CN110489519A - The session method and Related product of dialogue-based prediction model - Google Patents
The session method and Related product of dialogue-based prediction model Download PDFInfo
- Publication number
- CN110489519A CN110489519A CN201910605730.7A CN201910605730A CN110489519A CN 110489519 A CN110489519 A CN 110489519A CN 201910605730 A CN201910605730 A CN 201910605730A CN 110489519 A CN110489519 A CN 110489519A
- Authority
- CN
- China
- Prior art keywords
- session
- prediction model
- product
- target user
- historical
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 90
- 230000008569 process Effects 0.000 claims abstract description 24
- 238000012549 training Methods 0.000 claims description 42
- 230000006870 function Effects 0.000 claims description 21
- 238000005457 optimization Methods 0.000 claims description 21
- 238000004590 computer program Methods 0.000 claims description 10
- 238000004422 calculation algorithm Methods 0.000 claims description 8
- 238000003860 storage Methods 0.000 claims description 8
- 238000013459 approach Methods 0.000 claims description 7
- 230000005540 biological transmission Effects 0.000 claims description 5
- 238000004891 communication Methods 0.000 claims description 5
- 238000010801 machine learning Methods 0.000 claims description 5
- 230000002452 interceptive effect Effects 0.000 claims description 4
- 230000000977 initiatory effect Effects 0.000 claims description 3
- 238000012512 characterization method Methods 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 8
- 230000003993 interaction Effects 0.000 description 7
- 230000002996 emotional effect Effects 0.000 description 5
- 230000005236 sound signal Effects 0.000 description 5
- 230000009471 action Effects 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000011478 gradient descent method Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000007405 data analysis Methods 0.000 description 2
- 230000008030 elimination Effects 0.000 description 2
- 238000003379 elimination reaction Methods 0.000 description 2
- 230000008451 emotion Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000005192 partition Methods 0.000 description 2
- 230000001052 transient effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007717 exclusion Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000036651 mood Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000012360 testing method 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
- 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
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/06—Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/06—Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
- G10L15/063—Training
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Computer Interaction (AREA)
- Mathematical Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Databases & Information Systems (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Economics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Evolutionary Computation (AREA)
- Game Theory and Decision Science (AREA)
- Medical Informatics (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Machine Translation (AREA)
Abstract
The embodiment of the present application discloses the session method and Related product of a kind of dialogue-based prediction model, this method comprises: the selection target user from customer data base；It initiates a session request to the target user, receives the first return information that the target user is directed to the session request；First return information is predicted using session prediction model, obtains object statement；The object statement is sent to the target user, to establish session process with the target user.The application is conducive to improve the initiative of session process, improves user experience.
Description
Technical field
This application involves field of artificial intelligence, and in particular to a kind of session method and phase of dialogue-based prediction model
Close product.
Background technique
With the development of artificial intelligence technology, artificial intelligence product is gradually applied to each scene in life, for example, objective
The problem of intelligent robot can propose user is taken to answer, so that human cost is saved, still, current customer service intelligence
Answer is searched in the question that robot is all based on user from the answer library pre-set, and the answer searched is sent to
User；For another example, marketing robot only can question according to user to product, carry out product characteristic introduction.So existing intelligent machine
Device is all passive type to user's marketing product, conversates and exchanges with user without active, causes marketing efficiency low, user
It is poor to experience.
Summary of the invention
The embodiment of the present application provides the session method and Related product of a kind of dialogue-based prediction model, by introducing meeting
Prediction model is talked about, realizes that intelligent robot actively conversates with user and exchanges, improve marketing effectiveness, and then improve user's body
It tests.
In a first aspect, the embodiment of the present application provides a kind of session method of dialogue-based prediction model, comprising:
The selection target user from customer data base；
It initiates a session request to the target user, receives the target user for the first reply of the session request
Information；
First return information is predicted using session prediction model, obtains object statement；
The object statement is sent to the target user, to establish session process with the target user.
Second aspect, the embodiment of the present application provide a kind of session platform of dialogue-based prediction model, comprising:
Selecting unit, for the selection target user from customer data base；
Interactive unit receives the target user for the session for initiating a session request to the target user
First return information of request；
Predicting unit obtains object statement for predicting using session prediction model first return information；
Transmission unit, for sending the object statement to the target user, to establish session with the target user
Process.
The third aspect, the embodiment of the present application provide a kind of electronic equipment, including processor, memory, communication interface and
One or more programs, wherein one or more of programs are stored in the memory, and are configured by described
It manages device to execute, described program is included the steps that for executing the instruction in method as described in relation to the first aspect.
Fourth aspect, the embodiment of the present application provide a kind of computer readable storage medium, the computer-readable storage medium
Matter is stored with computer program, and the computer program makes the method for computer execution as described in relation to the first aspect.
5th aspect, the embodiment of the present application provide a kind of computer program product, and the computer program product includes depositing
The non-transient computer readable storage medium of computer program is stored up, the computer is operable to make computer to execute such as the
Method described in one side.
Implement the embodiment of the present application, has the following beneficial effects:
As can be seen that in the embodiment of the present application, actively initiated a session request to user by introducing session prediction model,
And the return information of user is predicted, corresponding object statement is automatically generated, is carried out independently to realize with user
Property exchange, improve the efficiency of product marketing, improve user experience.
Detailed description of the invention
In order to more clearly explain the technical solutions in the embodiments of the present application, make required in being described below to embodiment
Attached drawing is briefly described, it should be apparent that, the accompanying drawings in the following description is some embodiments of the present application, for ability
For the those of ordinary skill of domain, without creative efforts, it can also be obtained according to these attached drawings other attached
Figure.
Figure 1A is a kind of flow diagram of the session method of dialogue-based prediction model provided by the embodiments of the present application；
Figure 1B is a kind of process of method that session process is established by the outer paging system of intelligence provided by the embodiments of the present application
Schematic diagram；
Fig. 2 is the flow diagram of the session method of the dialogue-based prediction model of another kind provided by the embodiments of the present application；
Fig. 3 is the flow diagram of the session method of the dialogue-based prediction model of another kind provided by the embodiments of the present application；
Fig. 4 is a kind of structural schematic diagram of the session platform of dialogue-based prediction model provided by the embodiments of the present application；
Fig. 5 is that a kind of functional unit of session platform of dialogue-based prediction model provided by the embodiments of the present application forms frame
Figure.
Specific embodiment
Below in conjunction with the attached drawing in the embodiment of the present application, technical solutions in the embodiments of the present application carries out clear, complete
Site preparation description, it is clear that described embodiment is some embodiments of the present application, instead of all the embodiments.Based on this Shen
Please in embodiment, every other implementation obtained by those of ordinary skill in the art without making creative efforts
Example, shall fall in the protection scope of this application.
The description and claims of this application and term " first ", " second ", " third " and " in the attached drawing
Four " etc. are not use to describe a particular order for distinguishing different objects.In addition, term " includes " and " having " and it
Any deformation, it is intended that cover and non-exclusive include.Such as it contains the process, method of a series of steps or units, be
System, product or equipment are not limited to listed step or unit, but optionally further comprising the step of not listing or list
Member, or optionally further comprising other step or units intrinsic for these process, methods, product or equipment.
Referenced herein " embodiment " is it is meant that the special characteristic, result or the characteristic that describe can wrap in conjunction with the embodiments
It is contained at least one embodiment of the application.Each position in the description occur the phrase might not each mean it is identical
Embodiment, nor the independent or alternative embodiment with other embodiments mutual exclusion.Those skilled in the art explicitly and
Implicitly understand, embodiment described herein can be combined with other embodiments.
Session platform in the application may include smart phone (such as Android phone, iOS mobile phone, Windows
Phone mobile phone etc.), tablet computer, palm PC, laptop, mobile internet device MID (Mobile Internet
Devices, referred to as: MID) or wearable device etc., above-mentioned session platform is only citing, and non exhaustive, including but not limited to upper
Session platform is stated, certainly in practical applications, session platform is also not necessarily limited to above-mentioned realization form, such as can also include: intelligence
Car-mounted terminal, computer equipment etc..
A refering to fig. 1, Figure 1A are a kind of session method of dialogue-based prediction model provided by the embodiments of the present application, the party
Method is applied to session platform, and this method includes but is not limited to following steps:
101: session platform selection target user from customer data base.
Wherein, each user in customer data base has user tag, and the user tag is for characterizing whether user has
There is purchasing demand, i.e. whether user has the demand for buying product, and target user characterizes the user for user tag, and there is purchase to need
The user asked, wherein the user tag can carry out big data analysis by the consumption data to user and obtain, and establish user's mark
The process of label is the prior art, herein no longer narration in detail.
102: session platform initiates a session request to the target user, receives the target user and asks for the session
The first return information asked.
Optionally, session platform initiates a session request according to preset opening remarks to user, which can be user
The opening remarks or session inputted in advance are based on big data analysis for no reason, obtain target product, then automatically generate and target
The corresponding opening remarks of product.
103: session platform predicts first return information using session prediction model, obtains object statement.
Wherein, session platform predicts the first return information using preparatory trained prediction model, automatically generates
Revert statement, i.e. object statement, to establish the session process of session platform and target user.
104: session platform sends the object statement to the target user, to establish session stream with the target user
Journey.
It is understood that the session that can carry out N wheel with user exchanges after target session is sent to user, thus
Obtain the conversation recording with target user, wherein every wheel session exchange is consistent with above-mentioned implementation, repeats no more.
As can be seen that in the above-described embodiments, session platform actively initiates meeting to user by introducing session prediction model
Words request, and predicts the return information of user, automatically generates corresponding object statement, thus realize and user into
The exchange of row independence improves the efficiency of product marketing, improves user experience.
In some possible embodiments, above-mentioned session prediction model is constructed based on machine learning Integrated Algorithm, described
Method further include: before selection target user in customer data base, the session is predicted using N historical session data
Model optimizes training, and the N is positive integer；Then, first return information is predicted using session prediction model,
The realization process for obtaining object statement can be with are as follows: is replied using the session prediction model after newest optimization training described first and is believed
Breath is predicted, object statement is obtained.Training is optimized to session prediction model by historical session data, it is pre- to improve model
Survey the accuracy of sentence.
In some possible embodiments, every historical session data include product in above-mentioned N historical session data
Demand, the product know-how map built in advance and historical session record, the product know-how map include the attribute letter of product
Breath and non-attribute information；Then, trained realization process is optimized to the session prediction model using N historical session data
It can be with are as follows: using the product demand of historical session data each in N historical session data and product know-how map as input number
According to the session prediction model is input to, the corresponding prediction conversation recording of every historical session data is obtained；Based on every history
The corresponding prediction conversation recording of session data and historical session record, obtain loss function, are based on gradient descent method and the damage
It loses function and training is optimized to the session prediction model.
For example, product know-how map is constructed to each product first, i.e., by the attribute information of each product and non-category
Property information composition sequence, obtains the product know-how figure of product, for example, by the attribute information of product A, such as price, color, function,
It with non-attribute information, such as comments on, forms multidimensional sequence [price, color, function, comment], then, obtain the production in historical data
Product demand and product know-how map, and it regard the historical session record in the historical data as supervision message, mould is predicted to session
Type, which exercises supervision, optimizes training.
In some possible embodiments, it is replied using the session prediction model after newest optimization training described first
Information predicted, the realization process for obtaining object statement can be with are as follows: based on the session prediction model after newest optimization training, obtains
To first eigenvector, the first eigenvector believes corresponding product type for characterizing described first and replying；Based on described
Session prediction model after newest optimization training, obtains second feature vector, and the second feature vector is for characterizing described the
The corresponding product know-how map of one return information；The first eigenvector and the second feature vector are spliced, obtained
To target feature vector；The target feature vector is matched, matched first template vector is obtained；By first mould
The product type and product know-how map of plate vector form sentence, obtain object statement.
In some possible embodiments, the method also includes: obtain and the complete session of target user note
Record；Determine the target user to the satisfaction of any one product according to the complete conversation recording；To the target user
The purchase approach of target product is pushed, the target product is the product that satisfaction is more than or equal to threshold value.Obtain target
User carries out semantic analysis to the paragraph, obtains the satisfaction to the product to the intention paragraph of any one product, thus
When satisfaction is greater than threshold value, to target user with purchase approach is sent, marketing efficiency is improved.
Wherein, satisfaction can be obtained by the semantic mapping relations with satisfaction, which can be 70%, 80%
Or other values.
In some possible embodiments, when the conversation recording is voice conversation, according to the complete session
After record determines the target user to the satisfaction of any one product, the method also includes: obtain the session note
The call voice of record；Emotion identification is carried out to the call voice, obtains emotional information of the target user in conversation procedure, root
The corresponding probability of error of the emotional information is determined according to emotional information and the mapping relations of the probability of error；According to the probability of error
The satisfaction for adjusting target user, obtains the true satisfaction of each user.In the present embodiment, known by mood
Not, dynamic adjustment is carried out to satisfaction, makes to obtain true satisfaction and is more in line with actual conditions, improve the precision of marketing.
Optionally, in above-mentioned possible embodiment, Emotion identification is carried out to the call voice, obtains target user
The realization process of emotional information in conversation procedure can be with are as follows: according to trained voice in advance extract model (for example,
Hourglass model), voice extraction is carried out to the call voice, obtains the first audio signal comprising voice；According to independence
Component Analysis algorithm and Kalman filtering algorithm filter out the ambient noise in first audio signal, obtain only including voice
Second audio signal；The mute frame that second audio signal is filtered out according to voice activity detection algorithms obtains third audio letter
Number；The audio user frame in the third audio signal is obtained, Audio Matching is carried out to audio user frame, obtains target user's
Emotional information.
In some possible embodiments, above-mentioned session request includes: text session and/or voice conversation, in the meeting
When words request is text session, for example, the exchange that can be conversated by way of establishing dialog box, is voice in the session request
When session, exchanged for example, can be conversated by the outer paging system of intelligence with target user.
B refering to fig. 1, Figure 1B are a kind of side that session process is established by the outer paging system of intelligence provided by the embodiments of the present application
The flow diagram of method, this method are applied to the outer paging system of intelligence, the intelligence outside paging system include: core net, it is voice gateways, outer
Module, intelligent interaction module are exhaled, includes the trained session prediction model of above-mentioned optimization, the intelligence in the intelligent interaction module
Outer paging system is applied to session platform, and this method includes but is not limited to following steps:
101a: core net dials to user terminal, and user is waited to connect phone.
101b: after user connects, user terminal acquires user speech, sends core net for user speech.
101c: user speech is linked into outgoing call module by voice gateways by core net.
101d: outgoing call module is to intelligent interaction module incoming call user speech.
101e: intelligent interaction module predicts user speech using session prediction model, obtains object statement.
Intelligent interaction module first carries out noise elimination to user speech, converts text for the voice messaging after elimination noise
Then information carries out feature extraction by the session prediction model after newest optimization training, obtains target feature vector, finally,
Based on the matching to target feature vector, object statement is obtained.
101f: object statement is converted voice by intelligent interaction module, and voice is issued to voice gateways.
101g: voice is issued to user terminal by core net by voice gateways.
101h: user terminal receives voice, and plays.
As can be seen that in the present embodiment, session prediction model being introduced in intelligent interaction module, is predicted by session
Model automatic Prediction object statement promotes the intelligence and initiative of session platform to establish session process with user, improves
Marketing efficiency improves user experience.
Referring to Fig.2, Fig. 2 is the session method of the dialogue-based prediction model of another kind provided by the embodiments of the present application, the party
Method is applied to session platform, and this method includes but is not limited to following steps:
201: session platform optimizes training to session prediction model using N historical session data.
202: session platform selection target user from customer data base.
203: session platform initiates a session request to the target user, receives the target user and asks for the session
The first return information asked.
204: session platform carries out first return information using the session prediction model after newest optimization training pre-
It surveys, obtains object statement.
205: session platform sends the object statement to the target user, to establish session stream with the target user
Journey.
As can be seen that in the above-described embodiments, session platform actively initiates meeting to user by introducing session prediction model
Words request, and predicts the return information of user, automatically generates corresponding object statement, thus realize and user into
The exchange of row independence improves the efficiency of product marketing, improves user experience.
It should be noted that the specific implementation process of each step of method shown in Fig. 2 can be found in described in above-mentioned Figure 1A
The specific implementation process of method, no longer describes herein.
Refering to Fig. 3, Fig. 3 is the session method of the dialogue-based prediction model of another kind provided by the embodiments of the present application, the party
Method is applied to session platform, and this method includes but is not limited to following steps:
301: session platform optimizes training to session prediction model using N historical session data.
302: session platform selection target user from customer data base.
303: session platform initiates a session request to the target user, receives the target user and asks for the session
The first return information asked.
304: session platform carries out first return information using the session prediction model after newest optimization training pre-
It surveys, obtains object statement.
305: session platform sends the object statement to the target user, to establish session stream with the target user
Journey.
306: session platform obtains the complete conversation recording with the target user, is determined according to the complete conversation recording
Satisfaction of the target user to any one product.
307: session platform pushes the purchase approach of target product to the target user, and the target product is to be satisfied with journey
Degree is more than or equal to the product of threshold value.
As can be seen that in the above-described embodiment, session platform is actively initiated to user by introducing session prediction model
Session request, and the return information of user is predicted, corresponding object statement is automatically generated, thus realization and user
The exchange for carrying out independence improves the efficiency of product marketing, improves user experience；In addition, the satisfaction of target user is obtained, it is full
When sufficient condition, purchase approach is pushed to target user, and then improve user experience.
It should be noted that the specific implementation process of each step of method shown in Fig. 3 can be found in described in above-mentioned Figure 1A
The specific implementation process of method, no longer describes herein.
It is consistent with above-mentioned Figure 1A, Fig. 2, embodiment shown in Fig. 3, referring to Fig. 4, Fig. 4 provides for the embodiment of the present application
A kind of dialogue-based prediction model session platform 400 structural schematic diagram, as shown in figure 4, session platform 400 include processing
Device, memory, communication interface and one or more program, wherein said one or multiple programs be different from said one or
Multiple application programs, and said one or multiple programs are stored in above-mentioned memory, and are configured by above-mentioned processor
It executes, above procedure includes the instruction for executing following steps:
The selection target user from customer data base；
It initiates a session request to the target user, receives the target user for the first reply of the session request
Information；
First return information is predicted using session prediction model, obtains object statement；
The object statement is sent to the target user, to establish session process with the target user.
In some possible embodiments, the session prediction model is constructed based on machine learning Integrated Algorithm, above-mentioned
Program is also used to execute the instruction of following steps: before selection target user in customer data base, using N historical session
Data optimize training to the session prediction model, and the N is positive integer；In use session prediction model to described first
Return information is predicted, in terms of obtaining object statement, above procedure is specifically used for executing the instruction of following steps: using newest
Session prediction model after optimization training predicts first return information, obtains object statement.
In some possible embodiments, every historical session data include product in the N historical session data
Demand, the product know-how map built in advance and historical session record, the product know-how map include the attribute letter of product
Breath and non-attribute information；In terms of optimizing training to the session prediction model using N historical session data, above-mentioned journey
Sequence is specifically used for executing the instruction of following steps:
Using the product demand of historical session data each in N historical session data and product know-how map as input number
According to the session prediction model is input to, the corresponding prediction conversation recording of every historical session data is obtained；
It is recorded based on the corresponding prediction conversation recording of every historical session data and historical session, obtains loss function, base
Training is optimized to the session prediction model in gradient descent method and the loss function.
In some possible embodiments, the session prediction model after being trained using newest optimization is to described first time
Complex information is predicted, in terms of obtaining object statement, above procedure is specifically used for executing the instruction of following steps:
Based on the session prediction model after newest optimization training, first eigenvector is obtained, the first eigenvector is used
Corresponding product type is believed in characterizing described first and replying；
Based on the session prediction model after the newest optimization training, obtain second feature vector, the second feature to
Amount is for characterizing the corresponding product know-how map of first return information；
The first eigenvector and the second feature vector are spliced, target feature vector is obtained；
The target feature vector is matched, matched first template vector is obtained；
The product type of first template vector and product know-how map are formed into sentence, obtain object statement.
In some possible embodiments, above procedure is also used to execute the instruction of following steps:
Obtain the complete conversation recording with the target user；
Determine the target user to the satisfaction of any one product according to the complete conversation recording；
The purchase approach of target product is pushed to the target user, the target product is that satisfaction is greater than or waits
In the product of threshold value.
It is above-mentioned that mainly the scheme of the embodiment of the present application is described from the angle of method side implementation procedure.It is understood that
, in order to realize the above functions, it comprises execute the corresponding hardware configuration of each function and/or software module for set-top box.
Those skilled in the art should be readily appreciated that, in conjunction with each exemplary module and calculation of embodiment description presented herein
Method step, the application can be realized with the combining form of hardware or hardware and computer software.Some function is actually with hardware
Or computer software drives the mode of hardware to execute, the specific application and design constraint depending on technical solution.Specially
Industry technical staff can specifically realize described function to each using distinct methods, but this realization is not answered
Think beyond scope of the present application.
The embodiment of the present application can carry out the division of functional module according to above method example to set-top box, for example, can be with
Two or more functions, can also be integrated in a processing module by corresponding each each functional module of function division
In.Above-mentioned integrated module both can take the form of hardware realization, can also be realized in the form of software function module.It needs
It is noted that be schematical, only a kind of logical function partition to the division of module in the embodiment of the present application, it is practical real
It is current that there may be another division manner.
The session platform 500 of dialogue-based prediction model involved in above-described embodiment is shown refering to Fig. 5, Fig. 5
A kind of possible functional unit composition block diagram, session platform 500 includes: selecting unit 510, interactive unit 520, predicting unit
530 and transmission unit 540, in which:
Selecting unit 510, for the selection target user from customer data base；
Interactive unit 520 receives the target user for the meeting for initiating a session request to the target user
Talk about the first return information of request；
Predicting unit 530 obtains target language for predicting using session prediction model first return information
Sentence；
Transmission unit 540, for sending the object statement to the target user, to establish meeting with the target user
Talk about process.
In some possible embodiments, session platform 500 further includes training unit 550, the session prediction model
It is constructed based on machine learning Integrated Algorithm, training unit 550 is used for:
Before selection target user in customer data base, mould is predicted to the session using N historical session data
Type optimizes training, and the N is positive integer；First return information is being predicted using session prediction model, is being obtained
To in terms of object statement, predicting unit 530 is specifically used for: using the session prediction model after newest optimization training to described the
One return information is predicted, object statement is obtained.
In some possible embodiments, every historical session data include product in the N historical session data
Demand, the product know-how map built in advance and historical session record, the product know-how map include the attribute letter of product
Breath and non-attribute information；In terms of optimizing training to the session prediction model using N historical session data, training is single
Member 550, is specifically used for:
Using the product demand of historical session data each in N historical session data and product know-how map as input number
According to the session prediction model is input to, the corresponding prediction conversation recording of every historical session data is obtained；
It is recorded based on the corresponding prediction conversation recording of every historical session data and historical session, obtains loss function, base
Training is optimized to the session prediction model in gradient descent method and the loss function.
In some possible embodiments, the session prediction model after being trained using newest optimization is to described first time
Complex information is predicted, in terms of obtaining object statement, predicting unit 530 is specifically used for:
Based on the session prediction model after newest optimization training, first eigenvector is obtained, the first eigenvector is used
Corresponding product type is believed in characterizing described first and replying；
Based on the session prediction model after the newest optimization training, obtain second feature vector, the second feature to
Amount is for characterizing the corresponding product know-how map of first return information；
The first eigenvector and the second feature vector are spliced, target feature vector is obtained；
The target feature vector is matched, matched first template vector is obtained；
The product type of first template vector and product know-how map are formed into sentence, obtain object statement.
In some possible embodiments, transmission unit 540 are also used to:
Obtain the complete conversation recording with the target user；
Determine the target user to the satisfaction of any one product according to the complete conversation recording；
The purchase approach of target product is pushed to the target user, the target product is that satisfaction is greater than or waits
In the product of threshold value.
The embodiment of the present application also provides a kind of computer storage medium, and the computer-readable recording medium storage has calculating
Machine program, the computer program are executed by processor to realize that any one recorded in above method embodiment such as is based on meeting
Talk about some or all of the session method of prediction model step.
The embodiment of the present application also provides a kind of computer program product, and the computer program product includes storing calculating
The non-transient computer readable storage medium of machine program, the computer program are operable to that computer is made to execute such as above-mentioned side
Some or all of the session method for the dialogue-based prediction model of any one recorded in method embodiment step.
It should be noted that for the various method embodiments described above, for simple description, therefore, it is stated as a series of
Combination of actions, but those skilled in the art should understand that, the application is not limited by the described action sequence because
According to the application, some steps may be performed in other sequences or simultaneously.Secondly, those skilled in the art should also know
It knows, embodiment described in this description belongs to alternative embodiment, related actions and modules not necessarily the application
It is necessary.
In the above-described embodiments, it all emphasizes particularly on different fields to the description of each embodiment, there is no the portion being described in detail in some embodiment
Point, reference can be made to the related descriptions of other embodiments.
In several embodiments provided herein, it should be understood that disclosed device, it can be by another way
It realizes.For example, the apparatus embodiments described above are merely exemplary, such as the division of the unit, it is only a kind of
Logical function partition, there may be another division manner in actual implementation, such as multiple units or components can combine or can
To be integrated into another system, or some features can be ignored or not executed.Another point, shown or discussed is mutual
Coupling, direct-coupling or communication connection can be through some interfaces, the indirect coupling or communication connection of device or unit,
It can be electrical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, each functional unit in each embodiment of the application can integrate in one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also be realized in the form of software program module.
If the integrated unit is realized in the form of software program module and sells or use as independent product
When, it can store in a computer-readable access to memory.Based on this understanding, the technical solution of the application substantially or
Person says that all or part of the part that contributes to existing technology or the technical solution can body in the form of software products
Reveal and, which is stored in a memory, including some instructions are used so that a computer equipment
(can be personal computer, server or network equipment etc.) executes all or part of each embodiment the method for the application
Step.And memory above-mentioned includes: USB flash disk, read-only memory (ROM, Read-Only Memory), random access memory
The various media that can store program code such as (RAM, Random Access Memory), mobile hard disk, magnetic or disk.
Those of ordinary skill in the art will appreciate that all or part of the steps in the various methods of above-described embodiment is can
It is completed with instructing relevant hardware by program, which can store in a computer-readable memory, memory
May include: flash disk, read-only memory (English: Read-Only Memory, referred to as: ROM), random access device (English:
Random Access Memory, referred to as: RAM), disk or CD etc..
The embodiment of the present application is described in detail above, specific case used herein to the principle of the application and
Embodiment is expounded, the description of the example is only used to help understand the method for the present application and its core ideas；
At the same time, for those skilled in the art can in specific embodiments and applications according to the thought of the application
There is change place, in conclusion the contents of this specification should not be construed as limiting the present application.
Claims (10)
1. a kind of session method of dialogue-based prediction model characterized by comprising
The selection target user from customer data base；
It initiates a session request to the target user, receives the target user and replied for the first of the session request and believed
Breath；
First return information is predicted using session prediction model, obtains object statement；
The object statement is sent to the target user, to establish session process with the target user.
2. the method according to claim 1, wherein the session prediction model is based on machine learning Integrated Algorithm
Building, the method also includes: before selection target user in customer data base, using N historical session data to institute
It states session prediction model and optimizes training, the N is positive integer；
It is described that first return information is predicted using session prediction model, obtain object statement, comprising: using newest
Session prediction model after optimization training predicts first return information, obtains object statement.
3. according to the method described in claim 2, it is characterized in that, every historical session number in the N historical session data
According to product know-how map and the historical session record built including product demand, in advance, the product know-how map includes producing
The attribute information of product and non-attribute information；
It is described that training is optimized to the session prediction model using N historical session data, comprising:
The product demand of historical session data each in N historical session data and product know-how map is defeated as input data
Enter to the session prediction model, obtains the corresponding prediction conversation recording of every historical session data；
It is recorded based on the corresponding prediction conversation recording of every historical session data and historical session, obtains loss function, based on ladder
Degree descent method and the loss function optimize training to the session prediction model.
4. according to the method in claim 2 or 3, which is characterized in that the session using after newest optimization training is predicted
Model predicts first return information, obtains object statement, comprising:
Based on the session prediction model after newest optimization training, first eigenvector is obtained, the first eigenvector is used for table
It levies described first and replys and believe corresponding product type；
Based on the session prediction model after the newest optimization training, second feature vector is obtained, the second feature vector is used
In the corresponding product know-how map of characterization first return information；
The first eigenvector and the second feature vector are spliced, target feature vector is obtained；
The target feature vector is matched, matched first template vector is obtained；
The product type of first template vector and product know-how map are formed into sentence, obtain object statement.
5. method according to claim 1-4, which is characterized in that the method also includes:
Obtain the complete conversation recording with the target user；
Determine the target user to the satisfaction of any one product according to the complete conversation recording；
The purchase approach of target product is pushed to the target user, the target product is that satisfaction is more than or equal to threshold
The product of value.
6. a kind of session platform of dialogue-based prediction model characterized by comprising
Selecting unit, for the selection target user from customer data base；
Interactive unit receives the target user for the session request for initiating a session request to the target user
The first return information；
Predicting unit obtains object statement for predicting using session prediction model first return information；
Transmission unit, for sending the object statement to the target user, to establish session process with the target user.
7. session platform according to claim 6, which is characterized in that the session prediction model is integrated based on machine learning
Algorithm building, the session platform further include: training unit；
Before selection target user in customer data base, the training unit is used for using N historical session data to institute
It states session prediction model and optimizes training, the N is positive integer；
First return information is being predicted using session prediction model, in terms of obtaining object statement, the prediction is single
Member is specifically used for: being predicted using the session prediction model after newest optimization training first return information, obtains mesh
Poster sentence.
8. session platform according to claim 7, which is characterized in that every history meeting in the N historical session data
Words data include product demand, the product know-how map built in advance and historical session record, the product know-how map packet
Include the attribute information and non-attribute information of product；
In terms of optimizing training to the session prediction model using N historical session data, the training unit, specifically
For:
The product demand of historical session data each in N historical session data and product know-how map is defeated as input data
Enter to the session prediction model, obtains the corresponding prediction conversation recording of every historical session data；
It is recorded based on the corresponding prediction conversation recording of every historical session data and historical session, obtains loss function, based on ladder
Degree descent method and the loss function optimize training to the session prediction model.
9. a kind of session platform, which is characterized in that including processor, memory, communication interface and one or more program,
In, one or more of programs are stored in the memory, and are configured to be executed by the processor, described program
Include the steps that requiring the instruction in any one of 1-5 method for perform claim.
10. a kind of computer readable storage medium, which is characterized in that the computer-readable recording medium storage has computer journey
Sequence, the computer program are executed by processor to realize the method according to claim 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910605730.7A CN110489519B (en) | 2019-07-05 | 2019-07-05 | Session method based on session prediction model and related products |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910605730.7A CN110489519B (en) | 2019-07-05 | 2019-07-05 | Session method based on session prediction model and related products |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110489519A true CN110489519A (en) | 2019-11-22 |
| CN110489519B CN110489519B (en) | 2023-07-07 |
Family
ID=68546004
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910605730.7A Active CN110489519B (en) | 2019-07-05 | 2019-07-05 | Session method based on session prediction model and related products |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110489519B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111145754A (en) * | 2019-12-12 | 2020-05-12 | 深圳追一科技有限公司 | Voice input method, device, terminal equipment and storage medium |
| CN112735407A (en) * | 2020-12-24 | 2021-04-30 | 北京三快在线科技有限公司 | Conversation processing method and device |
| CN112905763A (en) * | 2021-02-03 | 2021-06-04 | 深圳市人马互动科技有限公司 | Session system development method, device, computer equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107767881A (en) * | 2016-08-15 | 2018-03-06 | ***通信有限公司研究院 | A kind of acquisition methods and device of the satisfaction of voice messaging |
| CA2988282A1 (en) * | 2016-12-22 | 2018-06-22 | Capital One Services, Llc | Systems and methods for customer sentiment prediction and depiction |
| US20180181873A1 (en) * | 2016-12-28 | 2018-06-28 | Facebook, Inc. | Personalized post session model for an online system |
| CN109493166A (en) * | 2018-10-23 | 2019-03-19 | 深圳智能思创科技有限公司 | A kind of construction method for e-commerce shopping guide's scene Task conversational system |
| CN109857848A (en) * | 2019-01-18 | 2019-06-07 | 深圳壹账通智能科技有限公司 | Interaction content generation method, device, computer equipment and storage medium |
-
2019
- 2019-07-05 CN CN201910605730.7A patent/CN110489519B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107767881A (en) * | 2016-08-15 | 2018-03-06 | ***通信有限公司研究院 | A kind of acquisition methods and device of the satisfaction of voice messaging |
| CA2988282A1 (en) * | 2016-12-22 | 2018-06-22 | Capital One Services, Llc | Systems and methods for customer sentiment prediction and depiction |
| US20180181873A1 (en) * | 2016-12-28 | 2018-06-28 | Facebook, Inc. | Personalized post session model for an online system |
| CN109493166A (en) * | 2018-10-23 | 2019-03-19 | 深圳智能思创科技有限公司 | A kind of construction method for e-commerce shopping guide's scene Task conversational system |
| CN109857848A (en) * | 2019-01-18 | 2019-06-07 | 深圳壹账通智能科技有限公司 | Interaction content generation method, device, computer equipment and storage medium |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111145754A (en) * | 2019-12-12 | 2020-05-12 | 深圳追一科技有限公司 | Voice input method, device, terminal equipment and storage medium |
| CN111145754B (en) * | 2019-12-12 | 2021-04-13 | 深圳追一科技有限公司 | Voice input method, device, terminal equipment and storage medium |
| CN112735407A (en) * | 2020-12-24 | 2021-04-30 | 北京三快在线科技有限公司 | Conversation processing method and device |
| CN112735407B (en) * | 2020-12-24 | 2023-07-21 | 北京三快在线科技有限公司 | Dialogue processing method and device |
| CN112905763A (en) * | 2021-02-03 | 2021-06-04 | 深圳市人马互动科技有限公司 | Session system development method, device, computer equipment and storage medium |
| CN112905763B (en) * | 2021-02-03 | 2023-10-24 | 深圳市人马互动科技有限公司 | Session system development method, device, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110489519B (en) | 2023-07-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9437215B2 (en) | Predictive video analytics system and methods | |
| CN110347863B (en) | Speaking recommendation method and device and storage medium | |
| CN110489519A (en) | The session method and Related product of dialogue-based prediction model | |
| CN109002515A (en) | A kind of method and apparatus of intelligent response | |
| CN110580516B (en) | Interaction method and device based on intelligent robot | |
| CN110265013A (en) | The recognition methods of voice and device, computer equipment, storage medium | |
| CN109922213A (en) | Data processing method, device, storage medium and terminal device when voice is seeked advice from | |
| CN109285011A (en) | A kind of smart client return visit method, device and equipment | |
| CN113840040B (en) | Man-machine cooperation outbound method, device, equipment and storage medium | |
| CN109739605A (en) | The method and apparatus for generating information | |
| CN111291151A (en) | Interaction method and device and computer equipment | |
| CN109215679A (en) | Dialogue method and device based on user emotion | |
| CN109271503A (en) | Intelligent answer method, apparatus, equipment and storage medium | |
| CN106356056B (en) | Audio recognition method and device | |
| CN113486166B (en) | Construction method, device and equipment of intelligent customer service robot and storage medium | |
| CN109474734A (en) | Recognition methods, device and the terminal of type of call | |
| CN110175240A (en) | Construction method, device and the storage medium of knowledge mapping relevant to outgoing call process | |
| CN108347531A (en) | Help seeking method, help seeking device, electronic equipment and computer readable storage medium | |
| CN113051405B (en) | Intelligent outbound knowledge graph construction method and device based on dialogue scene | |
| CN112331201A (en) | Voice interaction method and device, storage medium and electronic device | |
| CN115083412B (en) | Voice interaction method and related device, electronic equipment and storage medium | |
| CN111047362A (en) | Statistical management method and system for use activity of intelligent sound box | |
| CN110047473A (en) | A kind of man-machine collaboration exchange method and system | |
| CN108647139A (en) | Test method, device, storage medium and the electronic device of system | |
| CN117493582B (en) | Model result output method and device, electronic equipment and storage medium |
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 |