CN114242109A - Intelligent outbound method and device based on emotion recognition, electronic equipment and medium - Google Patents
Intelligent outbound method and device based on emotion recognition, electronic equipment and medium Download PDFInfo
- Publication number
- CN114242109A CN114242109A CN202111549523.8A CN202111549523A CN114242109A CN 114242109 A CN114242109 A CN 114242109A CN 202111549523 A CN202111549523 A CN 202111549523A CN 114242109 A CN114242109 A CN 114242109A
- Authority
- CN
- China
- Prior art keywords
- emotion
- outbound
- voice
- user
- voice data
- 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
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/27—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique
- G10L25/30—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique using neural networks
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
- G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
- G10L25/63—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination for estimating an emotional state
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/78—Detection of presence or absence of voice signals
- G10L25/87—Detection of discrete points within a voice signal
-
- 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
- H04M3/527—Centralised call answering arrangements not requiring operator intervention
Abstract
The invention relates to the technical field of artificial intelligence, and provides an intelligent outbound method, an intelligent outbound device, electronic equipment and a medium based on emotion recognition, wherein the method comprises the following steps: generating a dialing script in response to the received outbound command; activating a target telephone in the dialing script to automatically dial until the target telephone is connected; acquiring first voice data of a user from the recorded voice data of the target telephone every other preset period, and preprocessing the first voice data to obtain second voice data; inputting the second voice data into a pre-trained voice emotion classification model to obtain a voice emotion recognition result of the outbound user; and executing the outbound task based on the speech emotion recognition result of the outbound user. The invention executes the outbound task through the speech emotion recognition result of the outbound user, thereby improving the efficiency and accuracy of intelligent outbound. In addition, the invention also relates to the technical field of block chains, and the dialing script is stored in the block chain nodes.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to an intelligent outbound method, an intelligent outbound device, electronic equipment and a medium based on emotion recognition.
Background
With the development of artificial intelligence technology, the robot outbound system is widely applied to various service industries. In the prior art, when a product is recommended by a robot outbound system, emotional colors are lacked, the subjective emotion of a user cannot be judged in the process of recommending the product, and a product recommendation strategy is adjusted according to the subjective emotion of the user, so that a client is easy to generate aversive emotion when the product is recommended by the robot outbound system, and the recommendation efficiency and accuracy of the recommended product are low.
Disclosure of Invention
In view of the above, it is necessary to provide an intelligent outbound method, apparatus, electronic device and medium based on emotion recognition, which perform an outbound task by using a speech emotion recognition result of an outbound user, and improve efficiency and accuracy of intelligent outbound.
The first aspect of the invention provides an intelligent outbound method based on emotion recognition, which comprises the following steps:
responding to a received outbound command, and generating a dialing script, wherein the outbound command comprises an outbound task;
activating the target telephone in the dialing script to automatically dial until the target telephone is connected;
acquiring first voice data of a user from the recorded voice data of the target telephone every other preset period, and preprocessing the first voice data to obtain second voice data;
inputting the second voice data into a pre-trained voice emotion classification model to obtain a voice emotion recognition result of the outbound user;
and executing the outbound task based on the speech emotion recognition result of the outbound user.
Optionally, the preprocessing the first voice data to obtain second voice data includes:
detecting a silence segment in the first voice data;
cutting the first voice data according to the mute segments to obtain a plurality of first voice subdata;
sequencing each first voice subdata according to low frequency to high frequency;
after sorting, arranging a group of band-pass filters according to the size of critical bandwidth and a preset rule;
the band-pass filter filters each first voice subdata;
and obtaining the filtered multiple first voice subdata, and determining the filtered multiple first voice subdata as second voice data.
Optionally, before the inputting the second speech data into the pre-trained speech emotion classification model, the method further includes:
acquiring a plurality of emotion categories and a first voice sample set corresponding to each emotion category;
extracting a logarithmic Mel frequency spectrum diagram of each first voice sample in a first voice sample set corresponding to each emotion category;
segmenting the logarithmic Mel frequency spectrogram of each first voice sample in the first voice sample set corresponding to each emotion category according to the distribution characteristics of the sample sets of the emotion categories, and adjusting the sample data size of the segmented first voice sample set corresponding to each emotion category to obtain a second voice sample set of each emotion category;
adjusting model parameters in a pre-trained convolutional neural network model based on the second voice sample set of each emotion category to obtain an adjusted convolutional neural network model;
inputting a plurality of second voice sample sets of a plurality of emotion classes into the adjusted convolutional neural network model to obtain a segmented feature set;
dividing a training set and a testing set from the segmented feature set;
inputting the training set into a preset long-time and short-time memory network for training to obtain a speech emotion classification model;
inputting the test set into the speech emotion classification model for testing, and calculating a test passing rate;
if the test passing rate is larger than or equal to a preset passing rate threshold value, determining that the speech emotion classification model training is finished; and if the test passing rate is smaller than the preset passing rate threshold, increasing the number of training sets, and re-training the speech emotion classification model.
Optionally, the executing the outbound task based on the speech emotion recognition result of the outbound user includes:
when the voice emotion recognition result of the outbound user is positive emotion, executing an outbound task corresponding to the outbound user; or
And when the voice emotion recognition result of the outbound user is a negative emotion, performing emotion placation on the outbound user.
Optionally, when the voice emotion recognition result of the outbound user is a negative emotion, the emotion soothing to the outbound user comprises:
performing text conversion on the first voice data of the user to obtain a dialog text;
matching the conversation text by adopting a preset deep text matching method to obtain emotion subject keywords of the user;
determining a corresponding emotion placating strategy and a target knowledge base according to the emotion theme keywords of the user;
matching a placating tactic text from the target knowledge base randomly according to the emotion placating tactic, wherein the emotion placating tactic comprises the following steps: an emotion placating policy without a business scene and a business theme attribute placating policy.
Optionally, the executing the outbound task based on the speech emotion recognition result of the outbound user includes:
analyzing first voice data of the user, and generating response content according to the first voice data;
recognizing emotion categories in the voice emotion recognition result;
and adjusting the dialogue intonation of the response content according to the emotion category, and executing the outbound task based on the dialogue intonation.
Optionally, the generating response content according to the first voice includes:
analyzing the first voice data, and acquiring a key field in the conversation content and a corresponding service scene;
and generating response content according to a preset conversation generation rule based on the key field and the service scene.
A second aspect of the present invention provides an intelligent outbound device based on emotion recognition, the device comprising:
the generation module is used for responding to the received outbound command and generating a dialing script, wherein the outbound command comprises an outbound task;
the activation module is used for activating the target telephone in the dialing script to automatically dial until the target telephone is connected;
the preprocessing module is used for acquiring first voice data of a user from the recorded voice data of the target telephone every other preset period, and preprocessing the first voice data to obtain second voice data;
the input module is used for inputting the second voice data into a pre-trained voice emotion classification model to obtain a voice emotion recognition result of the outbound user;
and the execution module is used for executing the outbound task based on the speech emotion recognition result of the outbound user.
A third aspect of the invention provides an electronic device comprising a processor and a memory, the processor being configured to implement the intelligent call-out method based on emotion recognition when executing a computer program stored in the memory.
A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the intelligent call-out method based on emotion recognition.
In summary, the intelligent outbound method, apparatus, electronic device and medium based on emotion recognition generate a dialing script in response to a received outbound command, activate a target telephone in the dialing script to perform automatic dialing, collect first voice data of a user from recorded voice data of the target telephone every preset period, and preprocess the first voice data, thereby ensuring accuracy of second voice data. And inputting the second voice data into a pre-trained voice emotion classification model to obtain a voice emotion recognition result of the outbound user, so that the efficiency of obtaining the voice emotion recognition result is improved. And executing the outbound task based on the speech emotion recognition result of the outbound user, and adopting different outbound strategies according to different speech emotion recognition results, so that the method is more targeted, and further improves the efficiency and the accuracy of intelligent outbound.
Drawings
Fig. 1 is a flowchart of an intelligent outbound method based on emotion recognition according to an embodiment of the present invention.
Fig. 2 is a structural diagram of an intelligent outbound device based on emotion recognition according to a second embodiment of the present invention.
Fig. 3 is a schematic structural diagram of an electronic device according to a third embodiment of the present invention.
Detailed Description
In order that the above objects, features and advantages of the present invention can be more clearly understood, a detailed description of the present invention will be given below with reference to the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present invention and features of the embodiments may be combined with each other without conflict.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.
Example one
Fig. 1 is a flowchart of an intelligent outbound method based on emotion recognition according to an embodiment of the present invention.
In this embodiment, the intelligent outbound method based on emotion recognition may be applied to an electronic device, and for an electronic device that needs to perform intelligent outbound based on emotion recognition, the function of the intelligent outbound based on emotion recognition provided by the method of the present invention may be directly integrated on the electronic device, or may be run in the electronic device in the form of a Software Development Kit (SDK).
As shown in FIG. 1, the intelligent calling-out method based on emotion recognition specifically includes the following steps, and the order of the steps in the flowchart can be changed and some steps can be omitted according to different requirements.
And S11, responding to the received outbound command, and generating a dialing script, wherein the outbound command comprises an outbound task.
In the embodiment, in the service industry, the robot outbound system is widely applied, and when the robot outbound system receives an outbound command, a dialing script is generated in response to the outbound command. The outbound task refers to a service required for the current outbound, for example, service recommendation or product recommendation is required for the current outbound user.
It is emphasized that, in order to further ensure the privacy and security of the dialing script, the dialing script may also be stored in a node of a block chain.
In an alternative embodiment, the generating a dialing script in response to the received outbound command includes:
responding to the received outbound command, and acquiring a target telephone;
identifying a chat script and a receiver equipment code of a dialing interface operator corresponding to the target telephone;
determining a block identification code of a block chain matched with the equipment code of the receiving party from preset block identification codes;
and generating a dialing script according to the dialing interface, the chat script of the operator and the block identification code of the block chain.
In this embodiment, each target telephone corresponds to a dialing interface, the chat script of the operator is used to represent the dialing script of the operator, each receiver device code corresponds to a block identification code of a block chain, and the dialing script can be generated according to the dialing interface, the chat script of the operator and the block identification code of the block chain, so that the dialing script is generated in a targeted manner, and the accuracy of the dialing script is ensured.
And S12, activating the target telephone in the dialing script to automatically dial until the target telephone is connected.
In the embodiment, when the target telephone in the dialing script is detected to be activated, the target telephone is automatically and circularly dialed until the target telephone is connected, manual dialing is not needed, meanwhile, the phenomenon of error in manual dialing is avoided, the dialing efficiency is improved, and further, the efficiency and the accuracy of intelligent outbound are improved.
And S13, collecting first voice data of the user from the recorded voice data of the target telephone every other preset period, and preprocessing the first voice data to obtain second voice data.
In this embodiment, the preset collection period may be set to solve the problem that the emotion type of the user changes during the outbound call execution process and cannot be found in time, and the current state of the user is grasped in time by collecting the first voice data of the user from the recorded voice data of the target telephone every other preset period.
In an optional embodiment, the preprocessing the first voice data to obtain the second voice data includes:
detecting a silence segment in the first voice data;
cutting the first voice data according to the mute segments to obtain a plurality of first voice subdata;
sequencing each first voice subdata according to low frequency to high frequency;
after sorting, arranging a group of band-pass filters according to the size of critical bandwidth and a preset rule;
the band-pass filter filters each first voice subdata;
and obtaining the filtered multiple first voice subdata, and determining the filtered multiple first voice subdata as second voice data.
In this embodiment, in order to improve the efficiency of preprocessing the first voice data, a voice activity detection is performed on the first voice data by using a detection algorithm of VAD, a silence segment in the first voice data is detected, the silence segment is cut to obtain a plurality of first voice sub-data, a complete sentence is cut, interference of silence is reduced, and the accuracy of the obtained plurality of first voice sub-data is ensured.
In this embodiment, the first voice data may be cut using a library of librosa audio processing.
In this embodiment, an arrangement rule may be preset, for example, a group of band pass filters is arranged according to an arrangement rule from dense to sparse according to the size of the critical bandwidth after the sorting, when a call is made with a user, the environment background where the user is located is different, and particularly when the user makes a call in an environment with a large environment background noise, the recognition rate of the first speech sub data is affected. Therefore, in order to prevent the components with frequencies exceeding f/2 in the input signal of the input first voice sub-data from causing mixing and power frequency interference of 50HZ, the plurality of first voice sub-data are pre-filtered by a band-pass filter, the plurality of pre-filtered first voice sub-data are obtained, and the plurality of first voice sub-data are determined as second voice data.
The pre-filtering by the band-pass filter is prior art, and the invention will not be described in detail here.
In this embodiment, the first voice data is cut into the plurality of first sub-voice data for filtering processing by performing VAD detection and filtering processing on the first voice data, so that the speed of filtering processing is increased, the processing efficiency of the first voice data is improved, the accuracy of the obtained second voice data is ensured, and the intelligent outbound efficiency is further improved.
And S14, inputting the second voice data into a pre-trained voice emotion classification model to obtain a voice emotion recognition result of the calling user.
In this embodiment, a plurality of emotion categories may be preset, each emotion category corresponding to a plurality of key features, for example, when the emotion category is positive emotion, the corresponding key features include happy, curious, surprised, relaxed, and attentive; when the emotion classification is a negative emotion, the corresponding key features comprise sadness, repugnance, anger, fear and anger, the emotion features of the user can be identified through the voice data, and the emotion classification of the user is determined based on the emotion features.
In this embodiment, the speech emotion classification model includes a Convolutional Neural network and a long-time memory unit, and in order to solve the problem that the emotion classification model obtained by training is low in recognition efficiency due to unbalanced distribution of a sample set in the existing speech emotion recognition and the application problem of a conventional speech emotion classification model in a complex speech scene during speech emotion recognition, in this embodiment, a Convolutional Neural Network (CNN) and a long-time memory unit (LSTM) are combined to perform speech emotion recognition on a plurality of speech features, so that the problem of unbalanced distribution of the sample set is solved, and the accuracy of speech emotion recognition results is further improved.
In this embodiment, the speech emotion classification model is trained in advance before the second speech data is input into the pre-trained speech emotion classification model.
Specifically, the training process of the speech emotion classification model comprises the following steps:
acquiring a plurality of emotion categories and a first voice sample set corresponding to each emotion category;
extracting a logarithmic Mel frequency spectrum diagram of each first voice sample in a first voice sample set corresponding to each emotion category;
segmenting the logarithmic Mel frequency spectrogram of each first voice sample in the first voice sample set corresponding to each emotion category according to the distribution characteristics of the sample sets of the emotion categories, and adjusting the sample data size of the segmented first voice sample set corresponding to each emotion category to obtain a second voice sample set of each emotion category;
adjusting model parameters in a pre-trained convolutional neural network model based on the second voice sample set of each emotion category to obtain an adjusted convolutional neural network model;
inputting a plurality of second voice sample sets of a plurality of emotion classes into the adjusted convolutional neural network model to obtain a segmented feature set;
dividing a training set and a testing set from the segmented feature set;
inputting the training set into a preset long-time and short-time memory network for training to obtain a speech emotion classification model;
inputting the test set into the speech emotion classification model for testing, and calculating a test passing rate;
if the test passing rate is larger than or equal to a preset passing rate threshold value, determining that the speech emotion classification model training is finished; and if the test passing rate is smaller than the preset passing rate threshold, increasing the number of training sets, and re-training the speech emotion classification model.
In this embodiment, in a subsequent service process, the first voice data of the user collected in each period is used as a new sample, the new sample is added to the first voice sample set, the new voice sample set is analyzed to obtain a segmented feature set, and the speech emotion classification model is retrained based on the new segmented feature set. Namely, the speech emotion classification model is continuously updated, so that the accuracy of speech emotion classification is continuously improved.
S15, executing the calling task based on the voice emotion recognition result of the calling user.
In this embodiment, the speech emotion recognition result includes other emotion types such as a positive emotion or a negative emotion.
In an optional embodiment, the performing the outbound task based on the speech emotion recognition result of the outbound user comprises:
when the voice emotion recognition result of the outbound user is positive emotion, executing an outbound task corresponding to the outbound user; or
And when the voice emotion recognition result of the outbound user is a negative emotion, performing emotion placation on the outbound user.
In this embodiment, after the recognized speech emotion recognition result, the robot classifies the speech emotion recognition result according to content attributes of different emotion categories, and performs emotion soothing on the user, where different types of soothing correspond to different knowledge bases respectively, and the knowledge bases include standard response content in different service scenes. When the user is in the negative emotion, a soothing strategy is adopted for the user instead of directly recommending products, and when the fact that the emotion of the user is converted from the negative emotion to the positive emotion is detected, the user is subjected to an outbound task, so that on one hand, the experience degree of the user can be improved, and on the other hand, the accuracy rate of the outbound can be improved.
In an optional embodiment, when the voice emotion recognition result of the calling-out user is a negative emotion, the emotion soothing to the calling-out user comprises:
performing text conversion on the first voice data of the user to obtain a dialog text;
matching the conversation text by adopting a preset deep text matching method to obtain emotion subject keywords of the user;
determining a corresponding emotion placating strategy and a target knowledge base according to the emotion theme keywords of the user;
and matching a placating tactic text from the target knowledge base randomly according to the emotion placating strategy.
Specifically, the emotion placating strategy comprises the following steps: an emotion placating policy without a business scene and a business theme attribute placating policy.
Illustratively, without the emotion placating strategy of the business scene, when the user emotion topic keyword is "good and uncomfortable", the emotion placating strategy adopted by the robot is as follows: how are they about? What is not happy can be said to me.
Illustratively, the service theme attribute appeasing strategy is that when the user emotion theme keyword is "do not need, i are busy", the emotion appeasing strategy adopted by the robot is as follows: it is not good meaning and disturbing.
In the embodiment, when the voice emotion recognition result of the outbound user is a negative emotion, the outbound user is emotionally pacified instead of directly executing the outbound task, so that the phenomenon that conversation is directly refused due to the negative emotion is avoided, and the outbound efficiency and accuracy are improved.
In other optional embodiments, the performing the outbound task based on the speech emotion recognition result of the outbound user includes:
analyzing first voice data of the user, and generating response content according to the first voice data;
recognizing emotion categories in the voice emotion recognition result;
and adjusting the dialogue intonation of the response content according to the emotion category, and executing the outbound task based on the dialogue intonation.
In the embodiment, in the outbound conversation process, in order to enable the response content and the intonation to have the emotional color, the intonation and the emotional color of the robot conversation are adjusted according to the recognized emotion of the user, specifically, the first voice data (conversation content) of the user is analyzed, and the response content of the robot is generated, wherein the generation of the response content depends on the conversation content in the common service scene, the conversation intonation of the response content is adjusted according to the emotion type of the user, the response emotional color of the robot is given, the conversation with the emotional color is performed, and the outbound efficiency and the experience of the user are improved.
Further, the generating response content according to the first voice comprises:
analyzing the first voice data, and acquiring a key field in the conversation content and a corresponding service scene;
and generating response content according to a preset conversation generation rule based on the key field and the service scene.
In this embodiment, the first voice data of the user is analyzed, the obtained key field is interested in the heavy insurance, the corresponding business scenario is for purchasing insurance products, and the response content is generated for the user according to the historical dialogue data for purchasing the heavy insurance and the preset dialogue generation rule, wherein the preset dialogue generation rule is set according to the historical dialogue data generation rule.
In summary, in the intelligent outbound method based on emotion recognition according to this embodiment, after outbound dialing is performed, first voice data of a user is collected from recorded voice data of a target telephone every preset period, and the first voice data is preprocessed, so that accuracy of second voice data is ensured. And inputting the second voice data into a pre-trained voice emotion classification model to obtain a voice emotion recognition result of the outbound user, so that the efficiency of obtaining the voice emotion recognition result is improved. And executing the outbound task based on the speech emotion recognition result of the outbound user, and adopting different outbound strategies according to different speech emotion recognition results, so that the method is more targeted, and further improves the efficiency and the accuracy of intelligent outbound.
Example two
Fig. 2 is a structural diagram of an intelligent outbound device based on emotion recognition according to a second embodiment of the present invention.
In some embodiments, the emotion recognition based intelligent outbound device 20 may include a plurality of functional modules comprised of program code segments. The program code of each program segment of the emotion recognition based intelligent callout apparatus 20 can be stored in the memory of the electronic device and executed by the at least one processor to perform (see fig. 1 for details) the emotion recognition based intelligent callout function.
In this embodiment, the intelligent outbound device 20 based on emotion recognition may be divided into a plurality of functional modules according to the functions it performs. The functional module may include: a generation module 201, an activation module 202, a pre-processing module 203, an input module 204, and an execution module 205. The module referred to herein is a series of computer readable instruction segments stored in a memory that can be executed by at least one processor and that can perform a fixed function. In the present embodiment, the functions of the modules will be described in detail in the following embodiments.
The generating module 201 is configured to generate a dialing script in response to a received outbound command, where the outbound command includes an outbound task.
In the embodiment, in the service industry, the robot outbound system is widely applied, and when the robot outbound system receives an outbound command, a dialing script is generated in response to the outbound command. The outbound task refers to a service required for the current outbound, for example, service recommendation or product recommendation is required for the current outbound user.
In an alternative embodiment, the generating module 201, in response to the received outbound command, generates the dialing script including:
responding to the received outbound command, and acquiring a target telephone;
identifying a chat script and a receiver equipment code of a dialing interface operator corresponding to the target telephone;
determining a block identification code of a block chain matched with the equipment code of the receiving party from preset block identification codes;
and generating a dialing script according to the dialing interface, the chat script of the operator and the block identification code of the block chain.
In this embodiment, each target telephone corresponds to a dialing interface, the chat script of the operator is used to represent the dialing script of the operator, each receiver device code corresponds to a block identification code of a block chain, and the dialing script can be generated according to the dialing interface, the chat script of the operator and the block identification code of the block chain, so that the dialing script is generated in a targeted manner, and the accuracy of the dialing script is ensured.
It is emphasized that, in order to further ensure the privacy and security of the dialing script, the dialing script may also be stored in a node of a block chain.
And the activating module 202 is configured to activate the target telephone in the dialing script to perform automatic dialing until the target telephone is connected.
In the embodiment, when the target telephone in the dialing script is detected to be activated, the target telephone is automatically and circularly dialed until the target telephone is connected, manual dialing is not needed, meanwhile, the phenomenon of error in manual dialing is avoided, the dialing efficiency is improved, and further, the efficiency and the accuracy of intelligent outbound are improved.
The preprocessing module 203 is configured to collect first voice data of a user from the recorded voice data of the target telephone every preset period, and preprocess the first voice data to obtain second voice data.
In this embodiment, the preset collection period may be set to solve the problem that the emotion type of the user changes during the outbound call execution process and cannot be found in time, and the current state of the user is grasped in time by collecting the first voice data of the user from the recorded voice data of the target telephone every other preset period.
In an optional embodiment, the preprocessing module 203 performs preprocessing on the first voice data to obtain second voice data, and includes:
detecting a silence segment in the first voice data;
cutting the first voice data according to the mute segments to obtain a plurality of first voice subdata;
sequencing each first voice subdata according to low frequency to high frequency;
after sorting, arranging a group of band-pass filters according to the size of critical bandwidth and a preset rule;
the band-pass filter filters each first voice subdata;
and obtaining the filtered multiple first voice subdata, and determining the filtered multiple first voice subdata as second voice data.
In this embodiment, in order to improve the efficiency of preprocessing the first voice data, a voice activity detection is performed on the first voice data by using a detection algorithm of VAD, a silence segment in the first voice data is detected, the silence segment is cut to obtain a plurality of first voice sub-data, a complete sentence is cut, interference of silence is reduced, and the accuracy of the obtained plurality of first voice sub-data is ensured.
In this embodiment, the first voice data may be cut using a library of librosa audio processing.
In this embodiment, an arrangement rule may be preset, for example, a group of band pass filters is arranged according to an arrangement rule from dense to sparse according to the size of the critical bandwidth after the sorting, when a call is made with a user, the environment background where the user is located is different, and particularly when the user makes a call in an environment with a large environment background noise, the recognition rate of the first speech sub data is affected. Therefore, in order to prevent the components with frequencies exceeding f/2 in the input signal of the input first voice sub-data from causing mixing and power frequency interference of 50HZ, the plurality of first voice sub-data are pre-filtered by a band-pass filter, the plurality of pre-filtered first voice sub-data are obtained, and the plurality of first voice sub-data are determined as second voice data.
The pre-filtering by the band-pass filter is prior art, and the invention will not be described in detail here.
In this embodiment, the first voice data is cut into the plurality of first sub-voice data for filtering processing by performing VAD detection and filtering processing on the first voice data, so that the speed of filtering processing is increased, the processing efficiency of the first voice data is improved, the accuracy of the obtained second voice data is ensured, and the intelligent outbound efficiency is further improved.
And the input module 204 is configured to input the second speech data into a pre-trained speech emotion classification model to obtain a speech emotion recognition result of the outbound user.
In this embodiment, a plurality of emotion categories may be preset, each emotion category corresponding to a plurality of key features, for example, when the emotion category is positive emotion, the corresponding key features include happy, curious, surprised, relaxed, and attentive; when the emotion classification is a negative emotion, the corresponding key features comprise sadness, repugnance, anger, fear and anger, the emotion features of the user can be identified through the voice data, and the emotion classification of the user is determined based on the emotion features.
In this embodiment, the speech emotion classification model includes a Convolutional Neural network and a long-time memory unit, and in order to solve the problem that the emotion classification model obtained by training is low in recognition efficiency due to unbalanced distribution of a sample set in the existing speech emotion recognition and the application problem of a conventional speech emotion classification model in a complex speech scene during speech emotion recognition, in this embodiment, a Convolutional Neural Network (CNN) and a long-time memory unit (LSTM) are combined to perform speech emotion recognition on a plurality of speech features, so that the problem of unbalanced distribution of the sample set is solved, and the accuracy of speech emotion recognition results is further improved.
In this embodiment, the speech emotion classification model is trained in advance before the second speech data is input into the pre-trained speech emotion classification model.
Specifically, the training process of the speech emotion classification model comprises the following steps:
acquiring a plurality of emotion categories and a first voice sample set corresponding to each emotion category;
extracting a logarithmic Mel frequency spectrum diagram of each first voice sample in a first voice sample set corresponding to each emotion category;
segmenting the logarithmic Mel frequency spectrogram of each first voice sample in the first voice sample set corresponding to each emotion category according to the distribution characteristics of the sample sets of the emotion categories, and adjusting the sample data size of the segmented first voice sample set corresponding to each emotion category to obtain a second voice sample set of each emotion category;
adjusting model parameters in a pre-trained convolutional neural network model based on the second voice sample set of each emotion category to obtain an adjusted convolutional neural network model;
inputting a plurality of second voice sample sets of a plurality of emotion classes into the adjusted convolutional neural network model to obtain a segmented feature set;
dividing a training set and a testing set from the segmented feature set;
inputting the training set into a preset long-time and short-time memory network for training to obtain a speech emotion classification model;
inputting the test set into the speech emotion classification model for testing, and calculating a test passing rate;
if the test passing rate is larger than or equal to a preset passing rate threshold value, determining that the speech emotion classification model training is finished; and if the test passing rate is smaller than the preset passing rate threshold, increasing the number of training sets, and re-training the speech emotion classification model.
In this embodiment, in a subsequent service process, the first voice data of the user collected in each period is used as a new sample, the new sample is added to the first voice sample set, the new voice sample set is analyzed to obtain a segmented feature set, and the speech emotion classification model is retrained based on the new segmented feature set. Namely, the speech emotion classification model is continuously updated, so that the accuracy of speech emotion classification is continuously improved.
And the execution module 205 is used for executing the outbound task based on the speech emotion recognition result of the outbound user.
In this embodiment, the speech emotion recognition result includes other emotion types such as a positive emotion or a negative emotion.
In an alternative embodiment, the executing module 205 executing the outbound task based on the speech emotion recognition result of the outbound user includes:
when the voice emotion recognition result of the outbound user is positive emotion, executing an outbound task corresponding to the outbound user; or
And when the voice emotion recognition result of the outbound user is a negative emotion, performing emotion placation on the outbound user.
In this embodiment, after the recognized speech emotion recognition result, the robot classifies the speech emotion recognition result according to content attributes of different emotion categories, and performs emotion soothing on the user, where different types of soothing correspond to different knowledge bases respectively, and the knowledge bases include standard response content in different service scenes. When the user is in the negative emotion, a soothing strategy is adopted for the user instead of directly recommending products, and when the fact that the emotion of the user is converted from the negative emotion to the positive emotion is detected, the user is subjected to an outbound task, so that on one hand, the experience degree of the user can be improved, and on the other hand, the accuracy rate of the outbound can be improved.
In an optional embodiment, when the voice emotion recognition result of the calling-out user is a negative emotion, the emotion soothing to the calling-out user comprises:
performing text conversion on the first voice data of the user to obtain a dialog text;
matching the conversation text by adopting a preset deep text matching method to obtain emotion subject keywords of the user;
determining a corresponding emotion placating strategy and a target knowledge base according to the emotion theme keywords of the user;
and matching a placating tactic text from the target knowledge base randomly according to the emotion placating strategy.
Specifically, the emotion placating strategy comprises the following steps: an emotion placating policy without a business scene and a business theme attribute placating policy.
Illustratively, without the emotion placating strategy of the business scene, when the user emotion topic keyword is "good and uncomfortable", the emotion placating strategy adopted by the robot is as follows: how are they about? What is not happy can be said to me.
Illustratively, the service theme attribute appeasing strategy is that when the user emotion theme keyword is "do not need, i are busy", the emotion appeasing strategy adopted by the robot is as follows: it is not good meaning and disturbing.
In the embodiment, when the voice emotion recognition result of the outbound user is a negative emotion, the outbound user is emotionally pacified instead of directly executing the outbound task, so that the phenomenon that conversation is directly refused due to the negative emotion is avoided, and the outbound efficiency and accuracy are improved.
In other optional embodiments, the executing module 205, executing the outbound task based on the speech emotion recognition result of the outbound user, includes:
analyzing first voice data of the user, and generating response content according to the first voice data;
recognizing emotion categories in the voice emotion recognition result;
and adjusting the dialogue intonation of the response content according to the emotion category, and executing the outbound task based on the dialogue intonation.
In the embodiment, in the outbound conversation process, in order to enable the response content and the intonation to have the emotional color, the intonation and the emotional color of the robot conversation are adjusted according to the recognized emotion of the user, specifically, the first voice data (conversation content) of the user is analyzed, and the response content of the robot is generated, wherein the generation of the response content depends on the conversation content in the common service scene, the conversation intonation of the response content is adjusted according to the emotion type of the user, the response emotional color of the robot is given, the conversation with the emotional color is performed, and the outbound efficiency and the experience of the user are improved.
Further, the generating response content according to the first voice comprises:
analyzing the first voice data, and acquiring a key field in the conversation content and a corresponding service scene;
and generating response content according to a preset conversation generation rule based on the key field and the service scene.
In this embodiment, the first voice data of the user is analyzed, the obtained key field is interested in the heavy insurance, the corresponding business scenario is for purchasing insurance products, and the response content is generated for the user according to the historical dialogue data for purchasing the heavy insurance and the preset dialogue generation rule, wherein the preset dialogue generation rule is set according to the historical dialogue data generation rule.
In summary, the intelligent outbound device based on emotion recognition in this embodiment collects the first voice data of the user from the recorded voice data of the target phone every preset period after outbound dialing, preprocesses the first voice data, and ensures the accuracy of the second voice data. And inputting the second voice data into a pre-trained voice emotion classification model to obtain a voice emotion recognition result of the outbound user, so that the efficiency of obtaining the voice emotion recognition result is improved. And executing the outbound task based on the speech emotion recognition result of the outbound user, and adopting different outbound strategies according to different speech emotion recognition results, so that the method is more targeted, and further improves the efficiency and the accuracy of intelligent outbound.
EXAMPLE III
Fig. 3 is a schematic structural diagram of an electronic device according to a third embodiment of the present invention. In the preferred embodiment of the present invention, the electronic device 3 comprises a memory 31, at least one processor 32, at least one communication bus 33 and a transceiver 34.
It will be appreciated by those skilled in the art that the configuration of the electronic device shown in fig. 3 does not constitute a limitation of the embodiment of the present invention, and may be a bus-type configuration or a star-type configuration, and the electronic device 3 may include more or less other hardware or software than those shown, or a different arrangement of components.
In some embodiments, the electronic device 3 is an electronic device capable of automatically performing numerical calculation and/or information processing according to instructions set or stored in advance, and the hardware thereof includes but is not limited to a microprocessor, an application specific integrated circuit, a programmable gate array, a digital processor, an embedded device, and the like. The electronic device 3 may also include a client device, which includes, but is not limited to, any electronic product that can interact with a client through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device, for example, a personal computer, a tablet computer, a smart phone, a digital camera, and the like.
It should be noted that the electronic device 3 is only an example, and other existing or future electronic products, such as those that can be adapted to the present invention, should also be included in the scope of the present invention, and are included herein by reference.
In some embodiments, the memory 31 is used for storing program codes and various data, such as the intelligent calling-out device 20 based on emotion recognition installed in the electronic equipment 3, and realizes high-speed and automatic access to programs or data during the operation of the electronic equipment 3. The Memory 31 includes a Read-Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Programmable Read-Only Memory (EPROM), a One-time Programmable Read-Only Memory (OTPROM), an electronically Erasable rewritable Read-Only Memory (Electrically-Erasable Programmable Read-Only Memory (EEPROM)), an optical Read-Only disk (CD-ROM) or other optical disk Memory, a magnetic disk Memory, a tape Memory, or any other medium readable by a computer capable of carrying or storing data.
In some embodiments, the at least one processor 32 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or may be composed of a plurality of integrated circuits packaged with the same or different functions, including one or more Central Processing Units (CPUs), microprocessors, digital Processing chips, graphics processors, and combinations of various control chips. The at least one processor 32 is a Control Unit (Control Unit) of the electronic device 3, connects various components of the electronic device 3 by using various interfaces and lines, and executes various functions and processes data of the electronic device 3 by running or executing programs or modules stored in the memory 31 and calling data stored in the memory 31.
In some embodiments, the at least one communication bus 33 is arranged to enable connection communication between the memory 31 and the at least one processor 32 or the like.
Although not shown, the electronic device 3 may further include a power supply (such as a battery) for supplying power to each component, and optionally, the power supply may be logically connected to the at least one processor 32 through a power management device, so as to implement functions of managing charging, discharging, and power consumption through the power management device. The power supply may also include any component of one or more dc or ac power sources, recharging devices, power failure detection circuitry, power converters or inverters, power status indicators, and the like. The electronic device 3 may further include various sensors, a bluetooth module, a Wi-Fi module, and the like, which are not described herein again.
It is to be understood that the described embodiments are for purposes of illustration only and that the scope of the appended claims is not limited to such structures.
The integrated unit implemented in the form of a software functional module may be stored in a computer-readable storage medium. The software functional module is stored in a storage medium and includes several instructions to enable a computer device (which may be a personal computer, an electronic device, or a network device) or a processor (processor) to execute parts of the methods according to the embodiments of the present invention.
In a further embodiment, in conjunction with fig. 2, the at least one processor 32 may execute the operating device of the electronic device 3 and various installed applications (e.g. the emotion recognition based intelligent calling out device 20), program codes, etc., such as the above modules.
The memory 31 has program code stored therein, and the at least one processor 32 can call the program code stored in the memory 31 to perform related functions. For example, the modules illustrated in fig. 2 are program code stored in the memory 31 and executed by the at least one processor 32 to implement the functions of the modules for the purpose of intelligent callouts based on emotion recognition.
Illustratively, the program code may be partitioned into one or more modules/units that are stored in the memory 31 and executed by the processor 32 to accomplish the present application. The one or more modules/units may be a series of computer readable instruction segments capable of performing certain functions, which are used for describing the execution process of the program code in the electronic device 3. For example, the program code may be partitioned into a generation module 201, an activation module 202, a pre-processing module 203, an input module 204, and an execution module 205.
In one embodiment of the present invention, the memory 31 stores a plurality of computer readable instructions that are executed by the at least one processor 32 to implement the functionality of intelligent callouts based on emotion recognition.
Specifically, the at least one processor 32 may refer to the description of the relevant steps in the embodiment corresponding to fig. 1, and details are not repeated here.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is only one logical functional division, and other divisions may be realized in practice.
Further, the computer-readable storage medium may be non-volatile or volatile.
Further, the computer-readable storage medium mainly includes a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to the use of the blockchain node, and the like.
The block chain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, a consensus mechanism, an encryption algorithm and the like. A block chain (Blockchain), which is essentially a decentralized database, is a series of data blocks associated by using a cryptographic method, and each data block contains information of a batch of network transactions, so as to verify the validity (anti-counterfeiting) of the information and generate a next block. The blockchain may include a blockchain underlying platform, a platform product service layer, an application service layer, and the like.
The modules described as separate parts may or may not be physically separate, and parts displayed as modules may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.
In addition, functional modules in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional module.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned. Furthermore, it is obvious that the word "comprising" does not exclude other elements or that the singular does not exclude the plural. A plurality of units or means recited in the present invention may also be implemented by one unit or means through software or hardware. The terms first, second, etc. are used to denote names, but not any particular order.
Finally, it should be noted that the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.
Claims (10)
1. An intelligent outbound method based on emotion recognition, the method comprising:
responding to a received outbound command, and generating a dialing script, wherein the outbound command comprises an outbound task;
activating the target telephone in the dialing script to automatically dial until the target telephone is connected;
acquiring first voice data of a user from the recorded voice data of the target telephone every other preset period, and preprocessing the first voice data to obtain second voice data;
inputting the second voice data into a pre-trained voice emotion classification model to obtain a voice emotion recognition result of the outbound user;
and executing the outbound task based on the speech emotion recognition result of the outbound user.
2. The intelligent callout method based on emotion recognition as recited in claim 1, wherein said preprocessing the first speech data to obtain second speech data comprises:
detecting a silence segment in the first voice data;
cutting the first voice data according to the mute segments to obtain a plurality of first voice subdata;
sequencing each first voice subdata according to low frequency to high frequency;
after sorting, arranging a group of band-pass filters according to the size of critical bandwidth and a preset rule;
the band-pass filter filters each first voice subdata;
and obtaining the filtered multiple first voice subdata, and determining the filtered multiple first voice subdata as second voice data.
3. The intelligent emotion recognition-based callout method of claim 1, wherein prior to the inputting of the second speech data into a pre-trained speech emotion classification model, the method further comprises:
acquiring a plurality of emotion categories and a first voice sample set corresponding to each emotion category;
extracting a logarithmic Mel frequency spectrum diagram of each first voice sample in a first voice sample set corresponding to each emotion category;
segmenting the logarithmic Mel frequency spectrogram of each first voice sample in the first voice sample set corresponding to each emotion category according to the distribution characteristics of the sample sets of the emotion categories, and adjusting the sample data size of the segmented first voice sample set corresponding to each emotion category to obtain a second voice sample set of each emotion category;
adjusting model parameters in a pre-trained convolutional neural network model based on the second voice sample set of each emotion category to obtain an adjusted convolutional neural network model;
inputting a plurality of second voice sample sets of a plurality of emotion classes into the adjusted convolutional neural network model to obtain a segmented feature set;
dividing a training set and a testing set from the segmented feature set;
inputting the training set into a preset long-time and short-time memory network for training to obtain a speech emotion classification model;
inputting the test set into the speech emotion classification model for testing, and calculating a test passing rate;
if the test passing rate is larger than or equal to a preset passing rate threshold value, determining that the speech emotion classification model training is finished; and if the test passing rate is smaller than the preset passing rate threshold, increasing the number of training sets, and re-training the speech emotion classification model.
4. The intelligent outbound method based on emotion recognition as recited in claim 1, wherein said performing the outbound task based on the voice emotion recognition result of the outbound user comprises:
when the voice emotion recognition result of the outbound user is positive emotion, executing an outbound task corresponding to the outbound user; or
And when the voice emotion recognition result of the outbound user is a negative emotion, performing emotion placation on the outbound user.
5. The intelligent outbound method based on emotion recognition as claimed in claim 4, wherein said emotion placating the outbound user when the voice emotion recognition result of the outbound user is a negative emotion comprises:
performing text conversion on the first voice data of the user to obtain a dialog text;
matching the conversation text by adopting a preset deep text matching method to obtain emotion subject keywords of the user;
determining a corresponding emotion placating strategy and a target knowledge base according to the emotion theme keywords of the user;
matching a placating tactic text from the target knowledge base randomly according to the emotion placating tactic, wherein the emotion placating tactic comprises the following steps: an emotion placating policy without a business scene and a business theme attribute placating policy.
6. The intelligent outbound method based on emotion recognition as recited in claim 1, wherein said performing the outbound task based on the voice emotion recognition result of the outbound user comprises:
analyzing first voice data of the user, and generating response content according to the first voice data;
recognizing emotion categories in the voice emotion recognition result;
and adjusting the dialogue intonation of the response content according to the emotion category, and executing the outbound task based on the dialogue intonation.
7. The intelligent emotion recognition-based callout method of claim 6, wherein said generating response content from the first speech comprises:
analyzing the first voice data, and acquiring a key field in the conversation content and a corresponding service scene;
and generating response content according to a preset conversation generation rule based on the key field and the service scene.
8. An intelligent outbound device based on emotion recognition, the device comprising:
the generation module is used for responding to the received outbound command and generating a dialing script, wherein the outbound command comprises an outbound task;
the activation module is used for activating the target telephone in the dialing script to automatically dial until the target telephone is connected;
the preprocessing module is used for acquiring first voice data of a user from the recorded voice data of the target telephone every other preset period, and preprocessing the first voice data to obtain second voice data;
the input module is used for inputting the second voice data into a pre-trained voice emotion classification model to obtain a voice emotion recognition result of the outbound user;
and the execution module is used for executing the outbound task based on the speech emotion recognition result of the outbound user.
9. An electronic device, characterized in that the electronic device comprises a processor and a memory, the processor being configured to implement the intelligent call-out method based on emotion recognition according to any of claims 1 to 7 when executing the computer program stored in the memory.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, implements the intelligent call-out method based on emotion recognition according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111549523.8A CN114242109A (en) | 2021-12-17 | 2021-12-17 | Intelligent outbound method and device based on emotion recognition, electronic equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111549523.8A CN114242109A (en) | 2021-12-17 | 2021-12-17 | Intelligent outbound method and device based on emotion recognition, electronic equipment and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114242109A true CN114242109A (en) | 2022-03-25 |
Family
ID=80757958
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111549523.8A Pending CN114242109A (en) | 2021-12-17 | 2021-12-17 | Intelligent outbound method and device based on emotion recognition, electronic equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114242109A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116389644A (en) * | 2022-11-10 | 2023-07-04 | 八度云计算(安徽)有限公司 | Outbound system based on big data analysis |
| CN117714603A (en) * | 2024-02-01 | 2024-03-15 | 济南云上电子科技有限公司 | Outbound method, outbound device and readable storage medium |
-
2021
- 2021-12-17 CN CN202111549523.8A patent/CN114242109A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116389644A (en) * | 2022-11-10 | 2023-07-04 | 八度云计算(安徽)有限公司 | Outbound system based on big data analysis |
| CN116389644B (en) * | 2022-11-10 | 2023-11-03 | 八度云计算(安徽)有限公司 | Outbound system based on big data analysis |
| CN117714603A (en) * | 2024-02-01 | 2024-03-15 | 济南云上电子科技有限公司 | Outbound method, outbound device and readable storage medium |
| CN117714603B (en) * | 2024-02-01 | 2024-04-30 | 济南云上电子科技有限公司 | Outbound method, outbound device and readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10642889B2 (en) | Unsupervised automated topic detection, segmentation and labeling of conversations | |
| CN114242109A (en) | Intelligent outbound method and device based on emotion recognition, electronic equipment and medium | |
| CN110147726A (en) | Business quality detecting method and device, storage medium and electronic device | |
| CN106407178A (en) | Session abstract generation method and device | |
| CN111179935B (en) | Voice quality inspection method and device | |
| CN112001175B (en) | Flow automation method, device, electronic equipment and storage medium | |
| CN105991593B (en) | A kind of method and device identifying consumer's risk | |
| CN109977202A (en) | A kind of intelligent customer service system and its control method | |
| CN112468659B (en) | Quality evaluation method, device, equipment and storage medium applied to telephone customer service | |
| CN113129927B (en) | Voice emotion recognition method, device, equipment and storage medium | |
| US11133022B2 (en) | Method and device for audio recognition using sample audio and a voting matrix | |
| CN109977225A (en) | The analysis of public opinion method and device | |
| CN113436634A (en) | Voice classification method and device based on voiceprint recognition and related equipment | |
| CN111178701B (en) | Risk control method and device based on feature derivation technology and electronic equipment | |
| Noroozi et al. | Speech-based emotion recognition and next reaction prediction | |
| CN112632248A (en) | Question answering method, device, computer equipment and storage medium | |
| Rajan et al. | Conflictnet: End-to-end learning for speech-based conflict intensity estimation | |
| CN114078472A (en) | Training method and device for keyword calculation model with low false awakening rate | |
| CN113590771A (en) | Data mining method, device, equipment and storage medium | |
| Śniegula et al. | Study of machine learning methods for customer churn prediction in telecommunication company | |
| US10674952B1 (en) | Detection and management of memory impairment | |
| CN113241063B (en) | Algorithm parameter updating method, device, terminal and medium in voice recognition system | |
| CN113438374B (en) | Intelligent outbound call processing method, device, equipment and storage medium | |
| CN113221990A (en) | Information input method and device and related equipment | |
| CN113421590A (en) | Abnormal behavior detection method, device, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |