CN114844992A - Multilayer IVR navigation processing method - Google Patents
Multilayer IVR navigation processing method Download PDFInfo
- Publication number
- CN114844992A CN114844992A CN202210398533.4A CN202210398533A CN114844992A CN 114844992 A CN114844992 A CN 114844992A CN 202210398533 A CN202210398533 A CN 202210398533A CN 114844992 A CN114844992 A CN 114844992A
- Authority
- CN
- China
- Prior art keywords
- navigation
- ivr
- voice
- service
- execution
- 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
Images
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
- H04M3/5166—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
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
- G10L2015/223—Execution procedure of a spoken command
Abstract
The invention provides a method for multilayer IVR navigation processing. Receiving customer voice through a front-end processor, and identifying user requirements according to the customer voice; when the user requirement is successfully identified, directionally jumping to a corresponding navigation layer; and when the user requirement identification fails, sending the client voice to a manual processing end, controlling the jump of a navigation layer through the manual processing end, and updating the directional jump after the jump. The invention has the beneficial effects that: the invention solves the problem that the client can know the operation by listening to a large number of key prompts, saves the time of dialing by the client, and has the advantages that because the invention is based on the artificial intelligence technology, the artificial intelligence technology can not only realize intelligent navigation layer jump when executing, but also can carry out self-learning, and the keywords of the client in navigation are continuously updated through all the services processed.
Description
Technical Field
The invention relates to the technical field of voice service, in particular to a method for multilayer IVR navigation processing.
Background
The IVR service is a voice service value-added service, and after a user accesses the IVR system through a telephone, the IVR system provides a corresponding IVR service flow for the user. For example, after the user dials the designated phone number, the IVR system provides a menu broadcasting process to the user, and the user can enter an information query process, an information modification process, a manual service process and the like through corresponding selection operations.
However, the current IVR system provides a fixed IVR service flow, for example, after different users access the IVR system, the users enter the same menu broadcasting flow. Often, a user needs to operate for many times or listen to a voice broadcast for a long time to be able to enter a specific flow required by the user, which causes that the time for the user to access the IVR system is too long, and more system resources are occupied.
At present, when a client dials an enterprise phone, a multilayer IVR navigation condition is encountered, for example, 10086 prompts that a certain operation key 1 needs to be performed, and the key 2 … … client often needs to listen to a plurality of prompts and then knows how many keys need to be pressed, which wastes much time for the client.
Disclosure of Invention
The invention provides a method for multilayer IVR navigation processing, which is used for solving the problem that in the prior art, a key is required or a service flow is complicated to jump.
A method of multi-tier IVR navigation processing, comprising:
receiving customer voice through a front-end processor, and identifying user requirements according to the customer voice;
when the user requirements are successfully identified, directionally jumping to a corresponding navigation layer;
and when the user requirement identification fails, sending the client voice to a manual processing end, controlling the jump of a navigation layer through the manual processing end, and updating the directional jump after the jump.
Preferably, the method further comprises:
receiving a communication request of a user, and initializing a voice storage area of the front-end robot;
after the initialization, connecting the communication request and carrying out audio recording;
determining a mixed waveform of the audio according to the audio recording;
performing signal source separation on the mixed waveform to determine a plurality of audio signal sources;
determining the fundamental frequency of each signal source according to the signal sources, comparing the fundamental frequency with the mixed audio of the mixed waveform, and determining the contrast difference;
according to the contrast difference, deleting the noise audio frequency of the signal source with the difference reaching a preset threshold value;
and after the noise audio is deleted, audio splicing is carried out to generate new spliced voice.
Preferably, the method further comprises:
performing textualization on the spliced voice to determine a voice text;
according to the voice text, carrying out requirement item keyword calibration and determining requirement keywords;
according to the requirement keywords, carrying out requirement classification and determining a requirement type;
judging whether a corresponding requirement keyword exists in a requirement database or not according to the requirement type;
when the requirement keyword exists, determining the requirement of the user;
and when the requirement key words do not exist, sending the spliced voice to a manual processing end.
Preferably, the method further comprises:
pre-configuring a training set voice and a test set voice, and determining a training text of the training set voice and a test text of the test set voice;
acquiring a first acoustic feature of the training set voice, and determining a first keyword of the training text in the training set voice;
acquiring a second acoustic feature of the test set voice, and determining a second keyword of the test text in the test set voice;
establishing a deep network model based on four branches, and leading the first acoustic feature, the second acoustic feature, the first keyword and the second keyword into the deep network model of the four branches together for training in a fusion training mode to generate a text conversion model;
and performing textual conversion on the spliced voice through the text conversion model.
Preferably, the method further comprises:
acquiring all navigation processing services of IVR navigation and generating a navigation set;
generating a navigation list according to the navigation set;
establishing a keyword library of navigation service according to the navigation list;
establishing a contrast triggering mechanism in the keyword library, stacking the navigation services in each navigation list layer by layer to generate a multilayer navigation menu, and triggering each navigation service in the multilayer navigation menu through the contrast triggering mechanism.
Preferably, the method further comprises:
acquiring navigation information corresponding to each navigation layer in the navigation layers;
determining a corresponding service address of each navigation service end according to the navigation information;
performing service skipping according to the service address;
triggering the service end to record the user again according to the service jump;
forming a user order according to the user record;
and determining an order implementation flow according to the user order, and performing service processing according to the order implementation flow.
Preferably, the method further comprises:
acquiring the user order and determining a business execution process; wherein the content of the first and second substances,
responding to the business execution flow, sending a processing request for order execution to a supply server, and receiving an execution node corresponding to the business execution flow returned by the supply server; wherein the content of the first and second substances,
the provisioning server is further configured to determine an ID for each executing node;
initiating execution verification for the business execution flow to each execution node execution rule based on the ID of the execution node;
after receiving an execution verification result returned by each execution node, updating a service execution state in target order information associated with a user based on user information contained in the execution verification;
and sending the user order information after the service execution state is updated to the supply server so as to obtain an order completion response authorized by the supply server based on the order information after the service execution state is updated.
Preferably, the method further comprises:
receiving a communication request of a user, and carrying out an IVR flow configuration request according to the communication request;
obtaining configured IVR flow configuration information;
generating an IVR configuration page according to the configured IVR flow configuration information, wherein the IVR configuration page comprises configuration areas of all IVR branch flows, and the configuration area of each IVR branch flow comprises a node tree of the IVR branch flow;
receiving editing information in the IVR configuration page, and updating the IVR process configuration information according to the editing information;
obtaining IVR flow configuration information, wherein the IVR flow configuration information comprises execution conditions of all IVR branch flows, nodes included in all IVR branch flows and execution logics of all nodes in all IVR branch flows;
selecting an IVR branch flow meeting execution conditions according to the IVR flow configuration information and the incoming call request as an IVR branch flow to be executed;
and sequentially executing IVR instructions corresponding to the nodes according to the execution logic of the nodes in the IVR branch flow to be executed.
Preferably, the method further comprises:
acquiring navigation information of a manual processing end;
determining a corresponding navigation page and artificial voice keywords according to the navigation information;
storing a navigation page and artificial voice keywords for updating the navigation layer;
determining original navigation voice keywords according to the navigation page;
matching the original navigation voice keywords with the keywords of the artificial voice keywords;
and updating the original navigation voice keywords according to the keyword matching.
Preferably, the method further comprises:
calling a mobile phone number and customer information of a customer according to the front-end robot;
determining a business database of the client according to the client information;
judging whether the same service order exists in the service database according to the service database;
when the same service order exists, automatically classifying the customer requirements, automatically broadcasting a multi-level voice navigation menu for unclear description of customer information, and enabling options in the navigation of the customer voice selection menu to enter a problem processing or requirement handling processing system;
when the same service order does not exist, judging whether a corresponding navigation layer exists or not;
when the corresponding navigation layer exists, performing service processing according to the corresponding navigation layer;
and when the corresponding navigation layer does not exist, jumping to a manual processing end for manual processing.
The invention has the beneficial effects that: the invention solves the problem that the client can know the operation by listening to a large number of key prompts, saves the time of dialing by the client, and has the advantages that because the invention is based on the artificial intelligence technology, the artificial intelligence technology can not only realize intelligent navigation layer jump when executing, but also can carry out self-learning, and the keywords of the client in navigation are continuously updated through all the services processed.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and drawings.
The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:
FIG. 1 is a flow chart of a method of multi-tier IVR navigation processing in an embodiment of the invention;
FIG. 2 is a flow chart illustrating a process of navigating an order according to an embodiment of the present invention.
Detailed Description
The preferred embodiments of the present invention will be described in conjunction with the accompanying drawings, and it will be understood that they are described herein for the purpose of illustration and explanation and not limitation.
Example 1:
as shown in fig. 1, the present invention is a method for multi-layer IVR navigation processing, including:
receiving customer voice through a front-end processor, and identifying user requirements according to the customer voice;
when the user requirement is successfully identified, directionally jumping to a corresponding navigation layer;
and when the user requirement identification fails, sending the client voice to a manual processing end, controlling the skipping of a navigation layer through the manual processing end, and performing directional skipping updating after skipping.
The working principle of the technical scheme is as follows:
the invention relates to an intelligent method for executing IVR service technology by using artificial intelligence technology, in the process, a front-end robot receives the voice of a client, identifies the user requirement through voice analysis, directly jumps to a specified navigation layer, reduces the waiting time of the client, if the client requirement is not identified, the client requirement is sent to a processor through a message broadcasting mode, the client voice is directly listened to through voice conversion or customer service, the client requirement is manually processed and jumped to the specified navigation, and the client requirement is stored at the same time, so that the quick response is realized when the same similar requirement is met next time.
The beneficial effects of the above technical scheme are: the invention solves the problem that the client can know the operation by listening to a large number of key prompts, saves the time of dialing by the client, and has the advantages that because the invention is based on the artificial intelligence technology, the artificial intelligence technology can not only realize intelligent navigation layer jump when executing, but also can carry out self-learning, and the keywords of the client in navigation are continuously updated through all the services processed.
Example 2:
preferably, the method further comprises:
receiving a communication request of a user, and initializing a voice storage area of the front-end robot;
after the initialization, connecting the communication request and carrying out audio recording;
determining a mixed waveform of the audio according to the audio recording;
performing signal source separation on the mixed waveform to determine a plurality of audio signal sources;
determining the fundamental frequency of each signal source according to the signal sources, comparing the fundamental frequency with the mixed audio of the mixed waveform, and determining the contrast difference;
according to the contrast difference, deleting the noise audio frequency of the signal source with the difference reaching a preset threshold value;
and after the noise audio is deleted, audio splicing is carried out to generate new spliced voice.
The working principle of the technical scheme is as follows: the invention is a voice communication mode, so that the symptoms of weak signal and interference are necessarily existed. Because many noises occur, the present invention adopts a way of removing noises and performing speech concatenation in this case. In the process, the voice of the client is embodied in the form of voice waveforms, the voice of the client has voiceprint characteristics of the client, but when noise caused by signals does not have the voiceprint characteristics of the user, signal sources generated by the signals are different, and at the moment, the difference degree of the integral mixed audio and the audio of the voice of the client is definitely smaller through sound source division, namely fundamental frequency division, so that the difference degree is particularly large, namely noise is processed, and then the voice of the client is spliced to obtain clearer voice.
The beneficial effects of the above technical scheme are: the invention can integrally process the customer voice to obtain clearer customer voice, and can better recognize the customer requirements firstly, and can achieve the function of reaching the navigation layer more quickly due to better textualization.
Example 3:
preferably, the method further comprises:
performing textualization on the spliced voice to determine a voice text;
according to the voice text, carrying out requirement item keyword calibration and determining requirement keywords;
according to the requirement keywords, carrying out requirement classification and determining a requirement type;
judging whether a corresponding requirement keyword exists in a requirement database or not according to the requirement type;
when the requirement keyword exists, determining the requirement of the user;
and when the requirement key words do not exist, sending the spliced voice to a manual processing end.
The working principle of the technical scheme is as follows: the invention realizes the voice text conversation during navigation, and can realize clearer and clearer determination of the user requirement and extraction of the keywords in the user voice only through voice textualization. According to the keywords, the system directly jumps to the corresponding navigation layer, so that better navigation effect and navigation accuracy are achieved.
The beneficial effects of the above technical scheme are: compared with the prior art, the method and the device realize the skipping of the navigation layer through the keywords, the skipping can be more accurate, skipping errors can not occur, and secondly, the voice order recording of the client can be better carried out, order writing is realized, and artificial intelligent learning is facilitated.
Example 4:
preferably, the method further comprises:
pre-configuring a training set voice and a test set voice, and determining a training text of the training set voice and a test text of the test set voice;
acquiring a first acoustic feature of the training set voice, and determining a first keyword of the training text in the training set voice;
acquiring a second acoustic feature of the test set voice, and determining a second keyword of the test text in the test set voice;
establishing a deep network model based on four branches, and leading the first acoustic feature, the second acoustic feature, the first keyword and the second keyword into the deep network model of the four branches together for training in a fusion training mode to generate a text conversion model;
and performing textual conversion on the spliced voice through the text conversion model.
The working principle of the technical scheme is as follows: the invention is mainly based on artificial intelligence technology, so that in the process of voice navigation, the invention establishes a deep network of four branches. The deep network of the four tributaries is also a main technical purpose of the invention, and compared with the prior art that artificial intelligence learning is carried out only through texts or only through voice, the accuracy rate of single-item mode recognition is lower certainly. Secondly, although the two are fused for more accurate requirements and speech recognition, the technical effect of more accurate text conversion is achieved.
The beneficial effects of the above technical scheme are: the invention realizes the improvement of the accuracy of text conversion, firstly the improvement of the speech accuracy, the improvement of the textualization accuracy and the improvement of the demand recognition accuracy, and finally the learning effect is better when artificial intelligence is used for learning.
Example 5:
preferably, the method further comprises:
acquiring all navigation processing services of IVR navigation and generating a navigation set;
generating a navigation list according to the navigation set;
establishing a keyword library of navigation service according to the navigation list;
establishing a contrast triggering mechanism in the keyword library, stacking the navigation services in each navigation list layer by layer to generate a multilayer navigation menu, and triggering each navigation service in the multilayer navigation menu through the contrast triggering mechanism.
The working principle of the technical scheme is as follows: when the navigation is performed, the navigation function menu based on function execution is mainly used in the prior art, but a navigation layer with stacked layers is generated; each layer executes different navigation functions, so that each navigation function can be executed more accurately, and the navigation is more accurate.
The beneficial effects of the above technical scheme are: compared with the mode in the prior art, the method has better stereoscopic impression, and the stereoscopic impression can trigger different navigation functions more accurately and execute the navigation function more quickly.
Example 6:
preferably, as shown in fig. 2, the method further comprises:
acquiring navigation information corresponding to each navigation layer in the navigation layers;
determining a corresponding service address of each navigation service end according to the navigation information;
performing service skipping according to the service address;
triggering the service end to record the user again according to the service jump;
forming a user order according to the user record;
and determining an order implementation flow according to the user order, and performing service processing according to the order implementation flow.
The working principle of the technical scheme is as follows: after the navigation layer jumps, the corresponding navigation service is executed, the navigation service is necessarily recorded again when being executed, in the process of recording again, a corresponding order is generated according to the voice recognition of the recording, and the service flow of the user is determined according to the service order, so that the service processing is realized.
The beneficial effects of the above technical scheme are: after navigation, the method and the system can correspondingly process different services, and perform more refined and accurate execution on the services according to the order and the service flow.
Example 7:
preferably, the method further comprises:
acquiring the user order and determining a business execution process; wherein the content of the first and second substances,
responding to the business execution flow, sending a processing request for order execution to a supply server, and receiving an execution node corresponding to the business execution flow returned by the supply server; wherein the content of the first and second substances,
the provisioning server is further configured to determine an ID for each executing node;
initiating execution verification for the business execution flow to each execution node execution rule based on the ID of the execution node;
after receiving an execution verification result returned by each execution node, updating a service execution state in target order information associated with a user based on user information contained in the execution verification;
and sending the user order information after the service execution state is updated to the supply server so as to obtain an order completion response authorized by the supply server based on the order information after the service execution state is updated.
The working principle and the beneficial effects of the technical scheme are that: when executing each user order, the invention divides nodes according to the execution flow of the user service, which has the advantages of task flow of the service, realizing the progress supervision of the service after the task flow is realized, paying attention to the execution state of the service at all times after the service supervision is carried out, and updating the service state in real time through the execution state, thereby ensuring that the user can make clear the execution progress of the service and better assisting the user to realize the own requirement.
Example 8:
preferably, the method further comprises:
receiving a communication request of a user, and carrying out an IVR flow configuration request according to the communication request;
obtaining configured IVR flow configuration information;
generating an IVR configuration page according to the configured IVR flow configuration information, wherein the IVR configuration page comprises configuration areas of all IVR branch flows, and the configuration area of each IVR branch flow comprises a node tree of the IVR branch flow;
receiving editing information in the IVR configuration page, and updating the IVR process configuration information according to the editing information;
obtaining IVR flow configuration information, wherein the IVR flow configuration information comprises execution conditions of all IVR branch flows, nodes included in all IVR branch flows and execution logics of all nodes in all IVR branch flows;
selecting an IVR branch flow meeting execution conditions according to the IVR flow configuration information and the incoming call request as an IVR branch flow to be executed;
and sequentially executing IVR instructions corresponding to the nodes according to the execution logic of the nodes in the IVR branch flow to be executed.
The working principle of the technical scheme is as follows: the invention is IVR flow, so when receiving the voice of the user, the flow configuration of IVR must be carried out, in the configuration process, different services have corresponding IVR configuration pages, thus realizing clearer and clearer service division, and according to the service division, through the execution logic of different nodes. Executing the execution logic of each node and executing the corresponding IVR instruction.
The beneficial effects of the above technical scheme are: the invention makes the service more clear through IVR configuration, thereby being more accurate when implementing the requirement of the user.
Example 9:
preferably, the method further comprises:
acquiring navigation information of a manual processing end;
determining a corresponding navigation page and artificial voice keywords according to the navigation information;
storing a navigation page and artificial voice keywords for updating the navigation layer;
determining original navigation voice keywords according to the navigation page;
matching the original navigation voice keywords with the keywords of the artificial voice keywords;
and updating the original navigation voice keywords according to the keyword matching.
The working principle of the technical scheme is as follows: when the invention is used for manual processing, the manual processing is bound to be realized by automatic navigation, and voice recognition cannot be carried out, so that the voice requirement is determined.
The beneficial effects of the above technical scheme are: the invention can realize more diversification of manual automatic navigation through continuous updating, the automatic navigation can process more services, and the manual processing end can reduce more manpower requirements, thereby realizing double saving on cost and procedure.
In one embodiment, the navigation voice keyword updating method further comprises the following steps:
firstly, based on an original navigation voice keyword, establishing a word quantity fluctuation model based on the original navigation voice keyword:
wherein C represents a word quantity model of an original navigation voice keyword; n represents the total number of words of the navigation voice keyword;
the characteristic parameters represent the ith navigation voice keyword; h is i The number of synonyms representing the ith navigation voice keyword; w is a i A navigation weight representing the ith navigation voice keyword;
the table represents the word size fluctuation value of the navigation voice keyword (used for reflecting the fluctuation trend of the navigation voice keyword after more synonyms are identified, and the weight of each synonym is different according to the occurrence frequency in navigation).
According to the technical scheme, in order to supervise the fluctuation state of the word quantity in the continuous updating process and judge whether the navigation voice keywords are continuously increased, the state of the navigation voice keywords is judged according to the continuously increased result, and whether synonyms of each keyword are increased is judged, and the synonyms are only in the updated state when being increased;
acquiring updated navigation voice keywords, establishing times based on navigation errors and times based on navigation correctness in the INR navigation process, and generating an updated word navigation result;
establishing a quality model based on word quantity updating according to the word quantity fluctuation model and the updated word navigation result:
f represents the result of the quality evaluation; α represents the correct number of times navigation was performed using the updated synonym during navigation at ivr; γ represents the number of errors in navigation using the updated synonym during navigation at ivr; delta 1 And delta 2 Representing the number of original navigation speech keywords in the word volume trend chartThe first included angle value of the origin and the second included angle value of the updated number of the navigation voice keywords relative to the origin;
and judging whether the updating of the navigation voice keywords is successful or not according to the quality model and a preset updating detection time period. When the quality model obtained value is increased in each updating detection time period, the updating success is indicated, and when the quality model obtained value is decreased in any detection time period, the updating failure is indicated.
The technical scheme mainly aims to judge whether the updating result improves the quality of IVR navigation in the process of updating the keywords so as to achieve the aim of continuously optimizing the navigation voice keywords.
Example 10:
preferably, the method further comprises:
calling a mobile phone number and customer information of a customer according to the front-end robot;
determining a business database of the client according to the client information;
judging whether the same service order exists in the service database according to the service database;
when the same service order exists, automatically classifying the customer requirements, automatically broadcasting a multi-level voice navigation menu for unclear description of customer information, and enabling options in the navigation of the customer voice selection menu to enter a problem processing or requirement handling processing system;
when the same service order does not exist, judging whether a corresponding navigation layer exists or not;
when the corresponding navigation layer exists, performing service processing according to the corresponding navigation layer;
and when the corresponding navigation layer does not exist, jumping to a manual processing end for manual processing.
The working principle of the technical scheme is as follows: when the voice processing of the client is carried out, the invention mainly uses the front-end robot, so that the invention can finally realize automatic processing for the same service when order processing is carried out, and can realize manual processing for new tasks which are never executed by the client.
The beneficial effects of the above technical scheme are: the scheme is also a flow saving mode, the same service can realize a processing mode for saving the flow more, and the corresponding flow can be entered more quickly after the processing, but for the user requirements which are not processed, the user requirements can be automatically processed through manual processing when meeting again without manual work.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (10)
1. A method of multi-tier IVR navigation processing, comprising:
receiving customer voice through a front-end processor, and identifying user requirements according to the customer voice;
when the user requirement is successfully identified, directionally jumping to a corresponding navigation layer;
and when the user requirement identification fails, sending the client voice to a manual processing end, controlling the jump of a navigation layer through the manual processing end, and updating the directional jump after the jump.
2. The method of multi-tier IVR navigation processing as claimed in claim 1, the method further comprising:
receiving a communication request of a user, and initializing a voice storage area of the front-end robot;
after the initialization, connecting the communication request and carrying out audio recording;
determining a mixed waveform of the audio according to the audio recording;
performing signal source separation on the mixed waveform to determine a plurality of audio signal sources;
determining the fundamental frequency of each signal source according to the signal sources, comparing the fundamental frequency with the mixed audio of the mixed waveform, and determining the contrast difference;
according to the contrast difference, deleting the noise audio frequency of the signal source with the difference reaching a preset threshold value;
and after the noise audio is deleted, audio splicing is carried out to generate new spliced voice.
3. The method of multi-tier IVR navigation processing as claimed in claim 2, the method further comprising:
performing textualization on the spliced voice to determine a voice text;
according to the voice text, carrying out requirement item keyword calibration and determining requirement keywords;
according to the requirement keywords, carrying out requirement classification and determining a requirement type;
judging whether a corresponding requirement keyword exists in a requirement database or not according to the requirement type;
when the requirement keyword exists, determining the requirement of the user;
and when the requirement key words do not exist, sending the spliced voice to a manual processing end.
4. The method of multi-tier IVR navigation processing as claimed in claim 3, the method further comprising:
pre-configuring a training set voice and a test set voice, and determining a training text of the training set voice and a test text of the test set voice;
acquiring a first acoustic feature of the training set voice, and determining a first keyword of the training text in the training set voice;
acquiring a second acoustic feature of the test set voice, and determining a second keyword of the test text in the test set voice;
establishing a deep network model based on four branches, and leading the first acoustic feature, the second acoustic feature, the first keyword and the second keyword into the deep network model of the four branches together for training in a fusion training mode to generate a text conversion model;
and performing textual conversion on the spliced voice through the text conversion model.
5. The method of multi-tier IVR navigation processing as claimed in claim 1, the method further comprising:
acquiring all navigation processing services of IVR navigation and generating a navigation set;
generating a navigation list according to the navigation set;
establishing a keyword library of navigation service according to the navigation list;
establishing a contrast triggering mechanism in the keyword library, stacking the navigation services in each navigation list layer by layer to generate a multilayer navigation menu, and triggering each navigation service in the multilayer navigation menu through the contrast triggering mechanism.
6. The method of multi-tier IVR navigation processing as claimed in claim 1, the method further comprising:
acquiring navigation information corresponding to each navigation layer in the navigation layers;
determining a corresponding service address of each navigation service end according to the navigation information;
performing service skipping according to the service address;
triggering the service end to record the user again according to the service jump;
forming a user order according to the user record;
and determining an order implementation flow according to the user order, and performing service processing according to the order implementation flow.
7. The method of multi-tier IVR navigation processing as claimed in claim 6, the method further comprising:
acquiring the user order and determining a business execution process; wherein the content of the first and second substances,
responding to the business execution flow, sending a processing request for order execution to a supply server, and receiving an execution node corresponding to the business execution flow returned by the supply server; wherein the content of the first and second substances,
the provisioning server is further configured to determine an ID for each executing node;
initiating execution verification for the business execution flow to each execution node execution rule based on the ID of the execution node;
after receiving an execution verification result returned by each execution node, updating a service execution state in target order information associated with a user based on user information contained in the execution verification;
and sending the user order information after the service execution state is updated to the supply server so as to obtain an order completion response authorized by the supply server based on the order information after the service execution state is updated.
8. The method of multi-tier IVR navigation processing as claimed in claim 1, the method further comprising:
receiving a communication request of a user, and carrying out an IVR flow configuration request according to the communication request;
obtaining configured IVR flow configuration information;
generating an IVR configuration page according to the configured IVR flow configuration information, wherein the IVR configuration page comprises configuration areas of all IVR branch flows, and the configuration area of each IVR branch flow comprises a node tree of the IVR branch flow;
receiving editing information in the IVR configuration page, and updating the IVR process configuration information according to the editing information;
obtaining IVR flow configuration information, wherein the IVR flow configuration information comprises execution conditions of all IVR branch flows, nodes included in all IVR branch flows and execution logics of all nodes in all IVR branch flows;
selecting an IVR branch flow meeting execution conditions according to the IVR flow configuration information and the incoming call request as an IVR branch flow to be executed;
and sequentially executing IVR instructions corresponding to the nodes according to the execution logic of the nodes in the IVR branch flow to be executed.
9. The method of multi-tier IVR navigation processing as claimed in claim 1, the method further comprising:
acquiring navigation information of a manual processing end;
determining a corresponding navigation page and artificial voice keywords according to the navigation information;
storing a navigation page and artificial voice keywords for updating the navigation layer;
determining original navigation voice keywords according to the navigation page;
matching the original navigation voice keywords with the keywords of the artificial voice keywords;
and updating the original navigation voice keywords according to the keyword matching.
10. The method of multi-tier IVR navigation processing as claimed in claim 1, the method further comprising:
calling a mobile phone number and customer information of a customer according to the front-end robot;
determining a business database of the client according to the client information;
judging whether the same service order exists in the service database according to the service database;
when the same service order exists, automatically classifying the customer requirements, automatically broadcasting a multi-level voice navigation menu for unclear description of customer information, and enabling options in the navigation of the customer voice selection menu to enter a problem processing or requirement handling processing system;
when the same service order does not exist, judging whether a corresponding navigation layer exists or not;
when the corresponding navigation layer exists, performing service processing according to the corresponding navigation layer;
and when the corresponding navigation layer does not exist, jumping to a manual processing end for manual processing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210398533.4A CN114844992A (en) | 2022-04-15 | 2022-04-15 | Multilayer IVR navigation processing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210398533.4A CN114844992A (en) | 2022-04-15 | 2022-04-15 | Multilayer IVR navigation processing method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114844992A true CN114844992A (en) | 2022-08-02 |
Family
ID=82566247
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210398533.4A Pending CN114844992A (en) | 2022-04-15 | 2022-04-15 | Multilayer IVR navigation processing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114844992A (en) |
-
2022
- 2022-04-15 CN CN202210398533.4A patent/CN114844992A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107146612B (en) | Voice guidance method and device, intelligent equipment and server | |
| CN107038220B (en) | Method, intelligent robot and system for generating memorandum | |
| US8010343B2 (en) | Disambiguation systems and methods for use in generating grammars | |
| US9117453B2 (en) | Method and system for processing parallel context dependent speech recognition results from a single utterance utilizing a context database | |
| EP1602102B1 (en) | Management of conversations | |
| RU2391791C2 (en) | Machine training | |
| US7907705B1 (en) | Speech to text for assisted form completion | |
| CN104299623B (en) | It is used to automatically confirm that the method and system with disambiguation module in voice application | |
| US8064573B2 (en) | Computer generated prompting | |
| US6839671B2 (en) | Learning of dialogue states and language model of spoken information system | |
| US9978365B2 (en) | Method and system for providing a voice interface | |
| RU2383938C2 (en) | Improved calling subscriber identification based on speech recognition | |
| US8144838B2 (en) | Automated task classification system | |
| US20040260543A1 (en) | Pattern cross-matching | |
| CN108962233A (en) | Voice dialogue processing method and system for voice dialogue platform | |
| CN103714813A (en) | Phrase spotting systems and methods | |
| JP2021099874A (en) | Public scenario conversation backtracking method, apparatus, electronic device, recording medium, and computer program in multi-turn conversation | |
| EP3888081B1 (en) | Spoken language interface for network management | |
| CN112487810B (en) | Intelligent customer service method, device, equipment and storage medium | |
| CN115509485A (en) | Filling-in method and device of business form, electronic equipment and storage medium | |
| CN115148212A (en) | Voice interaction method, intelligent device and system | |
| US7460999B2 (en) | Method and apparatus for executing tasks in voice-activated command systems | |
| CN114844992A (en) | Multilayer IVR navigation processing method | |
| CN114328867A (en) | Intelligent interruption method and device in man-machine conversation | |
| CN114546862A (en) | Dialogue test method and device, electronic equipment and computer readable 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 |