CN108804567A - Improve method, equipment, storage medium and the device of intelligent customer service response rate - Google Patents
Improve method, equipment, storage medium and the device of intelligent customer service response rate Download PDFInfo
- Publication number
- CN108804567A CN108804567A CN201810499843.9A CN201810499843A CN108804567A CN 108804567 A CN108804567 A CN 108804567A CN 201810499843 A CN201810499843 A CN 201810499843A CN 108804567 A CN108804567 A CN 108804567A
- Authority
- CN
- China
- Prior art keywords
- customer service
- intelligent customer
- target criteria
- response rate
- service response
- 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.)
- Pending
Links
- 230000004044 response Effects 0.000 title claims abstract description 79
- 238000000034 method Methods 0.000 title claims abstract description 47
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 claims abstract description 13
- 230000015654 memory Effects 0.000 claims description 15
- 239000013598 vector Substances 0.000 claims description 15
- 238000007477 logistic regression Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 9
- 238000004891 communication Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 239000003795 chemical substances by application Substances 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000001035 drying Methods 0.000 description 1
- 230000008451 emotion Effects 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000036651 mood Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 238000012549 training Methods 0.000 description 1
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
- 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
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0203—Market surveys; Market polls
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
- G06F40/35—Discourse or dialogue representation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Mathematical Physics (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Strategic Management (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Human Computer Interaction (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Databases & Information Systems (AREA)
- Game Theory and Decision Science (AREA)
- Economics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of method, equipment, storage medium and device improving intelligent customer service response rate, this method includes:The dont answer problem for obtaining intelligent customer service, classifies to dont answer problem, obtains knowledge type problem;Calculate the first similarity between each knowledge type problem;Each knowledge type problem is clustered by affine propagation clustering algorithm according to the first similarity, obtains multiple clustering clusters;Using the cluster centre of each clustering cluster as target criteria problem, and obtains target criteria corresponding with target criteria problem and reply;Target criteria problem and the reply of corresponding target criteria are added in the knowledge base of intelligent customer service.In the present invention, by the way that the knowledge type problem in dont answer problem is clustered by affine propagation clustering algorithm, the cluster centre of each clustering cluster of acquisition is added to intelligent customer service as target criteria problem, cluster centre is most representational problem in each clustering cluster, to improve the response rate of intelligent customer service, user experience is promoted.
Description
Technical field
The present invention relates to intelligent customer service technical field more particularly to it is a kind of improve the method for intelligent customer service response rate, equipment,
Storage medium and device.
Background technology
With the development of science and technology, intelligent customer service system is increasingly paid attention to, but intelligent customer service cannot still be replaced completely at present
For artificial customer service, many problem intelligent customer services that client proposes can not be answered so that need to put into a large amount of artificial customer service to this
A little problems are answered, and cause personnel cost high, and user often cannot get standard and return when carrying out inquiry using intelligent customer service
It is multiple, and manual service of transferring, lead to poor user experience, therefore, how to improve the skill that the response rate of intelligent customer service is urgently to be resolved hurrily
Art problem.
The above is only used to facilitate the understanding of the technical scheme, and is not represented and is recognized that the above is existing skill
Art.
Invention content
The main purpose of the present invention is to provide it is a kind of improve the method for intelligent customer service response rate, equipment, storage medium and
Device, it is intended to solve the low technical problem of the response rate of intelligent customer service in the prior art.
To achieve the above object, the present invention provides a kind of method improving intelligent customer service response rate, the raising intelligence visitor
The method for taking response rate includes the following steps:
The dont answer problem for obtaining intelligent customer service, classifies to the dont answer problem, obtains knowledge type problem;
Calculate the first similarity between each knowledge type problem;
Each knowledge type problem is clustered by affine propagation clustering algorithm according to first similarity, is obtained multiple
Clustering cluster;
Using the cluster centre of each clustering cluster as target criteria problem, and obtain mesh corresponding with the target criteria problem
Mark standard is replied;
The target criteria problem and the reply of corresponding target criteria are added in the knowledge base of the intelligent customer service.
Preferably, the dont answer problem for obtaining intelligent customer service, classifies to the dont answer problem, obtains knowledge
Type problem, including:
The dont answer problem for obtaining intelligent customer service, calculates the frequency of the dont answer problem, is more than first by the frequency
The dont answer problem of predetermined threshold value is as pending problem;
The pending problem is classified by default disaggregated model, obtains knowledge type problem.
Preferably, first similarity calculated between each knowledge type problem, including:
Word segmentation processing is carried out to each knowledge type problem, to obtain word all in the knowledge type problem, described in calculating
The TF-IDF values of word;
It is the term vector formed with the TF-IDF values of word and word by each knowledge type problem representation;
The COS distance between each term vector is calculated, and using the COS distance as first between each knowledge type problem
Similarity.
Preferably, described using the cluster centre of the clustering cluster as target criteria problem, and obtain and the target mark
After the corresponding target criteria of quasi- problem is replied, the method for improving intelligent customer service response rate further includes:
Calculate second between the primary standard problem in the knowledge base of the target criteria problem and the intelligent customer service
Similarity;
The maximum value in second similarity is obtained, judges the maximum value whether more than the second predetermined threshold value;
If the maximum value is more than second predetermined threshold value, the corresponding primary standard problem conduct of the maximum value is obtained
Wait for supplementary question;
It is described that the target criteria problem and corresponding target criteria are replied to the knowledge base for being added to the intelligent customer service
In, including:
The target criteria problem and corresponding target criteria are replied and are added to institute in the knowledge base of the intelligent customer service
It states in the scaling problem list for waiting for supplementary question.
Preferably, the maximum value obtained in second similarity judges whether the maximum value is pre- more than second
If after threshold value, the method for improving intelligent customer service response rate further includes:
If the maximum value is less than second predetermined threshold value, by each clustering cluster other than target criteria problem
Other knowledge type problems are as the first scaling problem;
Calculate the third similarity between first scaling problem and corresponding target criteria problem;
N number of first scaling problem is selected in first scaling problem according to the third similarity descending order
As the scaling problem list of the target criteria problem, the N is the integer more than or equal to 1;
It is described that the target criteria problem and corresponding target criteria are replied to the knowledge base for being added to the intelligent customer service
In, including:
The scaling problem list of the target criteria problem, the target criteria problem and corresponding target criteria are replied
It is added in the form of question of independence in the knowledge base of the intelligent customer service.
Preferably, the third similarity calculated between first scaling problem and corresponding target criteria problem,
Including:
Calculate statistical nature, semantic feature and the theme between first scaling problem and corresponding target criteria problem
Feature;
The statistical nature, the semantic feature and the theme feature are polymerize by logistic regression, obtain institute
State the third similarity between the first scaling problem and corresponding target criteria problem.
Preferably, the statistical nature includes:Term co-occurrence rate, TF-IDF values, editing distance and Longest Common Substring.
In addition, to achieve the above object, the present invention also proposes a kind of equipment improving intelligent customer service response rate, the raising
The equipment of intelligent customer service response rate includes memory, processor and is stored on the memory and can transport on the processor
The program of capable raising intelligent customer service response rate, the program for improving intelligent customer service response rate are arranged for carrying out as described above
Raising intelligent customer service response rate method the step of.
In addition, to achieve the above object, the present invention also proposes a kind of storage medium, stores and is improved on the storage medium
The program of intelligent customer service response rate, the program for improving intelligent customer service response rate are realized as described above when being executed by processor
Raising intelligent customer service response rate method the step of.
In addition, to achieve the above object, the present invention also proposes a kind of device improving intelligent customer service response rate, the raising
The device of intelligent customer service response rate includes:Sort module, computing module, cluster module, acquisition module and add module;
The sort module, the dont answer problem for obtaining intelligent customer service are classified to the dont answer problem, are obtained
Obtain knowledge type problem;
The computing module, for calculating the first similarity between each knowledge type problem;
The cluster module, for according to first similarity by affine propagation clustering algorithm by each knowledge type problem
It is clustered, obtains multiple clustering clusters;
The acquisition module is used for using the cluster centre of each clustering cluster as target criteria problem, and is obtained and the mesh
The corresponding target criteria of typical problem is marked to reply;
The add module, for the target criteria problem and the reply of corresponding target criteria to be added to the intelligence
In the knowledge base of customer service.
In the present invention, the dont answer problem of intelligent customer service is obtained, is classified to the dont answer problem, knowledge type is obtained
Problem, to reject the non-knowledge type problem that classification obtains, the non-knowledge type problem does not have what standard was replied, without carrying out
Addition considers, to improve the quality for the target criteria problem subsequently added;By calculating the between the knowledge type problem
One similarity obtains multiple clustering clusters according to the first similarity by affine propagation clustering algorithm, will be in the cluster of each clustering cluster
The heart is added to as target criteria problem in the knowledge base of intelligent customer service, and the cluster centre is most represented in each clustering cluster
It the problem of property, is added in intelligent customer service as target criteria problem, the response rate of intelligent customer service can be effectively improved, carried
Rise user experience.
Description of the drawings
Fig. 1 is the device structure of the raising intelligent customer service response rate for the hardware running environment that the embodiment of the present invention is related to
Schematic diagram;
Fig. 2 is the flow diagram for the method first embodiment that the present invention improves intelligent customer service response rate;
Fig. 3 is the flow diagram for the method second embodiment that the present invention improves intelligent customer service response rate;
Fig. 4 is the flow diagram for the method 3rd embodiment that the present invention improves intelligent customer service response rate;
Fig. 5 is the structure diagram for the device first embodiment that the present invention improves intelligent customer service response rate.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific implementation mode
It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not intended to limit the present invention.
Referring to Fig.1, Fig. 1 is the raising intelligent customer service response rate for the hardware running environment that the embodiment of the present invention is related to
Device structure schematic diagram.
As shown in Figure 1, the equipment of the raising intelligent customer service response rate may include:Processor 1001, such as central processing
Device (Central Processing Unit, CPU), communication bus 1002, user interface 1003, network interface 1004, memory
1005.Wherein, communication bus 1002 is for realizing the connection communication between these components.User interface 1003 may include display
Shield (Display), optional user interface 1003 can also include standard wireline interface and wireless interface, for user interface
1003 wireline interface can be USB interface in the present invention.Network interface 1004 may include optionally standard wireline interface,
Wireless interface (such as Wireless Fidelity (WIreless-FIdelity, WI-FI) interface).Memory 1005 can be the random of high speed
Memory (Random Access Memory, RAM) memory is accessed, can also be stable memory (Non-volatile
Memory, NVM), such as magnetic disk storage.Memory 1005 optionally can also be the storage independently of aforementioned processor 1001
Device.
It will be understood by those skilled in the art that structure shown in Fig. 1 is not constituted to improving intelligent customer service response rate
The restriction of equipment may include either combining certain components or different component cloth than illustrating more or fewer components
It sets.
As shown in Figure 1, as may include that operating system, network are logical in a kind of memory 1005 of computer storage media
Believe module, Subscriber Interface Module SIM and the program for improving intelligent customer service response rate.
In the equipment of raising intelligent customer service response rate shown in Fig. 1, network interface 1004 is mainly used for connection backstage and takes
It is engaged in device, with the background server into row data communication;User interface 1003 is mainly used for connecting intelligent customer service equipment;It is described to carry
The equipment of high intelligent customer service response rate calls the raising intelligent customer service response rate stored in memory 1005 by processor 1001
Program, and execute it is provided in an embodiment of the present invention improve intelligent customer service response rate method.
Based on above-mentioned hardware configuration, propose that the present invention improves the embodiment of the method for intelligent customer service response rate.
With reference to Fig. 2, Fig. 2 is the flow diagram for the method first embodiment that the present invention improves intelligent customer service response rate, is carried
Go out the method first embodiment that the present invention improves intelligent customer service response rate.
In the first embodiment, the method for improving intelligent customer service response rate includes the following steps:
Step S10:The dont answer problem for obtaining intelligent customer service, classifies to the dont answer problem, obtains knowledge type
Problem.
It should be understood that the executive agent of the present embodiment is the equipment for improving intelligent customer service response rate, wherein the raising
The equipment of intelligent customer service response rate can be the electronic equipments such as PC, server.Dont answer problem is divided into two classes, it is a kind of
For knowledge type problem, one kind is non-knowledge type problem.Knowledge type problem can usually find corresponding standard response, can will be described
Knowledge type problem adds in knowledge base.The perception problem such as non-knowledge type problem, such as mood, emotion, does not have standard and replys,
Consider without being added, is not added to necessity of knowledge base.Therefore, we retain dont answer and ask after classification is handled
Knowledge type problem in topic is handled with doing subsequent addition to knowledge type problem.
It should be noted that classify to the dont answer problem, it can be by building vocabulary, the vocabulary can be regarded as
It is a huge set, all words are all stored in the vocabulary in the training set text of knowledge type problem, in building process
In can also calculate quantity, index and the type of each word, the type includes language material and label.In term vector space, these
Knowledge type problem has approximate term vector, these approximate term vectors correspond to identical tag along sort again, in repetitive exercise process
In, this correlation can be propagated constantly, until train accurate disaggregated model, then can by the disaggregated model to it is described not
Response problem is classified, and the knowledge type problem is obtained.
Step S20:Calculate the first similarity between each knowledge type problem.
In the concrete realization, each knowledge type problem can be segmented, calculates the TF-IDF values of word as word feature, it will be each
Knowledge type problem representation is term vector, calculates the COS distance between term vector as the first similarity between each knowledge type problem
Measurement.TF-IDF is actually:TF*IDF, TF word frequency (Term Frequency), the reverse document-frequency (Inverse of IDF
Document Frequency).TF indicates the frequency that entry occurs in document d.The main thought of IDF is:If including word
The document of t is fewer, and IDF is bigger, then illustrates that entry t has good class discrimination ability.High word in a certain particular document
The low document frequency of speech frequency rate and the word in entire collection of document, can produce the TF-IDF of high weight.Therefore,
TF-IDF tends to filter out common word, retains important word.In the present embodiment, it is described calculate each knowledge type problem it
Between the first similarity, specifically include:Word segmentation processing is carried out to each knowledge type problem, is owned to obtain in the knowledge type problem
Word, calculate the TF-IDF values of the word;It is to be formed with the TF-IDF values of word and word by each knowledge type problem representation
Term vector;The COS distance between each term vector is calculated, and using the COS distance as between each knowledge type problem
One similarity.
Step S30:Each knowledge type problem is clustered by affine propagation clustering algorithm according to first similarity,
Obtain multiple clustering clusters.
It will be appreciated that affine propagation clustering (Affinity Propagation, the AP) algorithm, is according to data point
Between similarity clustered, can be symmetrical, can also be asymmetric.The affine propagation clustering algorithm is not required to
It first to determine the number of cluster, but all data points all be regarded as the cluster centre on potential significance.According to each first phase
Similarity matrix can be built like degree measurement, using each knowledge type problem as node, the first similarity between each knowledge type problem
The value as matrix is measured, is then clustered by AP algorithms, obtains multiple clustering clusters.
Step S40:Using the cluster centre of each clustering cluster as target criteria problem, and obtain and the target criteria problem
Corresponding target criteria is replied.
It should be noted that cluster centre can be obtained from the clustering cluster, the cluster centre and corresponding clustering cluster
In other problems similitude it is most strong, most therefore the representativeness of the clustering cluster selects cluster centre as knowledge type problem
Target criteria problem, the other problems in clustering cluster can be used as the corresponding scaling problem of the target criteria problem.
It is returned it should be understood that knowledge type problem can usually find corresponding target criteria from existing data bank
It is multiple, existing data bank can be inquired according to the target criteria problem, inquire and matched with the target criteria problem
Data, as target criteria reply.When inquiring data matched with the target criteria problem, can manually it be done
In advance, matched data is arranged by artificial customer service, generates target criteria and replys.
Step S50:The target criteria problem and the reply of corresponding target criteria are added to knowing for the intelligent customer service
Know in library.
It will be appreciated that using most representative cluster centre in each clustering cluster as target criteria problem, by the mesh
Mark typical problem and the reply of corresponding target criteria are added in the knowledge base of the intelligent customer service so that are had in dont answer problem
Representative problem can be added into the knowledge base of intelligent customer service.
In the first embodiment, the dont answer problem for obtaining intelligent customer service classifies to the dont answer problem, obtains
Knowledge type problem, to reject the non-knowledge type problem that classification obtains, the non-knowledge type problem does not have what standard was replied, nothing
It need to be added consideration, to improve the quality for the target criteria problem subsequently added;By calculate the knowledge type problem it
Between the first similarity, multiple clustering clusters are obtained by affine propagation clustering algorithm according to the first similarity, by each clustering cluster
Cluster centre is added to as target criteria problem in the knowledge base of intelligent customer service, the cluster centre be in each clustering cluster most
Representative problem is added to as target criteria problem in intelligent customer service, can effectively improve answering for intelligent customer service
Rate is answered, user experience is promoted.
It is the flow diagram for the method second embodiment that the present invention improves intelligent customer service response rate, base with reference to Fig. 3, Fig. 3
In above-mentioned first embodiment shown in Fig. 2, propose that the present invention improves the second embodiment of the method for intelligent customer service response rate.
In a second embodiment, the step S10, including:
Step S101:The dont answer problem for obtaining intelligent customer service, calculates the frequency of the dont answer problem, by the frequency
More than the first predetermined threshold value dont answer problem as pending problem.
It will be appreciated that obtaining intelligent customer service dont answer problem, the number that each dont answer problem occurs is counted, to obtain
The frequency of each dont answer problem, the low dont answer problem of usual frequency belong to long-tail problem, and problem quality is usually relatively low, does not have
Referential, it is usual such issues that, we need not analyze.It is more than institute so first predetermined threshold value can be pre-set
The dont answer problem for stating the first predetermined threshold value is that client compares concern, is often asked, and problems have referential, make
Corresponding standard response is found if the pending problem belongs to knowledge type problem for pending problem, can be added to
In knowledge base, if next client inquires again and relevant issues, intelligent customer service can find corresponding standard response from knowledge base
Carry out response.
Step S102:The pending problem is classified by default disaggregated model, obtains knowledge type problem.
It should be noted that the default disaggregated model includes fast text (FastText) disaggregated model, it is described
FastText disaggregated models include three parts:Model framework, hierarchical classification device (Softmax) and Chinese language model (N-gram)
Feature.Hierarchical classification device Softmax is established on the basis of Huffman encodes, and is encoded, can greatly be reduced to label
The quantity of model prediction target greatly reduces the default disaggregated model and predicts knowledge type problem in the pending problem
Quantity.FastText adds N-gram features, and local word order is taken into account, to realize to the pending problem more
Accurate classification.
In a second embodiment, after the step S40, further include:
Step S401:Calculate primary standard problem in the knowledge base of the target criteria problem and the intelligent customer service it
Between the second similarity.
It should be understood that each primary standard problem in knowledge base can be segmented, the TF-IDF values of word, Jiang Geyuan are calculated
Beginning typical problem is expressed as the term vector formed with the TF-IDF values of word and word, calculates the corresponding word of each primary standard problem
COS distance between vectorial term vector corresponding with the target criteria problem, and using the COS distance as second phase
Like degree.
Step S402:The maximum value in second similarity is obtained, judges whether the maximum value is default more than second
Threshold value.
It will be appreciated that obtain the maximum value in second similarity, the corresponding primary standard problem of the maximum value with
The target criteria problem is most close, whether judges the maximum value more than the second predetermined threshold value, if being more than described second default
Threshold value illustrates that target criteria problem primary standard problem correlation corresponding with the maximum value in knowledge base is very high.
Step S403:If the maximum value is more than second predetermined threshold value, the corresponding original mark of the maximum value is obtained
Quasi- problem is used as and waits for supplementary question.
It should be noted that if the maximum value is more than second predetermined threshold value, for example second predetermined threshold value is
80%, if the maximum value is 81% more than 80%, then it is assumed that the corresponding primary standard problem of maximum value and the target
Typical problem correlation is very high, in some instances it may even be possible to substantially be the same problem, only be had differences in statement, can suggest will be described
Target criteria problem primary standard problem corresponding with the maximum value merges, and the corresponding primary standard of the maximum value is asked
The target criteria problem is then waited for that the scaling problem of supplementary question adds to knowledge base by topic as supplementary question is waited for as described in
In.
Correspondingly, the step S50, specifically includes:
Step S501:The target criteria problem and the reply of corresponding target criteria are added to knowing for the intelligent customer service
Know in the scaling problem list for waiting for supplementary question described in library.
In the concrete realization, if the maximum value in second similarity is more than second predetermined threshold value, described in explanation
The corresponding primary standard problem of maximum value and the target criteria problem correlation are very high, then can by the target criteria problem and
Corresponding target criteria reply is added in the scaling problem list for waiting for supplementary question described in the knowledge base of the intelligent customer service,
When user's inquiry is to described some problem when supplementary question or in the scaling problem list when supplementary question, intelligence visitor
Other problems in the scaling problem list for waiting for supplementary question can be also shown by clothes, to more fully be carried for user
For the related data information for wanting to ask questions.
In a second embodiment, it is more than the first predetermined threshold value by the frequency according to the frequency of the dont answer problem
Dont answer problem obtains knowledge type problem as pending problem, to which the pending problem is carried out classification, and frequency is low
Dont answer problem rejected, knowledge base supplement is improved while reducing the follow-up workload for determining target criteria problem
The quality of target criteria problem;Calculate the target criteria problem and the primary standard problem in the knowledge base of the intelligent customer service
Between the second similarity take the maximum value if maximum value in second similarity is more than second predetermined threshold value
The target criteria problem and the reply of corresponding target criteria are added to by corresponding primary standard problem as supplementary question is waited for
In the scaling problem list for waiting for supplementary question described in the knowledge base of the intelligent customer service, think to more fully provide to the user
The related data information asked questions.
It is the flow diagram for the method 3rd embodiment that the present invention improves intelligent customer service response rate, base with reference to Fig. 4, Fig. 4
In above-mentioned second embodiment shown in Fig. 3, propose that the present invention improves the 3rd embodiment of the method for intelligent customer service response rate.
In the third embodiment, after the step S402, further include:
Step S404:If the maximum value is less than second predetermined threshold value, by each clustering cluster in addition to target criteria
Other knowledge type problems except problem are as the first scaling problem.
If it should be understood that the maximum value in second similarity is less than second predetermined threshold value, described in explanation
The corresponding primary standard problem of maximum value and the target criteria problem correlation be not high, then can by the target criteria problem with
Question of independence form is added in the knowledge base.Each problem in the knowledge base is all with primary standard problem and corresponding
The form of scaling problem list stores, so as to some in user's inquiry to some primary standard problem or its scaling problem list
When problem, intelligent customer service can also open up the other problems in the primary standard problem and corresponding scaling problem list
Show, to more fully provide the related data information for wanting to ask questions to the user.
It is added in the knowledge base it will be appreciated that the target criteria problem can be used as question of independence form, then institute
Other knowledge type problems in the corresponding clustering cluster of target criteria problem other than the target criteria problem are stated as first
Scaling problem, first scaling problem can be used as the problems in the scaling problem list of the target criteria problem.
Step S405:Calculate the third similarity between first scaling problem and corresponding target criteria problem.
It should be noted that the possible quantity of the problems in each clustering cluster is more, it is not necessarily to all problems in each clustering cluster
Scaling problem all as the target criteria problem is added to together in the knowledge base.First extension can be then calculated to ask
Topic and the third similarity between corresponding target criteria problem ask the first larger extension of the numerical value of the third similarity
Inscribe the problems in the scaling problem list as the target criteria problem.
In the present embodiment, the step S405, including:Calculate first scaling problem and corresponding target criteria problem
Between statistical nature, semantic feature and theme feature;By logistic regression by the statistical nature, the semantic feature and institute
It states theme feature to be polymerize, obtains the third similarity between first scaling problem and corresponding target criteria problem.
Wherein, the statistical nature includes:Term co-occurrence rate, TF-IDF values, editing distance and Longest Common Substring.Described first can be calculated
Term co-occurrence rate, TF-IDF values, editing distance and Longest Common Substring between scaling problem and corresponding target criteria problem are made
For the statistical nature.Expanded based on shot and long term memory network (Long Short-Term Memory, LSTM) structure described first
The vector of exhibition problem and the target criteria problem calculates the cosine similarity between the vector, which is made
For the semantic feature.Model (Latent Dirichlet Allocation, LDA) is generated to described the by document subject matter
One scaling problem and corresponding target criteria problem carry out the generation of corresponding theme feature.Logistic regression (the Logistic
Regression, LR) on the basis of linear regression, a logical function has been applied mechanically, it can will be described by the logistic regression
Statistical nature, the semantic feature and the theme feature are polymerize, and first scaling problem and corresponding target are obtained
Third similarity between typical problem.
Step S406:N number of is selected in first scaling problem according to the third similarity descending order
Scaling problem list of one scaling problem as the target criteria problem, the N are the integer more than or equal to 1.
In the concrete realization, in order to avoid adding extension of the not high problem of multimass in the target criteria problem
In problem list, N number of can be usually selected in first scaling problem according to the third similarity descending order
One scaling problem, the N be integer more than or equal to 1, that is to say, that will be big with the target criteria problem similarity degree
Scaling problem list of N number of first scaling problem as the target criteria problem.
In the third embodiment, the step S50, specifically includes:
Step S502:By the scaling problem list of the target criteria problem, the target criteria problem and corresponding mesh
The reply of mark standard is added in the form of question of independence in the knowledge base of the intelligent customer service.
It should be noted that the maximum value in second similarity is less than second predetermined threshold value, described in explanation
The corresponding primary standard problem of maximum value and the target criteria problem correlation be not high, then by the expansion of the target criteria problem
Exhibition problem list, the target criteria problem and the reply of corresponding target criteria are added to the intelligence visitor in the form of question of independence
In the knowledge base of clothes, user next time again the problem of target criteria problem correlation described in inquiry when, the intelligent customer service can be from knowledge
The target criteria problem and corresponding scaling problem list are found in library, to realize response user.
In the third embodiment, if the maximum value is less than second predetermined threshold value, by each clustering cluster in addition to mesh
Other knowledge type problems except typical problem are marked as the first scaling problem, calculate first scaling problem and corresponding mesh
The third similarity between typical problem is marked, according to the third similarity descending order in first scaling problem
Select N number of first scaling problem as the scaling problem list of the target criteria problem, the N is whole more than or equal to 1
Number;The scaling problem list of the target criteria problem, the target criteria problem and corresponding target criteria are replied with only
Vertical problem form is added in the knowledge base of the intelligent customer service so that user next time, target criteria problem described in inquiry was related again
The problem of when, the intelligent customer service can find the target criteria problem and corresponding scaling problem list from knowledge base,
To realize response user, user experience is promoted.
In addition, the embodiment of the present invention also proposes a kind of storage medium, is stored on the storage medium and be improved intelligent customer service
The program of response rate, the program for improving intelligent customer service response rate realize raising intelligence as described above when being executed by processor
The step of method of energy customer service response rate.
In addition, with reference to Fig. 5, the embodiment of the present invention also proposes a kind of device improving intelligent customer service response rate, the raising
The device of intelligent customer service response rate includes:Sort module 10, computing module 20, cluster module 30, acquisition module 40 and addition mould
Block 50;
The sort module 10, the dont answer problem for obtaining intelligent customer service classify to the dont answer problem,
Obtain knowledge type problem;
The computing module 20, for calculating the first similarity between each knowledge type problem;
The cluster module 30, for being asked each knowledge type by affine propagation clustering algorithm according to first similarity
Topic is clustered, and multiple clustering clusters are obtained;
The acquisition module 40, for using the cluster centre of each clustering cluster be used as target criteria problem, and acquisition with it is described
The corresponding target criteria of target criteria problem is replied;
The add module 50, for the target criteria problem and the reply of corresponding target criteria to be added to the intelligence
In the knowledge base of energy customer service.
In the present embodiment, the dont answer problem of intelligent customer service is obtained, is classified to the dont answer problem, knowledge is obtained
Type problem, to reject classification obtain non-knowledge type problem, the non-knowledge type problem do not have standard reply, without into
Row addition considers, to improve the quality for the target criteria problem subsequently added;By between the calculating knowledge type problem
First similarity obtains multiple clustering clusters, by the cluster of each clustering cluster according to the first similarity by affine propagation clustering algorithm
Center is added to as target criteria problem in the knowledge base of intelligent customer service, and the cluster centre is most generation in each clustering cluster
The problem of table, is added to as target criteria problem in intelligent customer service, can effectively improve the response rate of intelligent customer service,
Promote user experience.
It should be noted that herein, the terms "include", "comprise" or its any other variant are intended to non-row
His property includes, so that process, method, article or system including a series of elements include not only those elements, and
And further include other elements that are not explicitly listed, or further include for this process, method, article or system institute it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including this
There is also other identical elements in the process of element, method, article or system.
The embodiments of the present invention are for illustration only, can not represent the quality of embodiment.If listing equipment for drying
Unit claim in, several in these devices can be embodied by the same hardware branch.Word first,
Second and the use of third etc. do not indicate that any sequence, can be title by these word explanations.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment side
Method can add the mode of required general hardware platform to realize by software, naturally it is also possible to by hardware, but in many cases
The former is more preferably embodiment.Based on this understanding, technical scheme of the present invention substantially in other words does the prior art
Going out the part of contribution can be expressed in the form of software products, which is stored in a storage medium
(such as read-only memory mirror image (Read Only Memory image, ROM)/random access memory (Random Access
Memory, RAM), magnetic disc, CD) in, including some instructions use so that a station terminal equipment (can be mobile phone, computer,
Server, air conditioner or network equipment etc.) execute method described in each embodiment of the present invention.
It these are only the preferred embodiment of the present invention, be not intended to limit the scope of the invention, it is every to utilize this hair
Equivalent structure or equivalent flow shift made by bright specification and accompanying drawing content is applied directly or indirectly in other relevant skills
Art field, is included within the scope of the present invention.
Claims (10)
1. a kind of method improving intelligent customer service response rate, which is characterized in that the method packet for improving intelligent customer service response rate
Include following steps:
The dont answer problem for obtaining intelligent customer service, classifies to the dont answer problem, obtains knowledge type problem;
Calculate the first similarity between each knowledge type problem;
Each knowledge type problem is clustered by affine propagation clustering algorithm according to first similarity, obtains multiple clusters
Cluster;
Using the cluster centre of each clustering cluster as target criteria problem, and obtain target mark corresponding with the target criteria problem
Standard is replied;
The target criteria problem and the reply of corresponding target criteria are added in the knowledge base of the intelligent customer service.
2. the method for improving intelligent customer service response rate as described in claim 1, which is characterized in that the acquisition intelligent customer service
Dont answer problem classifies to the dont answer problem, obtains knowledge type problem, including:
The dont answer problem for obtaining intelligent customer service, calculates the frequency of the dont answer problem, is more than first to preset by the frequency
The dont answer problem of threshold value is as pending problem;
The pending problem is classified by default disaggregated model, obtains knowledge type problem.
3. the method for improving intelligent customer service response rate as claimed in claim 2, which is characterized in that each knowledge type of calculating is asked
The first similarity between topic, including:
Word segmentation processing is carried out to each knowledge type problem and calculates the word to obtain word all in the knowledge type problem
TF-IDF values;
It is the term vector formed with the TF-IDF values of word and word by each knowledge type problem representation;
The COS distance between each term vector is calculated, and the COS distance is similar as first between each knowledge type problem
Degree.
4. the method for improving intelligent customer service response rate as claimed in claim 3, which is characterized in that described by the clustering cluster
Cluster centre is as target criteria problem, described and after obtaining target criteria reply corresponding with the target criteria problem
Improve intelligent customer service response rate method further include:
It calculates second similar between the target criteria problem and the primary standard problem in the knowledge base of the intelligent customer service
Degree;
The maximum value in second similarity is obtained, judges the maximum value whether more than the second predetermined threshold value;
If the maximum value is more than second predetermined threshold value, obtains the corresponding primary standard problem conduct of the maximum value and wait mending
Fill problem;
It is described that the target criteria problem and the reply of corresponding target criteria are added in the knowledge base of the intelligent customer service, it wraps
It includes:
The target criteria problem and the reply of corresponding target criteria are added to described in the knowledge base of the intelligent customer service and are waited for
In the scaling problem list of supplementary question.
5. the method for improving intelligent customer service response rate as claimed in claim 4, which is characterized in that described to obtain second phase
Like the maximum value in degree, after judging the maximum value whether more than the second predetermined threshold value, the raising intelligent customer service response rate
Method further include:
If the maximum value is less than second predetermined threshold value, by other in each clustering cluster other than target criteria problem
Knowledge type problem is as the first scaling problem;
Calculate the third similarity between first scaling problem and corresponding target criteria problem;
Selected in first scaling problem according to the third similarity descending order N number of first scaling problem as
The scaling problem list of the target criteria problem, the N are the integer more than or equal to 1;
It is described that the target criteria problem and the reply of corresponding target criteria are added in the knowledge base of the intelligent customer service, it wraps
It includes:
The scaling problem list of the target criteria problem, the target criteria problem and corresponding target criteria are replied with only
Vertical problem form is added in the knowledge base of the intelligent customer service.
6. the method for improving intelligent customer service response rate as claimed in claim 5, which is characterized in that described to calculate first expansion
Third similarity between exhibition problem and corresponding target criteria problem, including:
Statistical nature, semantic feature and the theme calculated between first scaling problem and corresponding target criteria problem is special
Sign;
The statistical nature, the semantic feature and the theme feature are polymerize by logistic regression, obtain described the
Third similarity between one scaling problem and corresponding target criteria problem.
7. the method for improving intelligent customer service response rate as claimed in claim 6, which is characterized in that the statistical nature includes:
Term co-occurrence rate, TF-IDF values, editing distance and Longest Common Substring.
8. a kind of equipment improving intelligent customer service response rate, which is characterized in that the equipment packet for improving intelligent customer service response rate
It includes:Memory, processor and the raising intelligent customer service response that is stored on the memory and can run on the processor
The program of rate, the program for improving intelligent customer service response rate are realized when being executed by the processor as appointed in claim 1 to 7
The step of method of raising intelligent customer service response rate described in one.
9. a kind of storage medium, which is characterized in that storage is improved the program of intelligent customer service response rate, institute on the storage medium
It states the program for improving intelligent customer service response rate and realizes raising as described in any one of claim 1 to 7 when being executed by processor
The step of method of intelligent customer service response rate.
10. a kind of device improving intelligent customer service response rate, which is characterized in that the device packet for improving intelligent customer service response rate
It includes:Sort module, computing module, cluster module, acquisition module and add module;
The sort module, the dont answer problem for obtaining intelligent customer service are classified to the dont answer problem, are known
Knowledge type problem;
The computing module, for calculating the first similarity between each knowledge type problem;
The cluster module, for being carried out each knowledge type problem by affine propagation clustering algorithm according to first similarity
Cluster, obtains multiple clustering clusters;
The acquisition module is used for using the cluster centre of each clustering cluster as target criteria problem, and is obtained and the target mark
The corresponding target criteria of quasi- problem is replied;
The add module, for the target criteria problem and the reply of corresponding target criteria to be added to the intelligent customer service
Knowledge base in.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810499843.9A CN108804567A (en) | 2018-05-22 | 2018-05-22 | Improve method, equipment, storage medium and the device of intelligent customer service response rate |
PCT/CN2018/092989 WO2019223058A1 (en) | 2018-05-22 | 2018-06-27 | Method, device and apparatus for improving response rate of intelligent customer service, and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810499843.9A CN108804567A (en) | 2018-05-22 | 2018-05-22 | Improve method, equipment, storage medium and the device of intelligent customer service response rate |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108804567A true CN108804567A (en) | 2018-11-13 |
Family
ID=64091417
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810499843.9A Pending CN108804567A (en) | 2018-05-22 | 2018-05-22 | Improve method, equipment, storage medium and the device of intelligent customer service response rate |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN108804567A (en) |
WO (1) | WO2019223058A1 (en) |
Cited By (25)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109688281A (en) * | 2018-12-03 | 2019-04-26 | 复旦大学 | A kind of intelligent sound exchange method and system |
CN109783582A (en) * | 2018-12-04 | 2019-05-21 | 平安科技(深圳)有限公司 | A kind of knowledge base alignment schemes, device, computer equipment and storage medium |
CN109800879A (en) * | 2018-12-21 | 2019-05-24 | 科大讯飞股份有限公司 | Construction of knowledge base method and apparatus |
CN109885651A (en) * | 2019-01-16 | 2019-06-14 | 平安科技(深圳)有限公司 | A kind of question pushing method and device |
CN110555101A (en) * | 2019-09-09 | 2019-12-10 | 浙江诺诺网络科技有限公司 | customer service knowledge base updating method, device, equipment and storage medium |
CN110941708A (en) * | 2019-11-04 | 2020-03-31 | 智器云南京信息科技有限公司 | Intelligent question and answer library establishing method, intelligent question and answer method and device and computer equipment |
CN110955766A (en) * | 2019-11-29 | 2020-04-03 | 支付宝(杭州)信息技术有限公司 | Method and system for automatically expanding intelligent customer service standard problem pairs |
CN111159370A (en) * | 2019-12-20 | 2020-05-15 | 中国建设银行股份有限公司 | Short-session new problem generation method, storage medium and man-machine interaction device |
CN111368060A (en) * | 2020-05-27 | 2020-07-03 | 支付宝(杭州)信息技术有限公司 | Self-learning method, device and system for conversation robot, electronic equipment and medium |
CN111382235A (en) * | 2018-12-27 | 2020-07-07 | 上海智臻智能网络科技股份有限公司 | Question-answer knowledge base optimization method and device |
CN111488442A (en) * | 2020-04-09 | 2020-08-04 | 深圳追一科技有限公司 | Data processing method, device, server and storage medium |
CN111611781A (en) * | 2020-05-27 | 2020-09-01 | 北京妙医佳健康科技集团有限公司 | Data labeling method, question answering method, device and electronic equipment |
CN111859985A (en) * | 2020-07-23 | 2020-10-30 | 平安普惠企业管理有限公司 | AI customer service model testing method, device, electronic equipment and storage medium |
CN112015899A (en) * | 2020-08-31 | 2020-12-01 | 中国银行股份有限公司 | Method and device for supplementing knowledge base, server and computer storage medium |
CN112115248A (en) * | 2020-09-27 | 2020-12-22 | 支付宝(杭州)信息技术有限公司 | Method and system for extracting conversation strategy structure from conversation corpus |
CN112148852A (en) * | 2020-09-10 | 2020-12-29 | 深圳康佳电子科技有限公司 | Intelligent customer service method and device, storage medium and computer equipment |
CN112541067A (en) * | 2020-12-15 | 2021-03-23 | 平安科技(深圳)有限公司 | Knowledge base problem mining method and device, electronic equipment and storage medium |
CN112597292A (en) * | 2020-12-29 | 2021-04-02 | 招联消费金融有限公司 | Question reply recommendation method and device, computer equipment and storage medium |
WO2021114834A1 (en) * | 2020-06-24 | 2021-06-17 | 平安科技(深圳)有限公司 | Customer service question update method and system, terminal device, and computer storage medium |
CN113139039A (en) * | 2021-05-06 | 2021-07-20 | 京东数字科技控股股份有限公司 | Dialogue data processing method, device, equipment and storage medium |
CN113392208A (en) * | 2020-03-12 | 2021-09-14 | 中国移动通信集团云南有限公司 | Method, device and storage medium for IT operation and maintenance fault processing experience accumulation |
CN113779050A (en) * | 2020-06-23 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Method and device for managing knowledge base of customer service robot |
US11615695B2 (en) | 2018-06-12 | 2023-03-28 | Intergraph Corporation | Coverage agent for computer-aided dispatch systems |
WO2023165172A1 (en) * | 2022-03-02 | 2023-09-07 | 北京沃东天骏信息技术有限公司 | E-mail tracking method and apparatus |
CN109885651B (en) * | 2019-01-16 | 2024-06-04 | 平安科技(深圳)有限公司 | Question pushing method and device |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102637192A (en) * | 2012-02-17 | 2012-08-15 | 清华大学 | Method for answering with natural language |
CN103810218A (en) * | 2012-11-14 | 2014-05-21 | 北京百度网讯科技有限公司 | Problem cluster-based automatic asking and answering method and device |
CN105955965A (en) * | 2016-06-21 | 2016-09-21 | 上海智臻智能网络科技股份有限公司 | Question information processing method and device |
CN105975460A (en) * | 2016-05-30 | 2016-09-28 | 上海智臻智能网络科技股份有限公司 | Question information processing method and device |
CN106547734A (en) * | 2016-10-21 | 2017-03-29 | 上海智臻智能网络科技股份有限公司 | A kind of question sentence information processing method and device |
CN107656948A (en) * | 2016-11-14 | 2018-02-02 | 平安科技(深圳)有限公司 | The problem of in automatically request-answering system clustering processing method and device |
CN107784105A (en) * | 2017-10-26 | 2018-03-09 | 平安科技(深圳)有限公司 | Construction of knowledge base method, electronic installation and storage medium based on magnanimity problem |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9563688B2 (en) * | 2014-05-01 | 2017-02-07 | International Business Machines Corporation | Categorizing users based on similarity of posed questions, answers and supporting evidence |
CN104036015A (en) * | 2014-06-24 | 2014-09-10 | 北京奇虎科技有限公司 | Electronic terminal question classification method and device, and solution provision method, system and device based on electronic terminal question classification device and method |
-
2018
- 2018-05-22 CN CN201810499843.9A patent/CN108804567A/en active Pending
- 2018-06-27 WO PCT/CN2018/092989 patent/WO2019223058A1/en active Application Filing
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102637192A (en) * | 2012-02-17 | 2012-08-15 | 清华大学 | Method for answering with natural language |
CN103810218A (en) * | 2012-11-14 | 2014-05-21 | 北京百度网讯科技有限公司 | Problem cluster-based automatic asking and answering method and device |
CN105975460A (en) * | 2016-05-30 | 2016-09-28 | 上海智臻智能网络科技股份有限公司 | Question information processing method and device |
CN105955965A (en) * | 2016-06-21 | 2016-09-21 | 上海智臻智能网络科技股份有限公司 | Question information processing method and device |
CN106547734A (en) * | 2016-10-21 | 2017-03-29 | 上海智臻智能网络科技股份有限公司 | A kind of question sentence information processing method and device |
CN107656948A (en) * | 2016-11-14 | 2018-02-02 | 平安科技(深圳)有限公司 | The problem of in automatically request-answering system clustering processing method and device |
CN107784105A (en) * | 2017-10-26 | 2018-03-09 | 平安科技(深圳)有限公司 | Construction of knowledge base method, electronic installation and storage medium based on magnanimity problem |
Cited By (33)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11615695B2 (en) | 2018-06-12 | 2023-03-28 | Intergraph Corporation | Coverage agent for computer-aided dispatch systems |
US11735028B2 (en) | 2018-06-12 | 2023-08-22 | Intergraph Corporation | Artificial intelligence applications for computer-aided dispatch systems |
CN109688281A (en) * | 2018-12-03 | 2019-04-26 | 复旦大学 | A kind of intelligent sound exchange method and system |
CN109783582A (en) * | 2018-12-04 | 2019-05-21 | 平安科技(深圳)有限公司 | A kind of knowledge base alignment schemes, device, computer equipment and storage medium |
CN109783582B (en) * | 2018-12-04 | 2023-08-15 | 平安科技(深圳)有限公司 | Knowledge base alignment method, device, computer equipment and storage medium |
CN109800879A (en) * | 2018-12-21 | 2019-05-24 | 科大讯飞股份有限公司 | Construction of knowledge base method and apparatus |
CN111382235A (en) * | 2018-12-27 | 2020-07-07 | 上海智臻智能网络科技股份有限公司 | Question-answer knowledge base optimization method and device |
CN109885651B (en) * | 2019-01-16 | 2024-06-04 | 平安科技(深圳)有限公司 | Question pushing method and device |
CN109885651A (en) * | 2019-01-16 | 2019-06-14 | 平安科技(深圳)有限公司 | A kind of question pushing method and device |
CN110555101A (en) * | 2019-09-09 | 2019-12-10 | 浙江诺诺网络科技有限公司 | customer service knowledge base updating method, device, equipment and storage medium |
CN110941708A (en) * | 2019-11-04 | 2020-03-31 | 智器云南京信息科技有限公司 | Intelligent question and answer library establishing method, intelligent question and answer method and device and computer equipment |
CN110941708B (en) * | 2019-11-04 | 2023-02-28 | 智器云南京信息科技有限公司 | Intelligent question and answer library establishing method, intelligent question and answer method and device and computer equipment |
CN110955766A (en) * | 2019-11-29 | 2020-04-03 | 支付宝(杭州)信息技术有限公司 | Method and system for automatically expanding intelligent customer service standard problem pairs |
CN111159370A (en) * | 2019-12-20 | 2020-05-15 | 中国建设银行股份有限公司 | Short-session new problem generation method, storage medium and man-machine interaction device |
CN113392208A (en) * | 2020-03-12 | 2021-09-14 | 中国移动通信集团云南有限公司 | Method, device and storage medium for IT operation and maintenance fault processing experience accumulation |
CN111488442A (en) * | 2020-04-09 | 2020-08-04 | 深圳追一科技有限公司 | Data processing method, device, server and storage medium |
CN111611781A (en) * | 2020-05-27 | 2020-09-01 | 北京妙医佳健康科技集团有限公司 | Data labeling method, question answering method, device and electronic equipment |
CN111368060A (en) * | 2020-05-27 | 2020-07-03 | 支付宝(杭州)信息技术有限公司 | Self-learning method, device and system for conversation robot, electronic equipment and medium |
CN111611781B (en) * | 2020-05-27 | 2023-08-18 | 北京妙医佳健康科技集团有限公司 | Data labeling method, question answering device and electronic equipment |
CN113779050A (en) * | 2020-06-23 | 2021-12-10 | 北京沃东天骏信息技术有限公司 | Method and device for managing knowledge base of customer service robot |
WO2021114834A1 (en) * | 2020-06-24 | 2021-06-17 | 平安科技(深圳)有限公司 | Customer service question update method and system, terminal device, and computer storage medium |
CN111859985A (en) * | 2020-07-23 | 2020-10-30 | 平安普惠企业管理有限公司 | AI customer service model testing method, device, electronic equipment and storage medium |
CN111859985B (en) * | 2020-07-23 | 2023-09-12 | 上海华期信息技术有限责任公司 | AI customer service model test method and device, electronic equipment and storage medium |
CN112015899B (en) * | 2020-08-31 | 2023-08-11 | 中国银行股份有限公司 | Method, device, server and computer storage medium for supplementing knowledge base |
CN112015899A (en) * | 2020-08-31 | 2020-12-01 | 中国银行股份有限公司 | Method and device for supplementing knowledge base, server and computer storage medium |
CN112148852A (en) * | 2020-09-10 | 2020-12-29 | 深圳康佳电子科技有限公司 | Intelligent customer service method and device, storage medium and computer equipment |
CN112115248A (en) * | 2020-09-27 | 2020-12-22 | 支付宝(杭州)信息技术有限公司 | Method and system for extracting conversation strategy structure from conversation corpus |
CN112115248B (en) * | 2020-09-27 | 2023-11-14 | 支付宝(杭州)信息技术有限公司 | Method and system for extracting dialogue strategy structure from dialogue corpus |
CN112541067A (en) * | 2020-12-15 | 2021-03-23 | 平安科技(深圳)有限公司 | Knowledge base problem mining method and device, electronic equipment and storage medium |
CN112597292A (en) * | 2020-12-29 | 2021-04-02 | 招联消费金融有限公司 | Question reply recommendation method and device, computer equipment and storage medium |
CN112597292B (en) * | 2020-12-29 | 2024-04-26 | 招联消费金融股份有限公司 | Question reply recommendation method, device, computer equipment and storage medium |
CN113139039A (en) * | 2021-05-06 | 2021-07-20 | 京东数字科技控股股份有限公司 | Dialogue data processing method, device, equipment and storage medium |
WO2023165172A1 (en) * | 2022-03-02 | 2023-09-07 | 北京沃东天骏信息技术有限公司 | E-mail tracking method and apparatus |
Also Published As
Publication number | Publication date |
---|---|
WO2019223058A1 (en) | 2019-11-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108804567A (en) | Improve method, equipment, storage medium and the device of intelligent customer service response rate | |
CN108021929B (en) | Big data-based mobile terminal e-commerce user portrait establishing and analyzing method and system | |
CN108182279B (en) | Object classification method, device and computer equipment based on text feature | |
CN105468742B (en) | The recognition methods of malice order and device | |
CN103117903B (en) | Surfing flow method for detecting abnormality and device | |
CN110598157B (en) | Target information identification method, device, equipment and storage medium | |
CN109509010B (en) | Multimedia information processing method, terminal and storage medium | |
EP2973038A1 (en) | Classifying resources using a deep network | |
CN103473036B (en) | A kind of input method skin method for pushing and system | |
CN111222647A (en) | Federal learning system optimization method, device, equipment and storage medium | |
CN110766438B (en) | Method for analyzing user behavior of power grid user through artificial intelligence | |
US20180012237A1 (en) | Inferring user demographics through categorization of social media data | |
CN110472016A (en) | Article recommended method, device, electronic equipment and storage medium | |
CN112579822A (en) | Video data pushing method and device, computer equipment and storage medium | |
CN114091472B (en) | Training method of multi-label classification model | |
CN109493186A (en) | The method and apparatus for determining pushed information | |
CN112330055A (en) | User complaint prediction method and device | |
CN105408894A (en) | Method and device for determining user identity category | |
CN106649380A (en) | Hot spot recommendation method and system based on tag | |
CN108108912A (en) | Method of discrimination, device, server and the storage medium of interactive low quality user | |
CN112925634A (en) | Heterogeneous resource scheduling method and system | |
CN110532448B (en) | Document classification method, device, equipment and storage medium based on neural network | |
CN112231299A (en) | Method and device for dynamically adjusting feature library | |
Yang et al. | Ott messages modeling and classification based on recurrent neural networks | |
CN108460475A (en) | Poor student's prediction technique and device based on network playing by students behavior |
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 |