamental problems.
However, existing intelligent customer service typically replies to a question with a single-service type, shallow answer analysis, such as for "why did the virtual machine fail? "the question is usually answered" network failure ". In practice, however, there may be further causes behind the "network failure" that simply learning of "network failure" does not help the user to resolve the problem. In the existing intelligent customer service question-answering, when the problem with a complex business background is faced, a specific solution for solving the problem cannot be provided for the user in a targeted manner, and the intelligent customer service question-answering is single in dimension and lacks in professionality.
Disclosure of Invention
The embodiment of the invention aims to provide a business question and answer data processing method and a business question and answer data processing device, so as to expand the depth and the breadth of answer data determined according to received question data and improve the accuracy of the answer data obtained in the question data processing process. The specific technical scheme is as follows:
in a first aspect of the present invention, a method for processing service question and answer data is provided, where the method may include:
receiving problem data;
calculating the association degree between the problem data and a service label to obtain a first weight ratio between the problem data and a service type corresponding to the service label;
determining a target service type corresponding to the problem data according to a first weight ratio between the problem data and the service types and a second weight ratio between the service types, wherein the second weight ratio is obtained by calculating the association degree between the service types;
determining answer data corresponding to the target service type as candidate answer data corresponding to the question data;
and calculating a third weight ratio between the question data and the candidate answer data, and determining target answer data corresponding to the question data according to the third weight ratio.
Optionally, the second weight ratio includes a longitudinal association weight ratio corresponding to different service types having a longitudinal association relationship, and/or a horizontal association weight ratio corresponding to different service types having a horizontal association relationship.
Optionally, the determining, according to a first weight ratio between the problem data and the service type and a second weight ratio between the service types, a target service type corresponding to the problem data includes:
determining a first target service type corresponding to the problem data according to a first weight ratio between the problem data and the service type;
determining a second target service type corresponding to the problem data according to a transverse correlation weight ratio and/or a longitudinal correlation weight ratio between the first target service type and other service types;
and taking the first target service type and the second target service type as target service types corresponding to the problem data.
Optionally, the calculating a degree of association between the question data and a service tag to obtain a first weight ratio between the question data and a service type corresponding to the service tag includes:
extracting problem features of the problem data;
calculating the correlation degree between the problem characteristics and the service label;
and determining the weight of the problem data corresponding to the service label according to the association degree, and obtaining the first weight ratio between the problem data and the service type corresponding to the service label.
Optionally, before receiving the question data, the method further includes:
calculating the association degree between answer data and the service label;
classifying the answer data according to the association degree between the answer data and the service label, and determining the service type corresponding to the answer data.
Optionally, the calculating a third weight ratio between the question data and the candidate answer data includes:
determining the weight of the answer data corresponding to the question data according to the association degree between the answer data and the service label and the association degree between the question data and the service label;
and according to the first weight, the weight of the answer data corresponding to the question data is compared and weighted, and a third weight ratio between the question data and the candidate answer data is calculated.
Optionally, before determining a second target service type corresponding to the problem data according to a horizontal association weight ratio and/or a vertical association weight ratio between the first target service type and another service type, the method includes:
determining an answer preset depth corresponding to the question data, wherein the answer preset depth is used for limiting the number of the second target service types;
the determining a second target service type corresponding to the problem data according to the horizontal correlation weight ratio and/or the vertical correlation weight ratio between the first target service type and other service types includes:
and determining a second target service type corresponding to the question data according to the transverse correlation weight ratio and/or the longitudinal correlation weight ratio between the first target service type and other service types and the preset depth of the answer.
In a second aspect of the present invention, there is also provided a question answer obtaining apparatus, which may include:
the problem data receiving module is used for receiving problem data;
the first weight calculation module is used for calculating the association degree between the problem data and the service label to obtain a first weight ratio between the problem data and the service type corresponding to the service label;
the target service determining module is used for determining a target service type corresponding to the problem data according to a first weight ratio between the problem data and the service types and a second weight ratio between the service types, wherein the second weight ratio is obtained by calculating the association degree between the service types;
a candidate answer determining module, configured to determine that answer data corresponding to the target service type is candidate answer data corresponding to the question data;
and the target answer determining module is used for calculating a third weight ratio between the question data and the candidate answer data and determining target answer data corresponding to the question data according to the third weight ratio.
Optionally, the second weight ratio includes a longitudinal association weight ratio corresponding to different service types having a longitudinal association relationship, and/or a transverse association weight ratio corresponding to different service types having a transverse association relationship;
optionally, the target service determination module includes:
the first target service determining submodule is used for determining a first target service type corresponding to the problem data according to a first weight ratio between the problem data and the service type;
a second target service determining submodule, configured to determine a second target service type corresponding to the problem data according to a horizontal association weight ratio and/or a vertical association weight ratio between the first target service type and another service type;
and the target service determining submodule is used for taking the first target service type and the second target service type as target service types corresponding to the problem data.
Optionally, the first weight calculating module includes:
the characteristic extraction submodule is used for extracting problem characteristics of the problem data;
the correlation calculation submodule is used for calculating the correlation degree between the problem characteristics and the service labels;
and the weight calculation submodule is used for determining the weight of the problem data corresponding to the service tag according to the association degree and obtaining the first weight ratio between the problem data and the service type corresponding to the service tag.
Optionally, the apparatus may further include:
the association calculation module is used for calculating the association degree between the answer data and the service label;
and the service determining module is used for classifying the answer data according to the association degree between the answer data and the service label and determining the service type corresponding to the answer data.
Optionally, the target answer determining module includes:
a weight determination submodule, configured to determine, according to a degree of association between the answer data and the service tag and a degree of association between the question data and the service tag, a weight of the question data corresponding to the answer data;
and the weight determining submodule is used for weighting the weight of the answer data corresponding to the question data according to the first weight comparison and calculating a third weight ratio between the question data and the candidate answer data.
Optionally, the apparatus further comprises:
and the depth determining module is used for determining answer preset depth corresponding to the question data, and the answer preset depth is used for limiting the number of the second target service types.
Optionally, the second target service determining sub-module is specifically configured to determine the second target service type corresponding to the question data according to a horizontal association weight ratio and/or a vertical association weight ratio between the first target service type and another service type, and the answer preset depth.
In yet another aspect of the present invention, there is also provided a computer-readable storage medium, in which a computer program is stored, which, when run on a computer, causes the computer to execute any one of the above-mentioned service question-answer data processing methods.
In another aspect of the present invention, there is also provided an electronic device, including a processor, a communication interface, a memory and a communication bus, where the processor, the communication interface and the memory complete communication with each other through the communication bus;
a memory for storing a computer program;
and the processor is used for realizing the service question-answering data processing method when executing the program stored in the memory.
In yet another aspect of the present invention, there is also provided a computer program product containing instructions which, when run on a computer, cause the computer to execute any of the above-mentioned service question-answer data processing methods.
According to the method for processing the service question-answer data, provided by the embodiment of the invention, after the question data is received, the first weight ratio is obtained according to the association degree between the question data and the service label, the target service type is determined according to the first weight ratio and the second weight ratio between the service types, the answer data corresponding to the target service type is used as the candidate answer data, the target answer data is determined according to the third weight ratio between the question data and the candidate answer data, and the target service type related to the question data can be determined according to the relationship between the service types while the relationship between the question data and the service types is considered. Because the target answer data is determined from the candidate answer data corresponding to the target service type, the dimensionality of the target answer data corresponding to the question data is expanded, the depth and the breadth of the answer data can be improved, the accuracy in obtaining the answer data corresponding to the question data is improved, the problem can be solved in a more professional manner, and the requirement of a user for obtaining the answer data is met.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below.
FIG. 1 is a flow chart illustrating steps of a method for processing service question answering data according to an embodiment of the present invention;
FIG. 2 is a flow chart of steps in another method for processing service question and answer data in an embodiment of the present invention;
FIG. 3 is a flowchart illustrating steps of another method for processing service question answering data according to an embodiment of the present invention;
FIG. 4 is a block diagram of a business question answering data processing apparatus according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of an electronic device in an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described below with reference to the drawings in the embodiments of the present invention.
The embodiment of the invention can be applied to a scene at a server side, and comprises the steps of determining corresponding answer data according to question data after receiving the question data such as question related texts, pictures and the like sent by a client, so as to realize the processing of the question and answer data in business, such as knowledge question and answer, customer service operation, a search engine and the like, wherein the search engine can receive keywords used for retrieval in a search bar of a user as the question data, and a search result returned according to the keywords as the answer data. For convenience of understanding, the following application scenario of customer service operation is taken as an example, that is, a scenario of solving a problem encountered by a user in a process of using a product in a manner of customer service question and answer.
Referring to fig. 1, a flowchart illustrating steps of a method for processing service question answering data in an embodiment of the present invention is shown, and as shown in fig. 1, the method includes:
step 101: receiving problem data;
step 102: calculating the association degree between the problem data and a service label to obtain a first weight ratio between the problem data and a service type corresponding to the service label;
step 103: determining a target service type corresponding to the problem data according to a first weight ratio between the problem data and the service types and a second weight ratio between the service types, wherein the second weight ratio is obtained by calculating the association degree between the service types;
step 104: determining answer data corresponding to the target service type as candidate answer data corresponding to the question data;
step 105: and calculating a third weight ratio between the question data and the candidate answer data, and determining target answer data corresponding to the question data according to the third weight ratio.
In the embodiment of the present invention, the problem data may be data that is input by the user and used to describe a problem encountered by the user in the service using process, and optionally, the problem data may be a text, a sentence, a voice, or the like, or a picture, a video, or the like, that is input by the user. Optionally, in the process of receiving the problem data, the problem data may be parsed, for example, a text sentence or a voice is converted into a text to obtain at least one keyword therein, or fault information in the picture or the video may be identified, for example, if the picture is a photograph of a display error prompt interface, an error code in the picture may be identified, and the like, and if the video is a screen recording or a video obtained by shooting in the display fault process, an image frame of the video may be extracted, and an error code or an error phenomenon in the image frame may be identified, for example, an error code is identified from the image frame, or a delay is determined from the number of repeated frames of the image frame, and the like. In the embodiment of the present invention, the form and the analysis manner of the problem data are not particularly limited.
In the embodiment of the invention, after receiving problem data, the correlation degree between the problem data and the service tags can be calculated, wherein the service tags are used for distinguishing different service types, one service type can correspond to at least one service tag, the service type can represent a primary reason causing a fault corresponding to the problem data, if the correlation degree of the problem data and the service tags such as virtual machine and ping obstruction is higher, the problem data belongs to the service type of virtual machine fault, or the correlation degree of the problem data and the service tags such as delay and ping obstruction is higher, the problem data problem is caused by load balancing fault, and the like, the service type and the corresponding service tags can be set by workers, and historical problem data can also be collected for analysis, extraction and classification.
In the embodiment of the present invention, the association degree may refer to a similarity between the problem data and the service tag, for example, by extracting a text feature of the service tag and a text feature of the problem data, calculating the similarity between the problem data and the service tag, where the higher the similarity is, the higher the association degree is. The target service type refers to a service type having an association relation with currently received problem data, and may be a service type with an association degree exceeding a preset threshold, and a cause causing the problem may be preliminarily determined from the target service type corresponding to the problem data, and if the association degree of the problem data with a service tag "virtual machine", "ping is not enabled", and the like exceeds the preset threshold, the service type "virtual machine fault" corresponding to the service tag "virtual machine", "ping is not enabled", and the like is determined as the target service type. Optionally, a first weight ratio between the problem data and the service type may be determined according to the association degree, and the first weight ratio may be a quantized association degree, so as to facilitate comparison of the association degrees of different service types and the problem data and determine the association relationship between different service types and the problem data.
In addition, since the problem phenomenon is displayed after the fault occurs, the problem phenomenon which can be observed and understood by the user may be only one of the problems caused by the fault, so that the same fault may cause a plurality of problems, and the same problem may be caused by a plurality of faults. At this time, the service type determined by the question data input by the user is low in pertinence, so that the target service type can be determined by the second weight ratio between the service type and the service type in addition to the first weight ratio, wherein the second weight ratio can represent the association relationship between the service types.
In the embodiment of the invention, the similarity between the service labels can be calculated according to different service labels corresponding to different service types before problem data is acquired, so that the association degree between the service types can be obtained in advance. And the correlation between the service types can be quantified after/before the problem data is acquired to obtain the second weight ratio, and the correlation between the service types is stable because the service labels between the service types generally do not frequently change greatly. And by the second weight ratio among the service types, the range of the target service type is expanded, and the accuracy of the target service type determined according to the problem data is ensured.
In the embodiment of the present invention, each service type corresponds to at least one answer data, and optionally, the answer data may be a specific reason for causing a problem corresponding to the service type, or a scheme for solving the problem, or the like. After the target service type corresponding to the question data is determined, answer data corresponding to the target service type may be used as candidate answer data corresponding to the question data. In the candidate answer data, determining target answer data corresponding to the question data by calculating a third weight ratio between the question data and the candidate answer data, wherein the third weight ratio can be the similarity between text features, semantic features, image features and the like of the question data and the text features, semantic features, image features and the like of the candidate answer data; or, each service type may correspond to at least one question-answer data pair, and a third weight ratio between the question data and answer data in the question-answer data pair is determined according to a similarity between the currently received question data and the question data in the question-answer data pair, which is not specifically limited in the embodiment of the present invention.
Referring to fig. 2, a flowchart illustrating steps of another method for processing service question answering data in the embodiment of the present invention is shown, as shown in fig. 2, on the basis of fig. 1, the method includes:
optionally, the second weight ratio includes a longitudinal association weight ratio corresponding to different service types having a longitudinal association relationship, and/or a horizontal association weight ratio corresponding to different service types having a horizontal association relationship.
In the embodiment of the present invention, the association relationship between the service types may include a horizontal association relationship and a vertical association relationship, where the horizontal association relationship exists between the service types in the same hierarchy and in parallel, and the vertical association relationship exists between the service types in different hierarchies and in the same category. Taking a cloud service as an example, the cloud service is generally divided into three levels, i.e., an Infrastructure as a service (IaaS), a Platform as (PaaS) and a Software as a service (SaaS). The virtual machine operating system, the load balancing, the virtualized storage and the like belong to an IaaS layer, so that a horizontal association relationship exists among service types such as a virtual machine operating system fault, a load balancing fault and a virtualized storage fault. The Java virtual machine belongs to the SaaS layer, the memory of the Java virtual machine belongs to the IaaS layer, and the Java virtual machine comprises the memory of the Java virtual machine. Therefore, it can be considered that there is a vertical correlation between the failure of the Java virtual machine and the failure of the memory of the Java virtual machine. Different business types can be linked through the transverse incidence relation and the longitudinal incidence relation among the business types, so that the dimensionality of answer data can be directionally expanded in the process of determining the answer data corresponding to the question data, and the accuracy of the answer data is improved.
In the embodiment of the present invention, the longitudinal association relationship may be a generic relationship, that is, the service type a and the service type B belong to different levels, and the service type a completely includes the service type B or the service type a is completely included by the service type B; the horizontal association relationship is a parallel relationship, that is, the service type A and the service type B belong to the same level, and the service type A and the service type B are partially overlapped or not overlapped. When the service types are labeled, the second weight ratios of the different service types can be respectively calculated according to the similarity between the different service labels and the incidence relation between the service types, for example, the longitudinal incidence relation corresponds to the longitudinal incidence weight ratio, and the transverse incidence relation corresponds to the transverse incidence weight ratio. Therefore, the service types are distinguished at different latitudes so as to select the target service type meeting the requirements of the user in the horizontal or vertical dimension.
Optionally, the step 103 includes:
substep 1031, determining a first target service type corresponding to the problem data according to a first weight ratio between the problem data and the service type;
a substep 1032, determining a second target service type corresponding to the problem data according to a horizontal correlation weight ratio and/or a vertical correlation weight ratio between the first target service type and other service types;
substep 1033, taking the first target service type and the second target service type as target service types corresponding to the problem data.
In this embodiment of the present invention, the first target service type may be considered as a service type with the highest similarity to the problem data and the closest correlation degree, and optionally, the first weight ratio may be obtained by calculating the similarity between the problem data and all service types, where the higher the similarity is, the closer the correlation degree is, and the larger/smaller the first weight ratio is, and the service type with the highest similarity represented by the first weight ratio is taken as the first target service type corresponding to the problem data. Because the similarity between the first target service type and the problem data is highest, the second target service type is determined according to the second weight ratio between the first target service type and other service types, and the dimensionality of the service types can be expanded on the premise of ensuring the accuracy.
In the embodiment of the present invention, other service types may be considered as service types other than the first target service type among all service types; or, all the service types may be primarily screened according to the first weight ratio, and after the service types lower than the screening threshold of the first weight ratio are discarded, the service types other than the first target service type with the highest first weight ratio among the remaining service types are used as other service types, so that resources required for subsequently determining the second target service type according to the second weight ratio among the service types are saved, and the efficiency of obtaining answer data is improved.
In the embodiment of the present invention, the second target service type is based on the first target service type, and when the second target service type is determined according to the other service types with the expanded second weight ratio, optionally, the first target task type is used as an origin, and according to the horizontal association weight ratio and/or the vertical association weight ratio of the first target service type and the other service types, the other service types with the horizontal association relationship with the first target service type may be determined only by the horizontal association weight ratio, the other service types with the vertical association relationship with the first target service type may be determined only by the vertical association weight ratio, or the horizontal association weight ratio and the vertical association weight ratio are determined alternately, so as to control the expansion direction of the target service type. For example, in practical applications, the longitudinal association relationship between general service types is stronger than the transverse association relationship, and therefore, it may be determined that one other service type is the second target service type through the longitudinal association weight ratio, and then determined that one other service type is the second target service type through the transverse association weight ratio, and so on.
For example, a weight of the problem data 1 may be set to be 1, at this time, a similarity between the problem data 1 and different service tags may be quantized, and since the higher the similarity is, it may be considered that the problem data 1 and the service tags are closer, the higher the similarity is, the smaller the quantization weight is, so as to obtain a first weight ratio of each service type with respect to the problem data 1, for example, the first weight ratio of the service type 1 is 3, the second weight ratio of the service type 2 is 5, the first weight ratio of the service type 3 is 9, and the first weight ratio of the service type 4 is 11. Alternatively, the preset threshold of the first weight ratio may be 10, and the service type with the first weight ratio greater than 10 is considered to have too low similarity with the problem data 1, and the service type 3 is preliminarily filtered.
Since the first weight ratio 3 of the service type 1 is the smallest among the service type 1, the service type 2, and the service type 3, it can be determined that the service type 1 is the first target service type. Or, a difference between each first weight ratio and the weight of the problem data 1 may be calculated to determine a first target service type, and a service label of the service type is considered to be closest to the problem data 1 when the difference is smaller, so that the service type 1 with the smallest difference may be determined as the first target service type.
Further, after the service type 1 is confirmed to be the first target service type, the first weight ratio 3 of the service type 1 may be used as an origin, the longitudinal association weight ratio of the service type 2 is determined to be 4, the transverse association weight ratio of the service type 3 is determined to be 7 according to the hierarchical relationship and the similarity between the service type 1 and the service types 2 and 3, and the second target service type corresponding to the first target service type may be determined according to the second weight ratios of different types and sizes.
In addition, the first weight ratio of each service type may be dynamically adjusted according to different problem data, for example, the weight of the problem data 2 is set to be 1, at this time, the similarity between the problem data 2 and different service tags may be quantized, and since the similarity between the problem data 2 and different service tags may be different from the similarity between the problem data 1, the first weight ratio of each service type with respect to the problem data 2 may also be different, for example, the first weight ratio of the service type 1 with respect to the problem data 2 is 8, the second weight ratio of the service type 2 is 5, the first weight ratio of the service type 3 is 6, the first weight ratio of the service type 4 is 3, and so on, thereby further ensuring the accuracy of each problem data corresponding to the target service type.
Optionally, before 1032, the method further comprises:
substep S11, determining an answer preset depth corresponding to the question data, where the answer preset depth is used to limit the number of the second target service types;
the step 1032 specifically includes determining a second target service type corresponding to the question data according to a horizontal association weight ratio and/or a vertical association weight ratio between the first target service type and other service types, and the preset depth of the answer.
In the embodiment of the present invention, the answer preset depth may be the number of second target service types determined for the question data by default or by the server, and since the first target service type is the service type with the highest degree of association with the question data, and the number is usually one, limiting the number of the second target service types can effectively limit the number of finally obtained target service types, thereby limiting the depth of answer data corresponding to the question data. Because there may be more than one cause for the problem data, and there may be reasons with too small granularity for positioning, the cause for the problem may be more accurately preliminarily positioned by adjusting the number of the target service types. Optionally, in the actual application process, a first weight ratio between the problem data and the service types may be considered to be higher in priority than a second weight ratio between the service types. When the number of the preset target service types is 1, only a first weight ratio between the problem data and the service types is considered, so that the target service types are determined, and when the number of the preset target service types is more than 1, a second weight ratio can be considered, so that the target service types corresponding to the problem data are expanded under the condition that the incidence relation between the problem data and the service types is ensured.
In the embodiment of the present invention, as can be seen from the above process, the target service type may include a first target service type and more than one second target service types associated with the first target service type. Therefore, when the number of target service types is limited, if the number of answer preset depths is 0, only one first target service type corresponding to the question data is determined; if the number of the preset depths of the answers is 1, only determining a first target service type corresponding to the question data and a second target service type longitudinally associated with the first target service type; if the number of the preset depths of the answers is 3, only one first target service type corresponding to the question data is determined, and two second target service types longitudinally associated with the first target service type and one second target service type transversely associated with the first target service type can be determined, so that the depth and the width of the service types corresponding to the question data can be flexibly regulated and controlled, answer data with different precisions can be provided for different question data, and the user requirements can be further met.
Optionally, the step 102 includes:
substep 1021, extracting problem features of the problem data;
a substep 1022 of calculating the degree of association between the problem feature and the business label;
and a substep 1023 of determining a weight value of the service tag corresponding to the question data according to the association degree, and obtaining the first weight ratio between the question data and the service type corresponding to the service tag.
In the embodiment of the present invention, when calculating the association degree between the question data and the answer data, the question features of the question data may be extracted first, wherein the question features may be text features extracted from the question data in a text form, such as keyword features, semantic features, and the like; or text data recognized from question data in the form of speech, and text features extracted from the text data; or text data identified from the question data in the form of pictures and text features extracted from the text data, or image features extracted from the question data in the form of pictures, such as color features, brightness features and the like of a display interface.
In the embodiment of the invention, the degree of correlation between the problem data and the service label can be obtained by calculating the similarity between the problem characteristic and the service label, wherein the degree of correlation can be the similarity of text characteristics, such as whether the keyword of the problem data comprises characters such as 'virtual machine', 'down', and the like, or whether the semantic characteristic comprises the meaning that the 'virtual machine' has a fault; or whether the color feature in the image feature of the problem data is close to the abnormal color in the service tag, whether the brightness feature is close to the abnormal brightness in the service tag, and the like, which are not specifically limited in the embodiment of the present invention.
In the embodiment of the present invention, the relevance may represent the degree of closeness between the problem data and the service tag, and may be smaller relevance represents higher closeness, or larger relevance represents higher closeness. When the degree of association is smaller, the degree of proximity is higher, different first weight ratios can be obtained by quantizing the degree of association, for example, the degree of association of the business type of the business label with the highest degree of association with the problem data is quantized to be 2 as the first weight ratio, the degree of association of the business type of the business label with the second highest degree of association with the problem data is quantized to be 3 as the first weight ratio, and the like,
Referring to fig. 3, a flowchart illustrating steps of another method for processing service question answering data in the embodiment of the present invention is shown, as shown in fig. 3, on the basis of fig. 1, the method includes:
optionally, before the step 101, the method may further include:
substep 1011, calculating the association degree between the answer data and the service label;
substep 1012, classifying the answer data according to the association degree between the answer data and the service label, and determining the service type corresponding to the answer data.
In the embodiment of the invention, in order to obtain target answer data corresponding to the received question data, the existing question data and answer data aiming at the question data can be collected, stored and updated, and then the existing knowledge base storing a large amount of answer data is classified and sorted in different service types in advance, for example, the similarity between the answer data and different service labels is calculated, so that the answer data under different service types are obtained, and the calculation of the similarity between the answer data and the service labels and the calculation of the similarity between the question data and the service labels are similar. The different service types can be set by workers, the answer data can be obtained by clustering service labels corresponding to the service types, or a knowledge base corresponding to the different service types can be pre-established, and the answer data corresponding to the service types is stored in the knowledge base corresponding to the service types according to the similarity with the service labels, so that the answer data is further classified and managed, the influence of the answer data among the different service types is avoided, and the data security is ensured.
For example, if the answer data is highly associated with the traffic labels "virtual machine", "ping disabled", "down", etc., it may be classified as the traffic type of the virtual machine fault, and if the answer data is highly associated with the traffic labels "SNAT fault", "VIP disabled", etc., it may be classified as the traffic type of the load balancing fault.
Optionally, in the step 105, calculating a third weight ratio between the question data and the candidate answer data includes:
substep 1051, determining a weight value of the answer data corresponding to the question data according to the association degree between the answer data and the service tag and the association degree between the question data and the service tag;
substep 1052, weighting the weight of the answer data corresponding to the question data according to the first weight ratio, and calculating a third weight ratio between the question data and the candidate answer data.
In the embodiment of the present invention, after determining the preliminary cause of the problem by determining the target service type corresponding to the problem data, the answer data corresponding to the target service type may be determined as candidate answer data of the problem data, and a third weight ratio between the problem data and the answer data is calculated. Optionally, since in the process of determining the first target traffic type, the first weight ratio of the problem data to all traffic types is calculated, wherein the first weight ratio of the target traffic type is also included. Therefore, a third weight ratio of the question data and the candidate answer data under the target service type can be calculated on the basis of the first weight ratio, for example, the similarity of the question data and the candidate answer data is weighted according to the magnitude of the first weight ratio, or the value of the question data and the candidate answer data and the value of the first weight ratio of the service type corresponding to the candidate answer data are added to determine the third weight ratio. On the basis of the correlation between the question data and the target service type, the correlation degree between the question data and the candidate answer data is calculated, the evaluation of the correlation degree between the question data and the answer data is influenced through the correlation degree between the question data and the service type, and the accuracy of obtaining the answer data is improved.
For example, according to the association degree between the answer data and the service tag corresponding to the service type and the association degree between the question data and the service tag of the same service type, the two association degrees can be further superimposed to determine the association degree between the answer data and the question data, so that the association degree between the answer data and the question data is determined while different service types are considered; or, the answer features of the answer data may be directly extracted, and the similarity between the answer features and the question features may be calculated, so as to obtain the association degree between the answer data and the question data. The relevance is quantized to obtain the weight of the answer data corresponding to the question data, the weight of the answer data 1 corresponding to the question data 1 is 3, the weight of the answer data 2 is 4, and the weight of the answer data 3 is 6.
Further, the weight may be weighted according to a first weight ratio of the service type, for example, the answer data 1 belongs to the service type 1, and the first weight ratio is 5; answer data 2 and answer data 3 belong to business type 2, the first weight ratio of the answer data 2 and the answer data 3 is 3, on the basis, the weight 3 of answer data 1 can be multiplied by 0.5 according to the first weight ratio 5 to obtain a third weight ratio of 1.5, the weights 4 and 6 of answer data 2 and answer data 3 are multiplied by 0.3 according to the first weight ratio to obtain a third weight ratio of 1.2 of answer data 2 and a third weight ratio of 1.8 of answer data 3; or, the weight may be added to the value of the first weight ratio of the service type corresponding to the candidate answer data, to obtain a third weight ratio of 8 for answer data 1, 7 for answer data 2, and 9 for answer data 3.
In the embodiment of the invention, after the target answer data is determined, the target answer data can be recommended to the user, so that the user can solve the problem corresponding to the input problem data according to the answer data. Optionally, for different target answer data, different recommendation manners may be adopted, for example, for target answer data with a third weight ratio above a preset threshold, or for target answer data determined from candidate answer data corresponding to the first target service type, answer data possibly corresponding to the question data may be notified in a dialog recommendation manner. Optionally, the user's mood can be analyzed according to the user input answer data, so as to select a corresponding dialect, generate a dialog according to the dialect with contents indicating a question, a solution and the like in the answer data, and recommend the dialog to the user, wherein the dialect means a way of language expression, different dialect templates such as "bad meaning causes your trouble" can be obtained in advance, according to information provided by your, the question you encounter may be caused by you and hopefully help you "or" timely feedback of thank you ", and through deep analysis, the reason why you caused your question may be caused may be you through question you solve" and the like, and then the user is recommended to be additionally filled according to the answer data. Therefore, the solution corresponding to the acquired problem data is guaranteed to be solved, the experience of the user in the process of acquiring the answer data can be improved, and emotion is avoided.
In addition, the target answer data determined from the candidate answer data corresponding to the second target service type may be in a simple recommendation manner, for example, after the dialog recommendation, the customer service answers "certainly, your question may also be · · · · · · · due to the following reason", and then main content summaries of the remaining target answer data are appended, and further, a selection operation of the user on the main content summaries of the remaining target answer data may be received, and the detailed content of the corresponding target answer data is displayed to the user according to the selection operation. In the embodiment of the present invention, the manner in which the target answer data is recommended to the user is not specifically limited. In addition, evaluation of the user on answer data, such as 'good-medium-poor' or 'all/part/unsolved problem' and the like, can be received, and the answer data corresponding to the service type is optimized and updated according to the evaluation.
In practical application, the embodiment of the invention can be applied to a customer service conversation scene, wherein b can be used1,b2… denote different traffic types, q1,q2… denotes various question data, a1,a2… represent different answer data. Classifying the answer data, wherein b is1Indicating virtual machine failure, with b2Indicating a load balancing failure. b1a1Representing answer data programmed first under a virtual machine failure, b2a1Indicating the answer data that was first programmed in under the load balancing failure.
And calculating the similarity of each service type according to the service label corresponding to each service type, thereby obtaining the association degree between each service type according to the similarity.
After receiving the problem data, calculating a first weight ratio W between the problem data and each service typeqb(q, b), and mixing Wqb(q,b)maxThe corresponding traffic type b1 serves as the first target traffic type.
And with the first weight ratio of b1 as an origin, quantizing the association degree between b1 and each other service type, and constructing a second weight ratio between each service type. Wherein b1 and b2 belong to different levels, and b1 comprises b2, and then, the longitudinal association weight ratio Wb (b1, b2) of b1 and b2 is constructed; b1 and b3 belong to the same hierarchy, and b1 and b3 do not overlap, so that the lateral association weight ratio Wb (b1, b3) of b1 and b3 is constructed.
Then according to the second weight ratio W between the first target service type and other service typesb(bm,bn) E.g. vertical correlation weight ratio Wb(b1,b2) The horizontal correlation weight ratio Wb(b1,b3) Etc. determining Wqb(q,b)max-1，Wqb(q,b)max-2The corresponding service type is a second target service type. And taking the first target service type and the second target service type as target service types corresponding to the problem data.
For example, taking cloud services as an example, the cloud services are divided into three levels, namely, PaaS, SaaS and IaaS from top to bottom, and taking 10 9 service types as an example, a corresponding service type table is established according to a second weight ratio between the service types as follows:
table 1 association table of each service type of cloud service
PaaS 5 8 11
SaaS 3 6 9
IaaS 1 4 7
In table 1, 3, 4, 5, 6, 7, 8, 9, and 11 indicate 9 service types, and the degree of similarity in the numbers indicates the degree of association between different service types. Optionally, since the difference between the service type No. 2 and the service type No. 1 is small, in a scenario of a complex service type, it is not beneficial to the mutual integration between the data models, and therefore, the service type No. 2 may not be compiled into table 1. The same applies to the service type 10. Optionally, table 1 may be correspondingly expanded according to the number of the service types and the number of the service layer levels.
After receiving the problem data, setting the weight of the problem data to be 1, calculating the similarity between the problem data and each service type to obtain the association degree between the problem data and the service type, and quantizing the association degree to determine the first weight ratio of each service type. Since the closer the association degree is, the higher the similarity is, it may be determined that the service type No. 1 with the lowest first weight ratio is the first target service type, and if the problem data is "virtual machine unavailable", the first target service type "network failure" is obtained according to the first weight ratio of 3.
However, the type 1 service is too single and superficial, so that the second target service type can be further determined. Since the first weight ratio of the service type No. 1 is 3, on this basis, it can be obtained that the longitudinal associated weight ratio of the service type No. 3 is 5, the transverse associated weight ratio of the service type No. 4 is 6, the total longitudinal associated party ratio of the service type No. 5 is 7, the transverse associated weight ratio of the service type No. 7 is 9, and the like. Further, when the answer preset depth is 2, it may be determined that the service type No. 3 is the second target service type, and then the service type No. 4 is the second target service type, and the like. And finally obtaining the target service type corresponding to the problem data.
Taking answer data corresponding to the target service type as candidate answer data corresponding to the question data, and constructing a third weight ratio W between the question data and the candidate answer dataaq(a, q) the third weight ratio is Wab(a, b, q) to calculate, Wab(a, b, q) represents the degree of correlation between candidate answer data corresponding to the target service type and question data in the knowledge base. Third weight ratio W corresponding to different question data for each answerab(a,b1,qn)、Wab(a,b1,qm)。
And determining target answer data according to the third weight ratio, and recommending the target answer data to the user.
According to the method for processing the service question-answer data, provided by the embodiment of the invention, after the question data is received, the first weight ratio is obtained according to the association degree between the question data and the service label, the target service type is determined according to the first weight ratio and the second weight ratio between the service types, the answer data corresponding to the target service type is used as the candidate answer data, the target answer data is determined according to the third weight ratio between the question data and the candidate answer data, and the target service type related to the question data can be determined according to the relationship between the service types while the relationship between the question data and the service types is considered. Because the target answer data is determined from the candidate answer data corresponding to the target service type, the dimensionality of the target answer data corresponding to the question data is expanded, the depth and the breadth of the answer data can be improved, the accuracy in obtaining the answer data corresponding to the question data is improved, the problem can be solved in a more professional manner, and the requirement of a user for obtaining the answer data is met.
Referring to fig. 4, a block diagram of a service question answering data processing device 400 according to an embodiment of the present invention is shown, and as shown in fig. 4, the device includes:
a question data receiving module 401, configured to receive question data;
a first weight calculation module 402, configured to calculate a degree of association between the problem data and a service tag, and obtain a first weight ratio between the problem data and a service type corresponding to the service tag;
a target service determining module 403, configured to determine a target service type corresponding to the problem data according to a first weight ratio between the problem data and the service type and a second weight ratio between the service types, where the second weight ratio is obtained by calculating a degree of association between the service types;
a candidate answer determining module 404, configured to determine that answer data corresponding to the target service type is candidate answer data corresponding to the question data;
a target answer determining module 405, configured to calculate a third weight ratio between the question data and the candidate answer data, and determine target answer data corresponding to the question data according to the third weight ratio.
Optionally, the second weight ratio includes a longitudinal association weight ratio corresponding to different service types having a longitudinal association relationship, and/or a horizontal association weight ratio corresponding to different service types having a horizontal association relationship.
Optionally, the target service determining module 403 includes:
the first target service determining submodule is used for determining a first target service type corresponding to the problem data according to a first weight ratio between the problem data and the service type;
a second target service determining submodule, configured to determine a second target service type corresponding to the problem data according to a horizontal association weight ratio and/or a vertical association weight ratio between the first target service type and another service type;
and the target service determining submodule is used for taking the first target service type and the second target service type as target service types corresponding to the problem data.
Optionally, the first weight calculating module 402 includes:
the characteristic extraction submodule is used for extracting problem characteristics of the problem data;
the correlation calculation submodule is used for calculating the correlation degree between the problem characteristics and the service labels;
and the weight calculation submodule is used for determining the weight of the problem data corresponding to the service tag according to the association degree and obtaining the first weight ratio between the problem data and the service type corresponding to the service tag.
Optionally, the apparatus may further include:
the association calculation module is used for calculating the association degree between the answer data and the service label;
and the service determining module is used for classifying the answer data according to the association degree between the answer data and the service label and determining the service type corresponding to the answer data.
Optionally, the target answer determining module 405 includes:
a weight determination submodule, configured to determine, according to a degree of association between the answer data and the service tag and a degree of association between the question data and the service tag, a weight of the question data corresponding to the answer data;
and the weight determining submodule is used for weighting the weight of the answer data corresponding to the question data according to the first weight comparison and calculating a third weight ratio between the question data and the candidate answer data.
Optionally, the apparatus further comprises:
and the depth determining module is used for determining answer preset depth corresponding to the question data, and the answer preset depth is used for limiting the number of the second target service types.
Optionally, the second target service determining sub-module is specifically configured to determine the second target service type corresponding to the question data according to a horizontal association weight ratio and/or a vertical association weight ratio between the first target service type and another service type, and the answer preset depth.
According to the method for processing the service question-answer data, provided by the embodiment of the invention, after the question data is received, the first weight ratio is obtained according to the association degree between the question data and the service label, the target service type is determined according to the first weight ratio and the second weight ratio between the service types, the answer data corresponding to the target service type is used as the candidate answer data, the target answer data is determined according to the third weight ratio between the question data and the candidate answer data, and the target service type related to the question data can be determined according to the relationship between the service types while the relationship between the question data and the service types is considered. Because the target answer data is determined from the candidate answer data corresponding to the target service type, the dimensionality of the target answer data corresponding to the question data is expanded, the depth and the breadth of the answer data can be improved, the accuracy in obtaining the answer data corresponding to the question data is improved, the problem can be solved in a more professional manner, and the requirement of a user for obtaining the answer data is met.
An embodiment of the present invention further provides an electronic device, as shown in fig. 5, which includes a processor 501, a communication interface 502, a memory 503 and a communication bus 504, where the processor 501, the communication interface 502 and the memory 503 complete mutual communication through the communication bus 504,
a memory 503 for storing a computer program;
the processor 501, when executing the program stored in the memory 503, implements the following steps:
receiving problem data;
calculating the association degree between the problem data and a service label to obtain a first weight ratio between the problem data and a service type corresponding to the service label;
determining a target service type corresponding to the problem data according to a first weight ratio between the problem data and the service types and a second weight ratio between the service types, wherein the second weight ratio is obtained by calculating the association degree between the service types;
determining answer data corresponding to the target service type as candidate answer data corresponding to the question data;
and calculating a third weight ratio between the question data and the candidate answer data, and determining target answer data corresponding to the question data according to the third weight ratio.
The communication bus mentioned in the above terminal may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown, but this does not mean that there is only one bus or one type of bus.
The communication interface is used for communication between the terminal and other equipment.
The Memory may include a Random Access Memory (RAM) or a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. Optionally, the memory may also be at least one memory device located remotely from the processor.
The Processor may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the device can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component.
In another embodiment of the present invention, a computer-readable storage medium is further provided, where instructions are stored in the computer-readable storage medium, and when the instructions are executed on a computer, the computer is caused to execute the service question and answer data processing method described in any one of the above embodiments.
In another aspect of the present invention, there is also provided an electronic device, including a processor, a communication interface, a memory and a communication bus, where the processor, the communication interface and the memory complete communication with each other through the communication bus;
a memory for storing a computer program;
and the processor is used for realizing the service question-answering data processing method when executing the program stored in the memory.
In another embodiment of the present invention, there is also provided a computer program product containing instructions, which when run on a computer, causes the computer to execute the method for processing service question and answer data according to any one of the above embodiments.
In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, cause the processes or functions described in accordance with the embodiments of the invention to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website site, computer, server, or data center to another website site, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that incorporates one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, 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 identical elements in a process, method, article, or apparatus that comprises the element.
All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.

Claims (10)

1. A method for processing service question and answer data, the method comprising:
receiving problem data;
calculating the association degree between the problem data and a service label to obtain a first weight ratio between the problem data and a service type corresponding to the service label;
determining a target service type corresponding to the problem data according to a first weight ratio between the problem data and the service types and a second weight ratio between the service types, wherein the second weight ratio is obtained by calculating the association degree between the service types;
determining answer data corresponding to the target service type as candidate answer data corresponding to the question data;
and calculating a third weight ratio between the question data and the candidate answer data, and determining target answer data corresponding to the question data according to the third weight ratio.
2. The method according to claim 1, wherein the second weight ratio comprises a longitudinal association weight ratio corresponding to different service types having a longitudinal association relationship and/or a horizontal association weight ratio corresponding to different service types having a horizontal association relationship;
the determining a target service type corresponding to the problem data according to a first weight ratio between the problem data and the service type and a second weight ratio between the service types includes:
determining a first target service type corresponding to the problem data according to a first weight ratio between the problem data and the service type;
determining a second target service type corresponding to the problem data according to a transverse correlation weight ratio and/or a longitudinal correlation weight ratio between the first target service type and other service types;
and taking the first target service type and the second target service type as target service types corresponding to the problem data.
3. The method of claim 1, wherein the calculating the association degree between the question data and the service label to obtain a first weight ratio between the question data and a service type corresponding to the service label comprises:
extracting problem features of the problem data;
calculating the correlation degree between the problem characteristics and the service label;
and determining the weight of the problem data corresponding to the service label according to the association degree, and obtaining the first weight ratio between the problem data and the service type corresponding to the service label.
4. The method of claim 1, wherein prior to receiving the issue data, the method further comprises:
calculating the association degree between answer data and the service label;
classifying the answer data according to the association degree between the answer data and the service label, and determining the service type corresponding to the answer data.
5. The method of claim 4, wherein calculating a third weight ratio between the question data and the candidate answer data comprises:
determining the weight of the answer data corresponding to the question data according to the association degree between the answer data and the service label and the association degree between the question data and the service label;
and according to the first weight, the weight of the answer data corresponding to the question data is compared and weighted, and a third weight ratio between the question data and the candidate answer data is calculated.
6. The method according to claim 2, wherein before determining the second target traffic type corresponding to the problem data according to the horizontal association weight ratio and/or the vertical association weight ratio between the first target traffic type and the other traffic types, the method comprises:
determining an answer preset depth corresponding to the question data, wherein the answer preset depth is used for limiting the number of the second target service types;
the determining a second target service type corresponding to the problem data according to the horizontal correlation weight ratio and/or the vertical correlation weight ratio between the first target service type and other service types includes:
and determining a second target service type corresponding to the question data according to the transverse correlation weight ratio and/or the longitudinal correlation weight ratio between the first target service type and other service types and the preset depth of the answer.
7. A business question-answer data apparatus, characterized in that said apparatus comprises:
the problem data receiving module is used for receiving problem data;
the first weight calculation module is used for calculating the association degree between the problem data and the service label to obtain a first weight ratio between the problem data and the service type corresponding to the service label;
the target service determining module is used for determining a target service type corresponding to the problem data according to a first weight ratio between the problem data and the service types and a second weight ratio between the service types, wherein the second weight ratio is obtained by calculating the association degree between the service types;
a candidate answer determining module, configured to determine that answer data corresponding to the target service type is candidate answer data corresponding to the question data;
and the target answer determining module is used for calculating a third weight ratio between the question data and the candidate answer data and determining target answer data corresponding to the question data according to the third weight ratio.
8. The apparatus according to claim 7, wherein the second weight ratio comprises a vertical association weight ratio corresponding to different service types having vertical association relationship and/or a horizontal association weight ratio corresponding to different service types having horizontal association relationship;
the target service determination module includes:
the first target service determining submodule is used for determining a first target service type corresponding to the problem data according to a first weight ratio between the problem data and the service type;
a second target service determining submodule, configured to determine a second target service type corresponding to the problem data according to a horizontal association weight ratio and/or a vertical association weight ratio between the first target service type and another service type;
and the target service determining submodule is used for taking the first target service type and the second target service type as target service types corresponding to the problem data.
9. An electronic device is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor and the communication interface are used for realizing mutual communication by the memory through the communication bus;
a memory for storing a computer program;
a processor for implementing the method of processing service question answering data according to any one of claims 1 to 6 when executing the program stored in the memory.
10. A computer-readable storage medium on which a computer program is stored, the program, when being executed by a processor, implementing the service question answering data processing method according to any one of claims 1 to 6.
CN202010301636.5A 2020-04-16 2020-04-16 Service question-answer data processing method and device Active CN111651567B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010301636.5A CN111651567B (en) 2020-04-16 2020-04-16 Service question-answer data processing method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010301636.5A CN111651567B (en) 2020-04-16 2020-04-16 Service question-answer data processing method and device

Publications (2)

Publication Number Publication Date
CN111651567A true CN111651567A (en) 2020-09-11
CN111651567B CN111651567B (en) 2023-09-22

Family

ID=72348251

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010301636.5A Active CN111651567B (en) 2020-04-16 2020-04-16 Service question-answer data processing method and device

Country Status (1)

Country Link
CN (1) CN111651567B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112948555A (en) * 2021-03-03 2021-06-11 北京奇艺世纪科技有限公司 Man-machine interaction method and device, electronic equipment and storage medium
CN113206759A (en) * 2021-04-27 2021-08-03 北京赛博云睿智能科技有限公司 Management method and system of intelligent PaaS service platform across different service domains
WO2022126965A1 (en) * 2020-12-18 2022-06-23 平安科技（深圳）有限公司 Intelligent question-answering method and apparatus, device, and storage medium

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108415980A (en) * 2018-02-09 2018-08-17 平安科技（深圳）有限公司 Question and answer data processing method, electronic device and storage medium
CN108595695A (en) * 2018-05-08 2018-09-28 和美（深圳）信息技术股份有限公司 Data processing method, device, computer equipment and storage medium
CN109918560A (en) * 2019-01-09 2019-06-21 平安科技（深圳）有限公司 A kind of answering method and device based on search engine
US20190371303A1 (en) * 2018-06-03 2019-12-05 Apple Inc. Providing semantically relevant answers to questions

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108415980A (en) * 2018-02-09 2018-08-17 平安科技（深圳）有限公司 Question and answer data processing method, electronic device and storage medium
CN108595695A (en) * 2018-05-08 2018-09-28 和美（深圳）信息技术股份有限公司 Data processing method, device, computer equipment and storage medium
US20190371303A1 (en) * 2018-06-03 2019-12-05 Apple Inc. Providing semantically relevant answers to questions
CN109918560A (en) * 2019-01-09 2019-06-21 平安科技（深圳）有限公司 A kind of answering method and device based on search engine

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2022126965A1 (en) * 2020-12-18 2022-06-23 平安科技（深圳）有限公司 Intelligent question-answering method and apparatus, device, and storage medium
CN112948555A (en) * 2021-03-03 2021-06-11 北京奇艺世纪科技有限公司 Man-machine interaction method and device, electronic equipment and storage medium
CN112948555B (en) * 2021-03-03 2023-09-01 北京奇艺世纪科技有限公司 Man-machine interaction method and device, electronic equipment and storage medium
CN113206759A (en) * 2021-04-27 2021-08-03 北京赛博云睿智能科技有限公司 Management method and system of intelligent PaaS service platform across different service domains
CN113206759B (en) * 2021-04-27 2022-10-14 北京赛博云睿智能科技有限公司 Management method and system of intelligent PaaS service platform across different service domains

Also Published As

Publication number Publication date
CN111651567B (en) 2023-09-22

Similar Documents

Publication Publication Date Title
CN106874441B (en) Intelligent question-answering method and device
US9449282B2 (en) System for determining and optimizing for relevance in match-making systems
US11651247B2 (en) Method for verifying lack of bias of deep learning AI systems
US7099855B1 (en) System and method for electronic communication management
US9218568B2 (en) Disambiguating data using contextual and historical information
CN111651567A (en) Service question and answer data processing method and device
CN107436916B (en) Intelligent answer prompting method and device
US10606910B2 (en) Ranking search results using machine learning based models
US9348901B2 (en) System and method for rule based classification of a text fragment
US20230237055A1 (en) Method and system for analytic based connections among user types in an online platform
US11775894B2 (en) Intelligent routing framework
CN112307770A (en) Sensitive information detection method and device, electronic equipment and storage medium
CN107291774B (en) Error sample identification method and device
CN116109373A (en) Recommendation method and device for financial products, electronic equipment and medium
CN112016850A (en) Service evaluation method and device
CN115062621A (en) Label extraction method and device, electronic equipment and storage medium
CN114548118A (en) Service conversation detection method and system
CN108415971B (en) Method and device for recommending supply and demand information by using knowledge graph
CN112199500A (en) Emotional tendency identification method and device for comments and electronic equipment
US20210224303A1 (en) Searching device and searching program
US11972467B2 (en) Question-answer expansion
CN113612777B (en) Training method, flow classification method, device, electronic equipment and storage medium
CN115618415A (en) Sensitive data identification method and device, electronic equipment and storage medium
CN114579876A (en) False information detection method, device, equipment and medium
JPWO2013129311A1 (en) Dissatisfaction extraction device, dissatisfaction extraction method, and dissatisfaction extraction program

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant