CN109600523A - Service hotline broadcasting method, device, computer equipment and storage medium - Google Patents
Service hotline broadcasting method, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN109600523A CN109600523A CN201811185764.7A CN201811185764A CN109600523A CN 109600523 A CN109600523 A CN 109600523A CN 201811185764 A CN201811185764 A CN 201811185764A CN 109600523 A CN109600523 A CN 109600523A
- Authority
- CN
- China
- Prior art keywords
- user
- information
- inlet wire
- business
- service
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M3/00—Automatic or semi-automatic exchanges
- H04M3/42—Systems providing special services or facilities to subscribers
- H04M3/50—Centralised arrangements for answering calls; Centralised arrangements for recording messages for absent or busy subscribers ; Centralised arrangements for recording messages
- H04M3/51—Centralised call answering arrangements requiring operator intervention, e.g. call or contact centers for telemarketing
- H04M3/5183—Call or contact centers with computer-telephony arrangements
- H04M3/5191—Call or contact centers with computer-telephony arrangements interacting with the Internet
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M3/00—Automatic or semi-automatic exchanges
- H04M3/42—Systems providing special services or facilities to subscribers
- H04M3/487—Arrangements for providing information services, e.g. recorded voice services or time announcements
- H04M3/493—Interactive information services, e.g. directory enquiries ; Arrangements therefor, e.g. interactive voice response [IVR] systems or voice portals
- H04M3/4936—Speech interaction details
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M3/00—Automatic or semi-automatic exchanges
- H04M3/42—Systems providing special services or facilities to subscribers
- H04M3/50—Centralised arrangements for answering calls; Centralised arrangements for recording messages for absent or busy subscribers ; Centralised arrangements for recording messages
- H04M3/51—Centralised call answering arrangements requiring operator intervention, e.g. call or contact centers for telemarketing
- H04M3/5166—Centralised call answering arrangements requiring operator intervention, e.g. call or contact centers for telemarketing in combination with interactive voice response systems or voice portals, e.g. as front-ends
Abstract
A kind of service hotline broadcasting method, device, computer equipment and storage medium based on prediction model proposed by the present invention, wherein method includes: to obtain the first user information by the inlet wire number that phone enters service hotline according to user；First user information is inputted in level-one prediction model and is calculated, the first probability value that user is intended to each single item business in the class of service handled is obtained, the level-one prediction model is based on gradient promotion decision tree and logistic regression algorithm is trained to obtain to specified sample set；The each single item business of correspondence first probability value is successively broadcasted according to the descending of first probability value, the business demand of subscriber phone inlet wire can be effectively predicted by the above method, the business for allowing user that can hear that oneself needs to handle, the user that can effectively shorten again finds the duration of services menu, improves user experience.
Description
Technical field
The present invention relates to prediction model technical field, a kind of service hotline broadcasting method, device, meter are especially related to
Calculate machine equipment and storage medium.
Background technique
Most of company is in order to facilitate client, or in order to provide better service, as communication common carrier, insurance company are usual
Service hotline system can be equipped in company, client can be by dialing service hotline to obtain respective service；Existing service heat
Line is usually to set fixed menu to be broadcasted, and allows user to carry out selection business in a menu, then jumps to next subdivision
Business, or by acquiring the product bought of user, and then recommend the series menu of this corresponding product, still, in this way
Can have that user hears is not oneself desired menu, or hears too many menu, so that user is difficult in a short time
The business for oneself wanting to handle is found, the phenomenon that user experience is bad is caused.
Summary of the invention
The main object of the present invention is to provide service hotline broadcasting method, the device, calculating of a kind of raising user experience
Machine equipment and storage medium.
The present invention proposes a kind of service hotline broadcasting method, comprising: according to user by phone enter service hotline into
Wire size code obtains the first user information, and first user information includes the product letter that the attribute information of user, user bought
Breath and user pass through the historical behavior trace information after phone inlet wire；
First user information is inputted in level-one prediction model and is calculated, user is obtained and is intended to the service class handled
First probability value of not middle each single item business, the level-one prediction model are based on gradient and promote decision tree and logistic regression algorithm
Specified sample set is trained to obtain；
The each single item business of correspondence first probability value is successively broadcasted according to the descending of first probability value.
Further, described that the first user information is obtained by the inlet wire number that phone enters service hotline according to user
Step, comprising:
Obtain the inlet wire number that user enters service hotline by phone；
According to first user information of the correspondence inlet wire number in the inlet wire number recalls information library.
Further, the descending according to first probability value is by each single item business of correspondence first probability value
After the step of successively being broadcasted, comprising:
Obtain the business of user's input；
Classification according to the business calls corresponding second level prediction model；
First user information is inputted in the second level prediction model and is calculated, each single item subordinate of the business is obtained
Second probability value of business；
The each single item subordinate business of correspondence second probability value is successively carried out according to the descending of second probability value
Casting.
Further, the descending according to first probability value is by each single item business of correspondence first probability value
After the step of successively being broadcasted, comprising:
The corresponding inlet wire information of the inlet wire number is recorded, the inlet wire information is to obtain according to input information after subscriber's drop
The subscriber information message arrived；
The inlet wire information is added to first user information to form second user information,
By the second user information training level-one prediction model to form new level-one prediction model, and will be described
Second user information is stored in described information storehouse.
Further, after the step of inlet wire number for obtaining user, comprising:
Judge whether the inlet wire number matches with the Subscriber Number that described information storehouse stores；
If it is not, then according in preset period inlet wire history information call hot spot service, the hot spot service be
In the business being ranked up according to the descending for handling number, the business in preset forward ranking；
The hot spot service is broadcasted.
Further, before the step of inlet wire history information according in preset period calls hot spot service,
Include:
Obtain the inlet wire history information in preset period；
The number that user handles each single item business is analyzed according to the inlet wire history information；
The business is successively sorted by the descending of number；
By the service marker preset in the sequence in ranking at the hot spot service.
Further, after described the step of being broadcasted the hot spot service, comprising:
Record the first inlet wire information of the inlet wire number, the first inlet wire information be according to user's first time inlet wire after
The subscriber information message that input information obtains；
Using the first inlet wire information as level-one prediction model described in sample training, obtain the inlet wire number for the first time into
The first inlet wire information is stored in described information storehouse by the new level-one prediction model after line.
The present invention also provides a kind of service hotline broadcast devices, comprising:
Acquiring unit, for obtaining the first user information by the inlet wire number that phone enters service hotline according to user,
First user information includes that the product information that the attribute information of user, user bought and user pass through after phone inlet wire
Historical behavior trace information；
Computing unit is calculated for inputting first user information in level-one prediction model, obtains user's meaning
First probability value of each single item business in the class of service that figure is handled, the level-one prediction model be based on gradient promoted decision tree with
And logistic regression algorithm is trained to obtain to specified sample set；
Unit is broadcasted, for the descending according to first probability value by each single item business of correspondence first probability value
Successively broadcasted.
The present invention also provides a kind of computer equipment, including memory and processor, the memory is stored with computer
The step of program, the processor realizes the above method when executing the computer program.
The present invention also provides a kind of computer readable storage mediums, are stored thereon with computer program, the computer
The step of above method is realized when program is executed by processor.
The invention has the benefit that the business demand of subscriber phone inlet wire can be effectively predicted by prediction model, allow
User can hear the duration for oneself needing the business handled and the user that effectively shortens to find services menu, improve user's body
It tests.
Detailed description of the invention
Fig. 1 is the step schematic diagram of service hotline broadcasting method in one embodiment of the invention；
Fig. 2 is the structural schematic block diagram of service hotline broadcast device in one embodiment of the invention；
Fig. 3 is the structural schematic block diagram of the computer equipment of one embodiment of the invention.
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 embodiment
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, the service hotline broadcasting method in the present embodiment, comprising:
Step S1: entering inlet wire number the first user information of acquisition of service hotline according to user by phone, and described the
One user information includes that the product information that the attribute information of user, user bought and user pass through the history after phone inlet wire
Action trail information；
Step S2: first user information being inputted in level-one prediction model and is calculated, and is obtained user and is intended to handle
Class of service in each single item business the first probability value, the level-one prediction model is based on gradient and promotes decision tree and logic
Regression algorithm is trained to obtain to specified sample set；
Step S3: according to first probability value descending by each single item business of correspondence first probability value successively into
Row casting.
In step sl, it before predicting service hotline, needs to obtain the information of prediction object, that is, needs to obtain
Above-mentioned first user information, first user information can be obtained according to user by the inlet wire number that phone enters service hotline,
Above-mentioned first user information includes but is not limited to: the attribute information of user, user bought product information and user passes through electricity
Historical behavior trace information after talking about inlet wire.Wherein the attribute information of user includes but is not limited to: name, gender, the age,
The purchase product information of location, occupation etc., user includes but is not limited to: the business once handled and the product bought, such as
Bought vehicle insurance, life insurance, handled credit card, bank card etc., historical behavior trace information includes but is not limited to: user passes through
Phone inlet wire and carried out action trail, such as activated credit card by phone, seeked advice from vehicle insurance purchase, inquire bank card
Remaining sum etc..
In one embodiment, above-mentioned steps S1, comprising:
Step S10: the inlet wire number that user enters service hotline by phone is obtained；
Step S11: according to first user of the correspondence inlet wire number in the inlet wire number recalls information library
Information.
When user wants to handle a certain business of some company, in order to save time energy, it is desirable to directly pass through phone
Hot line is completed, then user can make a phone call to handle into the service hotline of the said firm, the service hotline provided in the present embodiment
Broadcasting method can realize that, when user makes a phone call inlet wire, system can obtain the use by above-mentioned telephone service hot line system
The inlet wire number at family, in general, the user handled business in the said firm before inlet wire, and can stay when transacting business
Lower subscriber data, the i.e. data of user have been stored in information bank, it is to be understood that user may beat more than once
Phone inlet wire, so the first user information being stored in information bank includes the purchase of the attribute information of above-mentioned user, user
Product information and the historical behavior trace information of user etc..
In step s 11, all clients of the company using the telephone service hot line system are stored in above- mentioned information library
Information, these customer informations include the first above-mentioned user information, and when subscriber phone inlet wire, system can be according to the inlet wire number
The first user information that above-mentioned inlet wire number is corresponded in code recalls information library will specifically deposit in inlet wire number and information bank
The number of storage compares, and finds out the number with inlet wire numbers match, and each number wherein stored in information bank is corresponding
The first user information of one user is find corresponding user when finding the number with inlet wire numbers match
At this moment one user information calls the first user information of the user.
Above-mentioned steps S2, after the first user information for recalling corresponding above-mentioned inlet wire number in information bank, by these users
It is calculated in information input level-one prediction model, to obtain user is intended to each single item business in the class of service handled the
One probability value, above-mentioned business is that company is supplied to the business that client handles, including a variety of different classs of service, such as insurance company
Class of service can be provided for client has: producing danger, life insurance, bank, credit card, security etc., system can be to above-mentioned each single item industry
Business is broadcasted, so that user selects.
Wherein, above-mentioned level-one prediction model is by promoting decision tree GBDT (Gradient Boost based on gradient
Decision Tree) specified sample is obtained into collection row training with logistic regression LR algorithm, wherein GBDT is used to do feature group
It closes, LR is used to do the training of level-one prediction model, and above-mentioned level-one prediction model is that the training of both algorithm fusions obtains.Specifically
It says, it is a kind of integrated that above-mentioned GBDT, which can be used to do classify, return, such as can be used to predict the weather, predict buying intention
Learning method is set by grey iterative generation more, and each iteration all creates a decision tree in the gradient direction for reducing residual error, is owned
The conclusion of tree adds up as final result.In order to improve level-one prediction model stability, over-fitting, above-mentioned specified sample are prevented
Collection includes appropriate sample, such as tens of thousands of or hundreds of thousands sample, each sample for a user user information, in user information
Customer attribute information, the purchase product information of user and user's history action trail information can form thousands of Wesy Hu Te
Sign, such as " last time inlet wire interval number of days ", " last time inlet wire speech intention ", " are intended to " above-mentioned inlet wire interval times " for nearest one month
Score ", " whether being to produce dangerous user ", " whether wanting life insurance user ", " whether wanting credit card user " etc., in above-mentioned decision tree, often
The output of a leaf node is 0 or 1, i.e., the value of each user characteristics is 0 or 1, in training pattern, adjustable decision tree
Depth and number, then establish more decision trees, then the combination of each leaf child node of final output takes said combination
Value is trained using logistic regression algorithm, and above-mentioned logistic regression algorithm first can indicate independent variable x and dependent variable y with y=f (x)
Relationship, in the present embodiment, independent variable x is above-mentioned user characteristics, and dependent variable y is prediction result, i.e. user is intended to the industry handled
Business, finally by the logistic regression equation of the probability value of prediction result y, so that calculating user is intended to each single item business handled
The first probability value.
The first probability that user is intended to each single item business handled is calculated by level-one prediction model in above-mentioned steps S3
After value, according to these the first probability values, by the business of above-mentioned correspondence these the first probability values by the first probability value descending into
Row sequence, i.e., by business according to the first probability value from being ranked up to small sortord greatly, such first probability value is big
Business preferentially comes front, and then carrying out voice broadcast to user makes its accelerating selection to save the time of user, is promoted
Experience effect.
In a specific embodiment, such as 95511 service hotlines of safety group, when user beats the telephone number inlet wire,
System can get the inlet wire number of the user, and the corresponding inlet wire number is then found out in information bank according to the inlet wire number
User information recalls the user information when finding corresponding user information, is input to level-one prediction model and is counted
It calculates, obtains user and be intended to the first probability value of each single item business handled, if the first probability value of life insurance is 0.3, the of security
One probability value is 0.6, the first probability value of credit card is 0.2, is at this moment sorted from large to small by probability value: security, life insurance, credit
Card, then above-mentioned business successively can be carried out voice broadcast to user by system, and user is allowed to select.
In one embodiment, after above-mentioned steps S3, comprising:
Step S4: the business of user's input is obtained；
Step S5: the classification according to the business calls corresponding second level prediction model；
Step S6: first user information is inputted in the second level prediction model and is calculated, the every of the business is obtained
Second probability value of one subordinate's option；
Step S7: according to second probability value descending by each single item subordinate business of correspondence second probability value according to
It is secondary to be broadcasted.
In the present embodiment, since the company using above-mentioned telephone service hot line system may have multiple classs of service to produce
Product, such as insurance company have and produce danger, life insurance, bank, credit card, security class of service, and each business is handled with multiple
Subordinate's business, such as producing subordinate's business of handling of dangerous business includes to produce danger to report a case to the security authorities, produce danger purchase, produce dangerous setting loss, credit card
Subordinate's business of handling of business includes that creditable card sets password for inquiry, credit card activates, card of making a credit inquiry can use amount etc.
Deng.So user can input the business for oneself being intended to handle after in step s 4, carrying out voice broadcast to above-mentioned business,
At this moment system can obtain the business of user's input,, can be according to the business after the business for obtaining user's input in step S5
Classification calls corresponding second level prediction model, then inputs in second level prediction model user information and calculates, obtains above-mentioned industry
Second probability value of each single item subordinate's business of business, wherein the business of each classification respectively corresponds a second level prediction model,
Subordinate's business in the present embodiment is the business belonged in the above-mentioned scope of business, at this moment according to the second probability value by subordinate's business
Be ranked up according to according to the descending of the second probability value, i.e., by subordinate's business according to corresponding second probability value from large to small successively
Sequence, then successively broadcasts ordering subordinate's business.
Wherein above-mentioned second level prediction model is to train for single business, for training the data of the second level prediction model
For in user information user property and user handled the historical behavior track of the business, using algorithm and training process
It is consistent with above-mentioned level-one prediction model, it is similarly gradient and promotes decision tree GBDT and logistic regression algorithm LR, it is no longer superfluous here
It states.
Certainly, since subordinate's classification of business is uncertain, subordinate only a point level-one may be also possible to point a multiple grades, such as second level
Subordinate divides three-level again later, or divides level Four again, corresponding, can preset three-level prediction model, level Four prediction model, this programme is not
The series of subordinate's classification is limited, every level-one step can carry out prediction user's intention according to the method described above, to improve user's body
It tests.
In another embodiment, after above-mentioned steps S3, comprising:
Step S4 ': recording the inlet wire information of the corresponding user of the inlet wire number, the inlet wire information be according to user into
The subscriber information message that information obtains is inputted after line；
Step S5 ': the inlet wire information is added to first user information to form second user information；
Step S6 ': by the second user information training level-one prediction model to form new level-one prediction model,
And the second user information is stored in described information storehouse.
In the present embodiment, after step s 3, i.e., after above-mentioned business is carried out voice broadcast by system, user can be selected
Oneself is intended to the business handled, then the user will generate new historical behavior track, and lead to after the current inlet wire that finishes
The intention of user can be effectively obtained by crossing these historical behavior tracks, these all can serve as next user again inlet wire and carry out
The basis of prediction, so system records the inlet wire information of the corresponding user of inlet wire number, above-mentioned inlet wire information in step S4 '
The as historical behavior track of the current inlet wire of user, such as the secondary subscriber's drop handle production danger and report a case to the security authorities or activate to credit card
Deng.In step S5 ', after system gets above-mentioned carry out information by record, before these progress information can be added to
The first user information in, to form new second user information, and be stored in information bank so as to next inlet wire calling, together
When these second user information can be trained level-one prediction model, so that new level-one prediction model is formed, so that level-one
Prediction model is in newest casting state at any time.
In one embodiment, after above-mentioned steps S10, comprising:
Step S12: judge whether the inlet wire number matches with the Subscriber Number that described information storehouse stores；
Step S13: if it is not, then calling hot spot service, the hot spot according to the inlet wire history information in preset period
Business is the business in the business being ranked up according to the descending for handling number, in preset forward ranking；
Step S14: the hot spot series traffic item is broadcasted.
In the present embodiment, since the user of inlet wire number might not handle above-mentioned business or inlet wire was seeked advice from,
So not leaving historical behavior trace information, for such user, current hot spot service can be recommended, that is, compared more
The business that people handles or seeks advice from, above-mentioned hot spot service be in the business being ranked up according to the descending for handling number, it is preset
Business in forward ranking.It for above-mentioned such user, is not stored for information about in information bank, and training level-one prediction
Also without the information of the user in the user information of model, so after step slo, first judging inlet wire number and information inventory
Whether the Subscriber Number of storage matches, at this moment, can be with if it does not match, illustrate that the user of the inlet wire number is first time inlet wire
Hot spot service is called according to the inlet wire history information in preset period, in the present embodiment, above-mentioned preset period can be made by oneself
Justice setting, such as 15 days or one month, citing ground, if it is determined that inlet wire number is first time inlet wire, it at this moment can be in the database
The hot spot service in one month is called, then broadcasts these hot spot services.
Specifically, before the step of above-mentioned inlet wire history information according in preset period calls hot spot service,
Include:
Step S01: the inlet wire history information in preset period is obtained；
Step S02: the number that user handles each single item business is analyzed according to the inlet wire history information；
Step S03: the business is successively sorted by the descending of number；
Step S04: by the service marker preset in the sequence in ranking at the hot spot service.
It is understood that all inlet wire behaviors can all have record, these record storages are in database, in these notes
It can check which business inlet wire user once handled in record, these inlet wire history informations be obtained from database, so
The number for obtaining handling each business therefrom being analyzed afterwards, wherein business can successively be sorted by the descending of number, it then will be at this
In sequence, preset ranking in service marker be hot spot service, such as by successively sort by the descending of number the first eight in industry
Business is set as hot spot service, if what is obtained is the inlet wire historical record in preset period, what is obtained is in the preset period
Hot spot service.
Further, after above-mentioned S14, comprising:
Step S15: recording the first inlet wire information of the inlet wire number, and the first inlet wire information is according to user first
The subscriber information message that information obtains is inputted after secondary inlet wire；
Step S16: using the first inlet wire information as level-one prediction model described in sample training, the inlet wire number is obtained
New level-one prediction model after the first inlet wire of code, is stored in described information storehouse for the first inlet wire information.
After being broadcasted above-mentioned hot spot service, the first inlet wire information of the secondary subscriber's drop is recorded, i.e.,
The historical behavior information that the secondary inlet wire number corresponds to user is recorded, is then instructed these first inlet wire information as sample
Practice above-mentioned level-one prediction model, second level prediction model etc., to obtain the new level-one of the corresponding first inlet wire of inlet wire number
Prediction model, new second level prediction model；It is above-mentioned new level-one prediction model, new when user's inlet wire again of the secondary inlet wire
Second level prediction model can be used to predict that the user is intended to, meanwhile, above-mentioned first inlet wire information is stored in above- mentioned information library,
Facilitate calling.
Referring to Fig. 2, service hotline broadcast device in the present embodiment, comprising:
Information unit 100 is obtained, is used for obtaining first by the inlet wire number that phone enters service hotline according to user
Family information, first user information include that the product information that the attribute information of user, user bought and user pass through electricity
Historical behavior trace information after talking about inlet wire；
First computing unit 200 is calculated for inputting first user information in level-one prediction model, is obtained
User is intended to the first probability value of each single item business in the class of service handled, and the level-one prediction model is based on gradient promotion and determines
Plan tree and logistic regression algorithm are trained to obtain to specified sample set；
First casting unit 300, for the descending according to first probability value by the every of correspondence first probability value
One business is successively broadcasted.
In obtaining information unit 100, before predicting service hotline, need to obtain the information of prediction object,
Need to obtain above-mentioned first user information, which can enter the inlet wire of service hotline according to user by phone
Number obtains, and above-mentioned first user information includes but is not limited to: the attribute information of user, the purchase product information of user and use
The historical behavior trace information at family.Wherein the attribute information of user includes but is not limited to: name, gender, age, address, occupation
Deng the purchase product information of user includes but is not limited to: the business once handled and the product bought, vehicle was such as bought
Danger, life insurance handled credit card, bank card etc., and historical behavior trace information includes but is not limited to: user passes through phone inlet wire
And action trail was carried out, such as carry out activating credit card by phone, seeked advice from vehicle insurance purchase, inquiry bank card remaining sum
Deng.
In one embodiment, above-mentioned acquisition information unit 100, comprising:
Module is obtained, the inlet wire number of service hotline is entered for obtaining user by phone；
Calling module, for according to described first of the correspondence inlet wire number in the inlet wire number recalls information library
User information.
When user wants to handle a certain business of some company, in order to save time energy, it is desirable to directly pass through phone
Hot line is completed, then user can make a phone call to handle into the service hotline of the said firm, the service hotline provided in the present embodiment
Broadcasting method can realize that, when user makes a phone call inlet wire, system can obtain the use by above-mentioned telephone service hot line system
The inlet wire number at family, in general, the user handled business in the said firm before inlet wire, and can stay when transacting business
Lower subscriber data, the i.e. data of user have been stored in information bank, it is to be understood that user may beat more than once
Phone inlet wire, so the first user information being stored in information bank includes the purchase of the attribute information of above-mentioned user, user
Product information and the historical behavior trace information of user etc..
In calling module, all clients of the company using the telephone service hot line system are stored in above- mentioned information library
Information, these customer informations include the first above-mentioned user information, and when subscriber phone inlet wire, system can be according to the inlet wire number
The first user information that above-mentioned inlet wire number is corresponded in code recalls information library will specifically deposit in inlet wire number and information bank
The number of storage compares, and finds out the number with inlet wire numbers match, and each number wherein stored in information bank is corresponding
The first user information of one user is find corresponding user when finding the number with inlet wire numbers match
At this moment one user information calls the first user information of the user.
As described in above-mentioned calculating probability unit 200, the first user letter of corresponding above-mentioned inlet wire number is recalled from information bank
After breath, these user informations are inputted in level-one prediction model and are calculated, so that obtaining user is intended to the class of service handled
First probability value of middle each single item business, above-mentioned business is that company is supplied to the business that client handles, including a variety of different industry
Business classification, such as insurance company can provide class of service for client and have: produce danger, life insurance, bank, credit card, security, system can
To be broadcasted to above-mentioned each single item business, so that user selects.
Wherein, above-mentioned level-one prediction model is by promoting decision tree GBDT (Gradient Boost based on gradient
Decision Tree) specified sample is obtained into collection row training with logistic regression LR algorithm, wherein GBDT is used to do feature group
It closes, LR is used to do the training of level-one prediction model, and above-mentioned level-one prediction model is that the training of both algorithm fusions obtains.Specifically
It says, it is a kind of integrated that above-mentioned GBDT, which can be used to do classify, return, such as can be used to predict the weather, predict buying intention
Learning method is set by grey iterative generation more, and each iteration all creates a decision tree in the gradient direction for reducing residual error, is owned
The conclusion of tree adds up as final result.In order to improve level-one prediction model stability, over-fitting, above-mentioned specified sample are prevented
Collection includes appropriate sample, such as tens of thousands of or hundreds of thousands sample, each sample for a user user information, in user information
Customer attribute information, the purchase product information of user and user's history action trail information can form thousands of Wesy Hu Te
Sign, such as " last time inlet wire interval number of days ", " last time inlet wire speech intention ", " are intended to " above-mentioned inlet wire interval times " for nearest one month
Score ", " whether being to produce dangerous user ", " whether wanting life insurance user ", " whether wanting credit card user " etc., in above-mentioned decision tree, often
The output of a leaf node is 0 or 1, i.e., the value of each user characteristics is 0 or 1, in training pattern, adjustable decision tree
Depth and number, then establish more decision trees, then the combination of each leaf child node of final output takes said combination
Value is trained using logistic regression algorithm, and above-mentioned logistic regression algorithm first can indicate independent variable x and dependent variable y with y=f (x)
Relationship, in the present embodiment, independent variable x is above-mentioned user characteristics, and dependent variable y is prediction result, i.e. user is intended to the industry handled
Business, finally by the logistic regression equation of the probability value of prediction result y, so that calculating user is intended to each single item business handled
The first probability value.
It is above-mentioned first casting unit 300 as described in, by level-one prediction model be calculated user be intended to handle it is each
After first probability value of business, according to these the first probability values, by the business of above-mentioned correspondence these the first probability values by the
The descending of one probability value is ranked up, i.e., by business according to the first probability value from being ranked up to small sortord greatly, in this way
The big business of first probability value preferentially comes front, and then carry out voice broadcast to user makes to save the time of user
Its accelerating selection promotes experience effect.
In a specific embodiment, such as 95511 service hotlines of safety group, when user beats the telephone number inlet wire,
System can get the inlet wire number of the user, and the corresponding inlet wire number is then found out in information bank according to the inlet wire number
User information recalls the user information when finding corresponding user information, is input to level-one prediction model and is counted
It calculates, obtains user and be intended to the first probability value of each single item business handled, if the first probability value of life insurance is 0.3, the of security
One probability value is 0.6, the first probability value of credit card is 0.2, is at this moment sorted from large to small by probability value: security, life insurance, credit
Card, then above-mentioned business successively can be carried out voice broadcast to user by system, and user is allowed to select.
In one embodiment, above-mentioned service hotline broadcast device, further includes:
Business unit is obtained, for obtaining the business of user's input；
Calling model unit calls corresponding second level prediction model for the classification according to the business；
Second computing unit calculates for inputting first user information in the second level prediction model, obtains institute
State the second probability value of each single item subordinate's option of business；
Second casting unit, for the descending according to second probability value by each single item of correspondence second probability value
Subordinate's business is successively broadcasted.
In the present embodiment, since the company using above-mentioned telephone service hot line system may have multiple classs of service to produce
Product, such as insurance company have and produce danger, life insurance, bank, credit card, security class of service, and each business is handled with multiple
Subordinate's business, such as producing subordinate's business of handling of dangerous business includes to produce danger to report a case to the security authorities, produce danger purchase, produce dangerous setting loss, credit card
Subordinate's business of handling of business includes that creditable card sets password for inquiry, credit card activates, card of making a credit inquiry can use amount etc.
Deng.So user can input the business for oneself being intended to handle after in step s 4, carrying out voice broadcast to above-mentioned business,
At this moment system can obtain the business of user's input,, can be according to should after the business for obtaining user's input in calling model unit
The classification of business calls corresponding second level prediction model, then inputs in second level prediction model user information and calculates, obtains
Second probability value of each single item subordinate's business of above-mentioned business, wherein the business of each classification respectively corresponds a second level prediction
Model, subordinate's business in the present embodiment are the business belonged in the above-mentioned scope of business, at this moment according to the second probability value will under
Category business is ranked up according to according to the descending of the second probability value, i.e., by subordinate's business according to corresponding second probability value by greatly extremely
It is small successively to sort, then successively ordering subordinate's business is broadcasted.
Wherein above-mentioned second level prediction model is to train for single business, for training the data of the second level prediction model
For in user information user property and user handled the historical behavior track of the business, using algorithm and training process
It is consistent with above-mentioned level-one prediction model, it is similarly gradient and promotes decision tree GBDT and logistic regression algorithm LR, it is no longer superfluous here
It states.
Certainly, since subordinate's classification of business is uncertain, subordinate only a point level-one may be also possible to point a multiple grades, such as second level
Subordinate divides three-level again later, or divides level Four again, corresponding, can preset three-level prediction model, level Four prediction model, this programme is not
The series of subordinate's classification is limited, every level-one step can carry out prediction user's intention according to the method described above, to improve user's body
It tests.
In another embodiment, above-mentioned service hotline broadcast device, further includes:
First recording unit, for recording the inlet wire information of the corresponding user of the inlet wire number, the inlet wire information is
The subscriber information message obtained according to information is inputted after subscriber's drop；
First adding unit, for the inlet wire information to be added to first user information to form second user letter
Breath；
First training unit, for training the level-one prediction model to form new level-one the second user information
Prediction model, and the second user information is stored in described information storehouse.
In the present embodiment, after above-mentioned business is carried out voice broadcast by system, user, which can select, oneself to be intended to handle
Business, then the user will generate new historical behavior track, and pass through these historical behaviors after the current inlet wire that finishes
Track can effectively obtain the intention of user, these all can serve as next user inlet wire and the basis predicted, institute again
With in the first recording unit, system records the inlet wire information of the corresponding user of inlet wire number, and above-mentioned inlet wire information is user
The historical behavior track of current inlet wire, such as the secondary subscriber's drop handle production danger and report a case to the security authorities or activate to credit card.As above
It states described in the first adding unit, after system gets above-mentioned carry out information by record, these progress information can be added to
In the first user information before, to form new second user information, and it is stored in information bank so as to next inlet wire tune
With, while these second user information can be trained level-one prediction model, so that new level-one prediction model is formed, so that
Level-one prediction model is in newest casting state at any time.
In one embodiment, above-mentioned service hotline broadcast device, further includes:
Number unit is judged, for judging whether the inlet wire number matches with the Subscriber Number that described information storehouse stores；
Business unit is called, when the Subscriber Number for the inlet wire number and described information storehouse storage mismatches, according to
Inlet wire history information in preset period calls hot spot service, the hot spot service be according to handle the descending of number into
In the business of row sequence, the business in preset forward ranking；
Third broadcasts unit, for broadcasting the hot spot series traffic item.
In the present embodiment, since the user of inlet wire number might not handle above-mentioned business or inlet wire was seeked advice from,
So not leaving historical behavior trace information, for such user, the business of current hotspot can be recommended, that is, compared more
The business that people handles or seeks advice from, above-mentioned hot spot service be in the business being ranked up according to the descending for handling number, it is preset
Business in forward ranking.It for above-mentioned such user, is not stored for information about in information bank, and training level-one prediction
Also without the information of the user in the user information of model, so the Subscriber Number for first judging that inlet wire number is stored with information bank is
No matching, if it does not match, illustrate that the user of the inlet wire number is first time inlet wire, it at this moment, can be according in preset period
Inlet wire history information call hot spot service, in the present embodiment, above-mentioned preset period customized can be arranged, such as 15 days or
One month etc., citing ground if it is determined that inlet wire number is first time inlet wire, at this moment can call the heat in one month in the database
Point business, then broadcasts these hot spot services.
Specifically, above-mentioned service hotline broadcast device, further includes:
Recording unit is obtained, for obtaining the inlet wire history information in preset period；
Analysis times unit, for analyzing time that user handles each single item business according to the inlet wire history information
Number；
Business sequencing unit, for the business successively to sort by the descending of number；
Tagged traffic unit, the service marker for being preset in ranking in the sequence is at the hot spot service.
It is understood that all inlet wire behaviors can all have record, these record storages are in database, in these notes
It can check which business inlet wire user once handled in record, these inlet wire history informations be obtained from database, so
The number for obtaining handling each business therefrom being analyzed afterwards, wherein business can successively be sorted by the descending of number, it then will be at this
In sequence, preset ranking in service marker be hot spot service, such as by successively sort by the descending of number the first eight in industry
Business is set as hot spot service, if what is obtained is the inlet wire historical record in preset period, what is obtained is in the preset period
Hot spot service.
Further, above-mentioned service hotline broadcast device, further includes:
Second recording unit, for recording the first inlet wire information of the inlet wire number, the first inlet wire information is root
The subscriber information message obtained according to information is inputted after user's first time inlet wire；
Second training unit, for obtaining using the first inlet wire information as level-one prediction model described in sample training
The first inlet wire information is stored in described information storehouse by the new level-one prediction model after the first inlet wire of inlet wire number.
After being broadcasted above-mentioned hot spot service, the first inlet wire information of the secondary subscriber's drop is recorded, i.e.,
The historical behavior information that the secondary inlet wire number corresponds to user is recorded, is then instructed these first inlet wire information as sample
Practice above-mentioned level-one prediction model, second level prediction model etc., to obtain the new level-one of the corresponding first inlet wire of inlet wire number
Prediction model, new second level prediction model；It is above-mentioned new level-one prediction model, new when user's inlet wire again of the secondary inlet wire
Second level prediction model can be used to predict that the user is intended to, meanwhile, above-mentioned first inlet wire information is stored in above- mentioned information library,
Facilitate calling.
Referring to Fig. 3, a kind of computer equipment is also provided in the embodiment of the present invention, which can be server,
Its internal structure can be as shown in Figure 3.The computer equipment includes processor, the memory, network connected by system bus
Interface and database.Wherein, the processor of the Computer Design is for providing calculating and control ability.The computer equipment is deposited
Reservoir includes non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system, computer program
And database.The internal memory provides environment for the operation of operating system and computer program in non-volatile memory medium.It should
The database of computer equipment is for data such as storage service hot line broadcasting methods.The network interface of the computer equipment be used for
External terminal passes through network connection communication.To realize a kind of service hotline casting side when the computer program is executed by processor
Method.
Above-mentioned processor executes the step of above-mentioned service hotline broadcasting method: entering service hotline by phone according to user
Inlet wire number obtain the first user information, first user information includes the production that the attribute information of user, user bought
Product information and user pass through the historical behavior trace information after phone inlet wire；By first user information input level-one prediction
It is calculated in model, obtains the first probability value that user is intended to each single item business in the class of service handled, the level-one is pre-
Model is surveyed specified sample set is trained to obtain based on gradient promotion decision tree and logistic regression algorithm；According to described first
The descending of probability value successively broadcasts each single item business of correspondence first probability value.
Above-mentioned computer equipment, it is described that the first user is obtained by the inlet wire number that phone enters service hotline according to user
The step of information, comprising: obtain the inlet wire number that user enters service hotline by phone；It is called and is believed according to the inlet wire number
Cease first user information of the correspondence inlet wire number in library.
In one embodiment, after above-mentioned the step of being broadcasted the series menu, comprising: obtain user's input
Business；Classification according to the business calls corresponding second level prediction model；First user information is inputted described two
It is calculated in grade prediction model, obtains the second probability value of each single item subordinate's business of the business；According to second probability value
Descending each single item subordinate business of correspondence second probability value is successively broadcasted.
In one embodiment, above-mentioned to record the corresponding inlet wire information of the inlet wire number, according to the inlet wire information
The subscriber information message that information obtains is inputted after subscriber's drop；The inlet wire information is added to first user information with shape
At second user information, by the second user information training level-one prediction model to form new level-one prediction model,
And the second user information is stored in described information storehouse.
In one embodiment, after the step of above-mentioned acquisition user enters the inlet wire number of service hotline by phone,
It include: to judge whether the inlet wire number matches with the Subscriber Number that described information storehouse stores；If it is not, then according in preset period
Inlet wire history information call hot spot service, the hot spot service is according to handling the industry that the descending of number is ranked up
Business in business, in preset forward ranking；The hot spot service is broadcasted.
In one embodiment, the step of above-mentioned inlet wire history information according in preset period calls hot spot service
Before, comprising: obtain the inlet wire history information in preset period；User is analyzed according to the inlet wire history information to do
Manage the number of each single item business；The business is successively sorted by the descending of number；The institute that will be preset in the sequence in ranking
Service marker is stated into the hot spot service.
In one embodiment, after above-mentioned the step of being broadcasted the hot spot series traffic item, comprising: record institute
The first inlet wire information of inlet wire number is stated, the first inlet wire information is obtained according to input information after user's first time inlet wire
Subscriber information message；Using the first inlet wire information as level-one prediction model described in sample training, the inlet wire number is obtained
New level-one prediction model after first inlet wire, is stored in described information storehouse for the first inlet wire information.
It will be understood by those skilled in the art that structure shown in Fig. 3, only part relevant to application scheme is tied
The block diagram of structure does not constitute the restriction for the computer equipment being applied thereon to application scheme.
One embodiment of the invention also provides a kind of computer readable storage medium, is stored thereon with computer program, calculates
Machine program realizes a kind of service hotline broadcasting method when being executed by processor, specifically: service is entered by phone according to user
The inlet wire number of hot line obtains the first user information, and first user information includes that the attribute information of user, user bought
Product information and user pass through the historical behavior trace information after phone inlet wire；First user information is inputted into level-one
It is calculated in prediction model, obtains the first probability value that user is intended to each single item business in the class of service handled, described one
Grade prediction model is based on gradient promotion decision tree and logistic regression algorithm is trained to obtain to specified sample set；According to described
The descending of first probability value successively broadcasts each single item business of correspondence first probability value.
Above-mentioned computer readable storage medium, it is described to be obtained according to user by the inlet wire number that phone enters service hotline
The step of first user information, comprising: obtain the inlet wire number that user enters service hotline by phone；According to the inlet wire number
First user information of the correspondence inlet wire number in code recalls information library.
In one embodiment, after above-mentioned the step of being broadcasted the series menu, comprising: obtain user's input
Business；Classification according to the business calls corresponding second level prediction model；First user information is inputted described two
It is calculated in grade prediction model, obtains the second probability value of each single item subordinate's business of the business；According to second probability value
Descending each single item subordinate business of correspondence second probability value is successively broadcasted.
In one embodiment, above-mentioned to record the corresponding inlet wire information of the inlet wire number, according to the inlet wire information
The subscriber information message that information obtains is inputted after subscriber's drop；The inlet wire information is added to first user information with shape
At second user information, by the second user information training level-one prediction model to form new level-one prediction model,
And the second user information is stored in described information storehouse.
In one embodiment, after the step of above-mentioned acquisition user enters the inlet wire number of service hotline by phone,
It include: to judge whether the inlet wire number matches with the Subscriber Number that described information storehouse stores；If it is not, then according in preset period
Inlet wire history information call hot spot service, the hot spot service is according to handling the industry that the descending of number is ranked up
Business in business, in preset forward ranking；The hot spot service is broadcasted.
In one embodiment, the step of above-mentioned inlet wire history information according in preset period calls hot spot service
Before, comprising: obtain the inlet wire history information in preset period；User is analyzed according to the inlet wire history information to do
Manage the number of each single item business；The business is successively sorted by the descending of number；The institute that will be preset in the sequence in ranking
Service marker is stated into the hot spot service.
In one embodiment, after above-mentioned the step of being broadcasted the hot spot series traffic item, comprising: record institute
The first inlet wire information of inlet wire number is stated, the first inlet wire information is obtained according to input information after user's first time inlet wire
Subscriber information message；Using the first inlet wire information as level-one prediction model described in sample training, the inlet wire number is obtained
New level-one prediction model after first inlet wire, is stored in described information storehouse for the first inlet wire information.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can store and a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
Any reference used in provided herein and embodiment to memory, storage, database or other media,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, mono- diversified forms of RAM can obtain,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double speed are according to rate SDRAM (SSRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
It should be noted that, in this document, the terms "include", "comprise" or its any other variant are intended to non-row
His property includes, so that the process, device, article or the method that include a series of elements not only include those elements, and
And further include other elements that are not explicitly listed, or further include for this process, device, article or method institute it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including being somebody's turn to do
There is also other identical elements in the process, device of element, article or method.
The above description is only a preferred embodiment of the present invention, is not intended to limit the scope of the invention, all utilizations
Equivalent structure or equivalent flow shift made by description of the invention and accompanying drawing content is applied directly or indirectly in other correlations
Technical field, be included within the scope of the present invention.
Claims (10)
1. a kind of service hotline broadcasting method characterized by comprising
The first user information, the first user information packet are obtained by the inlet wire number that phone enters service hotline according to user
It includes the attribute information of user, the product information that user bought and user and passes through the historical behavior track letter after phone inlet wire
Breath；
First user information is inputted in level-one prediction model and is calculated, user is obtained and is intended in the class of service handled
First probability value of each single item business, the level-one prediction model are based on gradient and promote decision tree and logistic regression algorithm to finger
Determine sample set to be trained to obtain；
The each single item business of correspondence first probability value is successively broadcasted according to the descending of first probability value.
2. service hotline broadcasting method according to claim 1, which is characterized in that described to be entered according to user by phone
The inlet wire number of service hotline obtains the step of the first user information, comprising:
Obtain the inlet wire number that user enters service hotline by phone；
According to first user information of the correspondence inlet wire number in the inlet wire number recalls information library.
3. service hotline broadcasting method according to claim 1, which is characterized in that described according to first probability value
After the step of descending is successively broadcasted each single item business of correspondence first probability value, comprising:
Obtain the business of user's input；
Classification according to the business calls corresponding second level prediction model；
First user information is inputted in the second level prediction model and is calculated, each single item subordinate's business of the business is obtained
The second probability value；
The each single item subordinate business of correspondence second probability value is successively broadcasted according to the descending of second probability value.
4. service hotline broadcasting method according to claim 1, which is characterized in that described according to first probability value
After the step of descending is successively broadcasted each single item business of correspondence first probability value, comprising:
The corresponding inlet wire information of the inlet wire number is recorded, the inlet wire information is obtained according to input information after subscriber's drop
Subscriber information message；
The inlet wire information is added to first user information to form second user information,
By the second user information training level-one prediction model to form new level-one prediction model, and by described second
User information is stored in described information storehouse.
5. service hotline broadcasting method according to claim 2, which is characterized in that the acquisition user is entered by phone
After the step of inlet wire number of service hotline, comprising:
Judge whether the inlet wire number matches with the Subscriber Number that described information storehouse stores；
If it is not, then calling hot spot service according to the inlet wire history information in preset period, the hot spot service is in foundation
It handles in the business that the descending of number is ranked up, presets the business in ranking；
The hot spot service is broadcasted.
6. service hotline broadcasting method according to claim 5, which is characterized in that the inlet wire according in preset period
History information called before the step of hot spot service, comprising:
Obtain the inlet wire history information in preset period；
The number that user handles each single item business is analyzed according to the inlet wire history information；
The business is successively sorted by the descending of number；
By the service marker preset in the sequence in ranking at the hot spot service.
7. service hotline broadcasting method according to claim 5, which is characterized in that described to broadcast the hot spot service
After the step of report, comprising:
Record the first inlet wire information of the inlet wire number, the first inlet wire information is according to inputting after user's first time inlet wire
The subscriber information message that information obtains；
Using the first inlet wire information as level-one prediction model described in sample training, after obtaining the first inlet wire of inlet wire number
New level-one prediction model, the first inlet wire information is stored in described information storehouse.
8. a kind of service hotline broadcast device characterized by comprising
Acquiring unit, it is described for obtaining the first user information by the inlet wire number that phone enters service hotline according to user
First user information includes that the product information that the attribute information of user, user bought and user pass through going through after phone inlet wire
History action trail information；
Computing unit is calculated for inputting first user information in level-one prediction model, is obtained user and is intended to do
First probability value of each single item business in the class of service of reason, the level-one prediction model are based on gradient and promote decision tree and patrol
It collects regression algorithm specified sample set is trained to obtain；
Broadcast unit, for the descending according to first probability value by each single item business of correspondence first probability value successively
It is broadcasted.
9. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In the step of processor realizes any one of claims 1 to 7 the method when executing the computer program.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any one of claims 1 to 7 is realized when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811185764.7A CN109600523A (en) | 2018-10-11 | 2018-10-11 | Service hotline broadcasting method, device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811185764.7A CN109600523A (en) | 2018-10-11 | 2018-10-11 | Service hotline broadcasting method, device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109600523A true CN109600523A (en) | 2019-04-09 |
Family
ID=65957384
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811185764.7A Pending CN109600523A (en) | 2018-10-11 | 2018-10-11 | Service hotline broadcasting method, device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109600523A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110233946A (en) * | 2019-06-17 | 2019-09-13 | 三角兽（北京）科技有限公司 | Execute outbound call service method, electronic equipment and computer readable storage medium |
| CN110572521A (en) * | 2019-09-20 | 2019-12-13 | 中国银行股份有限公司 | Voice broadcasting method and device |
| CN111787164A (en) * | 2020-07-16 | 2020-10-16 | 阳光保险集团股份有限公司 | Incoming call prompting method and device, electronic equipment and computer readable medium |
| CN111833072A (en) * | 2019-08-28 | 2020-10-27 | 北京嘀嘀无限科技发展有限公司 | Broadcast item pushing method, pushing device and readable storage medium |
| CN114025044A (en) * | 2021-10-15 | 2022-02-08 | 中国人民财产保险股份有限公司 | Control method and device of call center and electronic equipment |
| CN115529387A (en) * | 2022-09-16 | 2022-12-27 | 中国建设银行股份有限公司 | Service index generation method, device, server and storage medium |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101094271A (en) * | 2007-07-20 | 2007-12-26 | 华为技术有限公司 | Method and device for creating menu of automatic voice responsion system |
| CN103139404A (en) * | 2013-01-25 | 2013-06-05 | 西安电子科技大学 | System and method for generating interactive voice response display menu based on voice recognition |
| CN103873706A (en) * | 2012-12-18 | 2014-06-18 | 北京裕丰大通科技有限公司 | Dynamic intelligent voice recognition IVR (Interactive Voice Response) service system |
| CN103886487A (en) * | 2014-03-28 | 2014-06-25 | 焦点科技股份有限公司 | Individualized recommendation method and system based on distributed B2B platform |
| CN104052883A (en) * | 2013-03-13 | 2014-09-17 | 华为技术有限公司 | Method, device and system for obtaining menu information |
| US20160103833A1 (en) * | 2012-06-01 | 2016-04-14 | Google Inc. | Ranking query results using biometric parameters |
| CN105704106A (en) * | 2014-11-27 | 2016-06-22 | 腾讯科技（深圳）有限公司 | Visualization IVR realization method and mobile terminal |
| CN106572272A (en) * | 2015-10-12 | 2017-04-19 | 阿里巴巴集团控股有限公司 | IVR voice menu determination method and apparatus |
| CN106780052A (en) * | 2017-01-10 | 2017-05-31 | 上海诺悦智能科技有限公司 | Method and system are recommended in insurance service based on classification customer behavior analysis |
| CN106998397A (en) * | 2016-01-25 | 2017-08-01 | 平安科技（深圳）有限公司 | The voice broadcast method and system of multi-service type |
| CN107704946A (en) * | 2017-08-29 | 2018-02-16 | 平安科技（深圳）有限公司 | Electronic installation, Voice Navigation needing forecasting method and storage medium |
| CN108595526A (en) * | 2018-03-28 | 2018-09-28 | 北京奇艺世纪科技有限公司 | Resource recommendation method and device |
-
2018
- 2018-10-11 CN CN201811185764.7A patent/CN109600523A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101094271A (en) * | 2007-07-20 | 2007-12-26 | 华为技术有限公司 | Method and device for creating menu of automatic voice responsion system |
| US20160103833A1 (en) * | 2012-06-01 | 2016-04-14 | Google Inc. | Ranking query results using biometric parameters |
| CN103873706A (en) * | 2012-12-18 | 2014-06-18 | 北京裕丰大通科技有限公司 | Dynamic intelligent voice recognition IVR (Interactive Voice Response) service system |
| CN103139404A (en) * | 2013-01-25 | 2013-06-05 | 西安电子科技大学 | System and method for generating interactive voice response display menu based on voice recognition |
| CN104052883A (en) * | 2013-03-13 | 2014-09-17 | 华为技术有限公司 | Method, device and system for obtaining menu information |
| CN103886487A (en) * | 2014-03-28 | 2014-06-25 | 焦点科技股份有限公司 | Individualized recommendation method and system based on distributed B2B platform |
| CN105704106A (en) * | 2014-11-27 | 2016-06-22 | 腾讯科技（深圳）有限公司 | Visualization IVR realization method and mobile terminal |
| CN106572272A (en) * | 2015-10-12 | 2017-04-19 | 阿里巴巴集团控股有限公司 | IVR voice menu determination method and apparatus |
| CN106998397A (en) * | 2016-01-25 | 2017-08-01 | 平安科技（深圳）有限公司 | The voice broadcast method and system of multi-service type |
| CN106780052A (en) * | 2017-01-10 | 2017-05-31 | 上海诺悦智能科技有限公司 | Method and system are recommended in insurance service based on classification customer behavior analysis |
| CN107704946A (en) * | 2017-08-29 | 2018-02-16 | 平安科技（深圳）有限公司 | Electronic installation, Voice Navigation needing forecasting method and storage medium |
| CN108595526A (en) * | 2018-03-28 | 2018-09-28 | 北京奇艺世纪科技有限公司 | Resource recommendation method and device |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110233946A (en) * | 2019-06-17 | 2019-09-13 | 三角兽（北京）科技有限公司 | Execute outbound call service method, electronic equipment and computer readable storage medium |
| CN111833072A (en) * | 2019-08-28 | 2020-10-27 | 北京嘀嘀无限科技发展有限公司 | Broadcast item pushing method, pushing device and readable storage medium |
| CN110572521A (en) * | 2019-09-20 | 2019-12-13 | 中国银行股份有限公司 | Voice broadcasting method and device |
| CN111787164A (en) * | 2020-07-16 | 2020-10-16 | 阳光保险集团股份有限公司 | Incoming call prompting method and device, electronic equipment and computer readable medium |
| CN111787164B (en) * | 2020-07-16 | 2021-06-15 | 阳光保险集团股份有限公司 | Incoming call prompting method and device, electronic equipment and computer readable medium |
| CN114025044A (en) * | 2021-10-15 | 2022-02-08 | 中国人民财产保险股份有限公司 | Control method and device of call center and electronic equipment |
| CN115529387A (en) * | 2022-09-16 | 2022-12-27 | 中国建设银行股份有限公司 | Service index generation method, device, server and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109600523A (en) | Service hotline broadcasting method, device, computer equipment and storage medium | |
| CN107403326A (en) | A kind of Insurance Fraud recognition methods and device based on teledata | |
| CN107958317A (en) | A kind of method and apparatus that crowdsourcing participant is chosen in crowdsourcing project | |
| CN108073659A (en) | A kind of love and marriage object recommendation method and device | |
| CN106156809A (en) | For updating the method and device of disaggregated model | |
| CN109272408A (en) | Vehicle loan financial product intelligent recommendation method, apparatus, equipment and medium | |
| CN110516910A (en) | Declaration form core based on big data protects model training method and core protects methods of risk assessment | |
| CN109300039A (en) | The method and system of intellectual product recommendation are carried out based on artificial intelligence and big data | |
| CN107038511A (en) | A kind of method and device for determining risk assessment parameter | |
| CN109492191A (en) | Calculate method, apparatus, computer equipment and the storage medium of probability of insuring | |
| CN116862668B (en) | Intelligent collecting accelerating method for improving collecting accelerating efficiency | |
| CN108416684A (en) | A kind of credibility appraisal procedure, device and the server of account main body | |
| CN109840778A (en) | The recognition methods of fraudulent user and device, readable storage medium storing program for executing | |
| CN110415036A (en) | Determination method, apparatus, computer equipment and the storage medium of user gradation | |
| CN108776846A (en) | Recommend method, apparatus, computer equipment and storage medium | |
| CN106095939A (en) | The acquisition methods of account authority and device | |
| CN107274141A (en) | A kind of event-handling method and the network equipment | |
| CN111898931A (en) | Variable-based strategic wind control engine implementation method and device and computer equipment | |
| CN112035775A (en) | User identification method and device based on random forest model and computer equipment | |
| CN111626864A (en) | Information pushing method and device, storage medium and electronic device | |
| CN116501979A (en) | Information recommendation method, information recommendation device, computer equipment and computer readable storage medium | |
| KR100601885B1 (en) | Call recommender system for outbound call center | |
| CN111091460A (en) | Data processing method and device | |
| CN111986018A (en) | Bill collection prompting method and device based on preset collection prompting system and electronic equipment | |
| Terko et al. | Credit scoring model implementation in a microfinance context |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190409 |
|
| RJ01 | Rejection of invention patent application after publication |