WO2020253362A1 - 基于情绪分析的业务处理方法、装置、设备及存储介质 - Google Patents

基于情绪分析的业务处理方法、装置、设备及存储介质 Download PDF

Info

Publication number
WO2020253362A1
WO2020253362A1 PCT/CN2020/086161 CN2020086161W WO2020253362A1 WO 2020253362 A1 WO2020253362 A1 WO 2020253362A1 CN 2020086161 W CN2020086161 W CN 2020086161W WO 2020253362 A1 WO2020253362 A1 WO 2020253362A1
Authority
WO
WIPO (PCT)
Prior art keywords
target
information
preset
voice information
target user
Prior art date
Application number
PCT/CN2020/086161
Other languages
English (en)
French (fr)
Inventor
陈娟
傅婧
黄忆丁
张鹭
钱尼丽
Original Assignee
深圳壹账通智能科技有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 深圳壹账通智能科技有限公司 filed Critical 深圳壹账通智能科技有限公司
Publication of WO2020253362A1 publication Critical patent/WO2020253362A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33Querying
    • G06F16/332Query formulation
    • G06F16/3329Natural language query formulation or dialogue systems
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33Querying
    • G06F16/3331Query processing
    • G06F16/334Query execution
    • G06F16/3343Query execution using phonetics
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33Querying
    • G06F16/3331Query processing
    • G06F16/334Query execution
    • G06F16/3344Query execution using natural language analysis
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35Clustering; Classification
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00Handling natural language data
    • G06F40/20Natural language analysis
    • G06F40/205Parsing
    • G06F40/211Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00Handling natural language data
    • G06F40/20Natural language analysis
    • G06F40/279Recognition of textual entities
    • G06F40/289Phrasal analysis, e.g. finite state techniques or chunking
    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00Speech recognition
    • G10L15/26Speech to text systems
    • GPHYSICS
    • G10MUSICAL INSTRUMENTS; ACOUSTICS
    • G10LSPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/48Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
    • G10L25/51Speech 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
    • G10L25/63Speech 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

Definitions

  • This application relates to the field of artificial intelligence, and in particular to a business processing method, device, equipment and computer-readable storage medium based on sentiment analysis.
  • the main purpose of this application is to provide a business processing method, device, equipment, and computer-readable storage medium based on sentiment analysis, aiming to solve the technical problems of low self-service terminal business processing efficiency and poor user experience.
  • the business processing method based on sentiment analysis includes the following steps:
  • the terminal Upon receiving the service handling request sent by the terminal, obtain the target voice information of the target user, convert the target voice information into corresponding text information, and determine the target voice information based on the preset emotion recognition model, the target voice information and the text information. State the emotional category of the target user;
  • the present application also provides a business processing device based on sentiment analysis, and the business processing device based on sentiment analysis includes:
  • the emotion determination module is used to obtain the target voice information of the target user when receiving the service handling request sent by the terminal, and convert the target voice information into corresponding text information, based on the preset emotion recognition model and the target voice information And text information to determine the emotional category of the target user;
  • a speech technique determining module configured to determine the business processing flow and target service speech technique corresponding to the target user based on the text information and the emotion category;
  • the processing guidance module is used to broadcast the target service language and guide the target user to complete the corresponding target service according to the service processing flow.
  • the present application also provides a business processing device based on sentiment analysis.
  • the business processing device based on sentiment analysis includes a processor, a memory, and stored in the memory and can be used by the processor.
  • the executed business processing program based on sentiment analysis wherein when the business processing program based on sentiment analysis is executed by the processor, the steps of the above-mentioned business processing method based on sentiment analysis are implemented.
  • the present application also provides a computer-readable storage medium that stores a business processing program based on sentiment analysis, wherein the business processing program based on sentiment analysis is processed by the processor When executed, the steps of the business processing method based on sentiment analysis as described above are implemented.
  • This application provides a service processing method based on sentiment analysis, that is, when a service processing request sent by a terminal is received, target voice information of a target user is obtained, and the target voice information is converted into corresponding text information, based on preset emotions Recognition model, the target voice information and text information, determine the emotional category of the target user; based on the text information and the emotional category, determine the business processing flow and target service words corresponding to the target user; The target service language is described, and the target user is guided to complete the corresponding target service according to the business processing flow.
  • this application determines the user’s emotional category based on the user’s voice information, then determines the corresponding business processing flow and corresponding target service language based on the emotional category, and then guides the target based on the business processing flow and target service language
  • the user completes the corresponding target business, improves business processing efficiency, improves user experience, and solves the technical problems of low efficiency of existing self-service terminal business processing and poor user experience.
  • FIG. 1 is a schematic diagram of the hardware structure of a business processing device based on sentiment analysis involved in the solution of an embodiment of the application;
  • FIG. 2 is a schematic flowchart of a first embodiment of a business processing method based on sentiment analysis in this application;
  • FIG. 3 is a schematic flowchart of a second embodiment of a business processing method based on sentiment analysis according to this application;
  • FIG. 4 is a schematic flowchart of a third embodiment of a business processing method based on sentiment analysis of this application;
  • FIG. 5 is a schematic diagram of functional modules of a first embodiment of a business processing apparatus based on sentiment analysis according to this application.
  • the business processing method based on sentiment analysis involved in the embodiments of the present application is mainly applied to business processing equipment based on sentiment analysis.
  • the business processing equipment based on sentiment analysis may be a PC, a portable computer, a mobile terminal, or other devices with display and processing functions.
  • FIG. 1 is a schematic diagram of the hardware structure of a service processing device based on sentiment analysis involved in the solution of an embodiment of the application.
  • the service processing device based on sentiment analysis may include a processor 1001 (for example, a CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005.
  • the communication bus 1002 is used to realize the connection and communication between these components;
  • the user interface 1003 may include a display (Display), an input unit such as a keyboard (Keyboard);
  • the network interface 1004 may optionally include a standard wired interface, a wireless interface (Such as WI-FI interface);
  • the memory 1005 can be a high-speed RAM memory or a non-volatile memory, such as a disk memory.
  • the memory 1005 can optionally be a storage device independent of the aforementioned processor 1001 .
  • FIG. 1 does not constitute a limitation on the business processing equipment based on sentiment analysis, and may include more or less components than shown in the figure, or combine certain components, or different The layout of the components.
  • the memory 1005 as a computer-readable storage medium in FIG. 1 may include an operating system, a network communication module, and a business processing program based on sentiment analysis.
  • the network communication module is mainly used to connect to the server and perform data communication with the server; and the processor 1001 can call the business processing program based on sentiment analysis stored in the memory 1005, and execute the sentiment analysis based on the embodiment of this application. Business processing methods.
  • the embodiment of the application provides a business processing method based on sentiment analysis.
  • the business processing method based on sentiment analysis includes the following steps:
  • Step S10 When receiving the service processing request sent by the terminal, obtain the target voice information of the target user, and convert the target voice information into corresponding text information, based on a preset emotion recognition model, the target voice information and the text information To determine the emotional category of the target user;
  • the client triggers a service handling request on the terminal
  • the terminal sends the service handling request to the server
  • the server responds to the service handling request, that is, when the server receives the service handling request, the server sends a preset initial voice to the terminal
  • the preset initial voice refers to the pre-set recommended voice.
  • the preset initial voice is: Dear customer, welcome to the xxx platform, is there anything I can serve you?
  • the terminal receives the preset initial voice sent by the server and plays it.
  • the terminal collects the customer's voice information through the preset voice collection device as target voice information.
  • the preset voice collection device refers to a preset voice collection device.
  • the preset voice collection device may be a recording device on the terminal, and the terminal sends the collected target voice information for the preset initial voice to the server. Receive the target voice information sent by the terminal to handle business according to the target voice information.
  • the step of converting the target voice information into corresponding text information includes: inputting the target voice information into a preset voice recognition model, and performing target voice recognition through the preset voice recognition model to obtain The text information corresponding to the target voice information.
  • the target voice information input by the user may also be pre-processed, and the pre-processing includes: word segmentation, part-of-speech tagging, named entity recognition, reference disambiguation, and similar word expansion.
  • the word segmentation, part-of-speech tagging, and named entity recognition refer to word segmentation, part-of-speech tagging, and named entity recognition using a natural language processing tool trained by a deep neural network.
  • the word segmentation refers to segmenting the sequence of Chinese characters into word sequences.
  • the part-of-speech tagging refers to the identification and tagging based on the part-of-speech of words.
  • the part-of-speech includes: nouns, adverbs, adjectives, verbs, pronouns, etc., for example, NT stands for time nouns, V stands for verbs, NN stands for spoken nouns, PU stands for calibration symbols, AD stands for adverbs, PN stands for pronouns, etc.
  • the named entity recognition refers to a named entity such as a person's name, a place name, and an organization's name in the recognition sentence.
  • the named entity includes 3 categories, such as entity, time, and number, and 7 categories, such as person's name, place name, and organization name , Time, date, currency and percentage.
  • the reference disambiguation refers to the elimination of the reference ambiguity of the personal pronouns, which is accomplished by using dependency parsing (dependency parsing, DP) to identify the dependency relationship between the components in the language unit and reveal its syntactic structure.
  • dependency parsing dependency parsing
  • the expansion of similar sentences refers to the expansion of similar words using Word2vec technology.
  • the emotion categories include: anger, anxiety, anger, happiness, disappointment, surprise, curiosity, etc.
  • the emotion category of the user may also be identified based on the emotion analysis method of the dictionary and the pre-trained emotion classification model based on deep learning.
  • the dictionary-based sentiment analysis method refers to the construction of an sentiment analysis dictionary by professionals with grammatical sensitivity. According to the sentiment analysis dictionary constructed: positive sentiment dictionary, negative sentiment dictionary and neutral sentiment dictionary, a sentence is used The vocabulary expressing emotion is divided into three categories, and then the number of positive, negative and neutral emotion words in the sentence is compared to evaluate the emotion category of the sentence.
  • the emotion classification models based on deep learning include, but are not limited to, Long Short-Term Memory (LSTM) models, Support Vector Machines (SVM) models, Random Forests (RF) models, and A pre-trained model in the Naive Bayesian Model (NBM) model.
  • LSTM Long Short-Term Memory
  • SVM Support Vector Machines
  • RF Random Forests
  • NBM Naive Bayesian Model
  • the training step of the preset emotion recognition model is: obtaining preset training voice information and corresponding emotion category information, and training the preset training voice information through a neural network model skip-gram or a continuous bag of words model , Get the low-dimensional word vector; by looking up the word vector table, the low-dimensional word vector is converted into a corresponding vector, and the vector expression is converted into a feature vector through convolution and pooling operations.
  • the vector uses the Hard Tanh function to extract nonlinear features to obtain the final features of the preset training voice information;
  • the preset emotion recognition model is generated according to the final feature of the preset training voice information and the corresponding emotion category information.
  • Step S20 based on the text information and the emotion category, determine the business processing flow and target service words corresponding to the target user;
  • a library of service words corresponding to different emotion categories can be established in advance.
  • the service words corresponding to the angry emotion category are simple and friendly words.
  • the business you need to handle is X.
  • Two processes are required. The first process is X and the second process is X. Please check whether the above information is correct.
  • the service language corresponding to the anxious emotion category is a comfort and guidance type. In order to prevent network fraud or phone fraud, you can also broadcast anti-fraud promotion information.
  • different business processing procedures can also be set according to different emotional categories, such as recommending the most efficient business processing procedures with the least time-consuming or the least business process nodes for users in the anxious and angry emotional categories, for happy and joyful users It is recommended to add a business handling process with wealth management product recommendation or other related business introductions to increase the sales volume of wealth management products.
  • the server determines the target business that the target user needs to handle according to the text information, such as transfer, loan, bank card or credit card, etc. Then, according to the emotional category of the target user, the target service language and business processing flow suitable for serving the target user are determined.
  • Step S30 Broadcast the target service language, and guide the target user to complete the corresponding target service according to the service processing flow.
  • each operation node corresponding to the business processing flow displays the corresponding interface to be operated, and simultaneously plays the corresponding target service words, such as "place the ID card in the front ID card recognition area".
  • Guide the target user to handle the target business through the target service language that conforms to the current emotional category of the target user. That is, the target service language and the business processing flow corresponding to the target user are determined, and the guidance information of each process step corresponding to the business processing flow is correspondingly added to the target service language.
  • the business types with the same business process that is, the same business types with the same process guidance information, can use the same set of service words, and then replace the specific business name or business number.
  • the current payee information input by the target user may also be compared with the historical payee information corresponding to the target user to determine whether the current payee is a frequent contact of the target user for transactions. If it is not, then broadcast the anti-fraud warning information. If there are frequent fraud crimes recently and the current payee is not a frequent contact person, please confirm again whether to perform the current transfer operation, etc.
  • the background color of the current operation interface can also be displayed according to the emotional category of the target user, so as to further enhance the user experience.
  • This embodiment provides a service processing method based on sentiment analysis, that is, when a service processing request sent by a terminal is received, target voice information of a target user is obtained, and the target voice information is converted into corresponding text information, based on a preset Emotion recognition model, the target voice information and text information, determine the emotion category of the target user; based on the text information and the emotion category, determine the business processing flow and target service words corresponding to the target user; broadcast
  • the target service language is used to guide the target user to complete the corresponding target service according to the business processing flow.
  • this application determines the user’s emotional category based on the user’s voice information, then determines the corresponding business processing flow and corresponding target service language based on the emotional category, and then guides the target based on the business processing flow and target service language
  • the user completes the corresponding target business, improves business processing efficiency, improves user experience, and solves the technical problems of low efficiency of existing self-service terminal business processing and poor user experience.
  • Fig. 3 is a schematic flowchart of a second embodiment of a business processing method based on sentiment analysis in this application.
  • the step S30 includes:
  • Step S31 Broadcast the target service speech, and receive the feedback voice information of the target user based on the target service speech feedback, and input the feedback speech information into a preset speech recognition model, and pass the A preset voice recognition model obtains feedback text information corresponding to the feedback voice information;
  • the target user while playing the target service speech, the target user’s feedback voice information based on the target service speech is acquired in real time, and the feedback speech information is converted into corresponding feedback through a preset speech recognition model.
  • Text information is: separately collecting the voice data of the preset dialect language system and the common language system, and extracting the voice feature parameters corresponding to each of the voice data, and composing each of the voice feature parameters into a voice feature Set; extract each of the voice feature parameters in the voice feature set in a preset proportion, and construct an initial voice recognition model through each of the voice feature parameters; train the initial voice recognition model through an iterative algorithm, and obtain the training result
  • the speech recognition model whose recognition accuracy rate is higher than the preset threshold is used as the preset speech recognition model.
  • Step S32 According to the feedback text information, determine whether the business processing flow is the business processing flow requested by the target user;
  • the feedback text information it is determined whether the currently determined target service language and business processing flow are the target service language corresponding to the target business requested by the target user and the business processing flow corresponding to the target business. If the feedback text information is "No", “Exit”, etc., that is, the guidance information of the business processing process broadcast by the target service voice is not the business process corresponding to the target business that the target user needs to handle. If the feedback text information is "confirmation”, "thank you”, etc., that is, the business process of the target service speech broadcast is the business process corresponding to the target business that the target user needs to handle.
  • step S33 if it is not the business processing flow requested by the target user, re-determine the business processing flow corresponding to the target business to be processed and the corresponding service skills according to the feedback voice information.
  • the feedback voice information entered by the target user can be re-based, if it is not the feedback voice information for handling transfers or savings , Re-execute the operation of determining the target service words and the business processing flow corresponding to the savings business.
  • the feedback voice information of the target user is an exit instruction, the voice information of the target user is re-entered to perform the business processing flow corresponding to the target service and the determination operation of the target service language.
  • FIG. 4 is a schematic flowchart of a third embodiment of a business processing method based on sentiment analysis in this application.
  • the step S10 specifically includes:
  • Step S11 when receiving the service processing request sent by the terminal, obtain the face information of the target user, and judge whether the age of the target user is greater than a preset threshold according to the face information;
  • the target user in order to prevent middle-aged and elderly people from being unable to use self-service terminals to cause business processing obstacles, when it is detected that the age of the target user exceeds a preset threshold, such as 50 years old, manual assistance can be provided to the target user.
  • a preset threshold such as 50 years old
  • manual assistance can be provided to the target user.
  • the face information of the target user is acquired through a camera, and it is determined whether the target user corresponding to the face information is a user requiring assistance, that is, the person requiring assistance is a user whose age exceeds a preset threshold.
  • Step S12 If the age of the target user is greater than the preset threshold, generate assistance request information according to the terminal identification of the current service terminal, and send the assistance request information to the staff, so as to remind the relevant staff of the target user Provide manual guidance;
  • the terminal identifier of the current service terminal that the target user handles the service is generated to carry the The assistance request information of the terminal identification. If the self-service terminal No. 3 is currently waiting to assist the user in business processing, please go to assist as soon as possible, and send the assistance request information to the relevant terminal of the staff, such as the terminal corresponding to the lobby manager. In this way, relevant staff are reminded in time to help some older middle-aged and elderly people to handle business with autonomous terminals in time to improve business handling efficiency.
  • Step S13 If the age of the target user is not greater than a preset threshold, obtain target voice information of the target user, and convert the target voice information into corresponding text information, based on the preset emotion recognition model and the target voice information And text information to determine the emotional category of the target user.
  • the user’s emotional category is determined according to the user’s voice information, and then the corresponding business processing flow and corresponding target service language are determined according to the emotional category, and then processed according to the business
  • the process and target service words guide the target user to complete the corresponding target business.
  • step S10 it further includes:
  • emergency request information is generated according to the terminal identification of the current service terminal, and the emergency request information is sent to the staff, so as to remind the relevant staff to provide the target user with Quick assistance.
  • the target user is also possible to provide quick assistance to users in emergency situations, such as providing an emergency service window for users whose bank cards are stolen, so as to quickly handle the business of freezing bank cards for the target users. That is, when it is determined that the emotional category of the target user is an anxious emotional category, the emergency request information is sent to the relevant staff according to the terminal identification of the current business terminal where the target user handles the business, so that the relevant staff can go to confirm and manually check. And when it is determined that the business that the target user really needs to handle is indeed an emergency business, the target user is provided with quick assistance to help the target user quickly complete the target business.
  • the embodiment of the present application also provides a business processing device based on sentiment analysis.
  • FIG. 5 is a schematic diagram of the functional modules of the first embodiment of the business processing apparatus based on sentiment analysis in this application.
  • the business processing device based on sentiment analysis includes:
  • the emotion determination module 10 is configured to obtain target voice information of a target user when receiving a service handling request sent by a terminal, and convert the target voice information into corresponding text information, based on a preset emotion recognition model and the target voice Information and text information to determine the emotional category of the target user;
  • the speech technique determining module 20 is configured to determine the business processing flow and target service speech technique corresponding to the target user based on the text information and the emotion category;
  • the processing guidance module 30 is configured to broadcast the target service language, and guide the target user to complete the corresponding target service according to the service processing flow.
  • the business processing device based on sentiment analysis further includes:
  • the information training module is used to obtain preset training voice information and corresponding emotional category information, and train preset training voice information through neural network model skip-gram or continuous word bag model to obtain low-dimensional word vectors;
  • the feature extraction module is used to convert a low-dimensional word vector into a corresponding vector by looking up a word vector table, convert the vector expression into a feature vector through convolution and pooling operations, and use the feature vector
  • the Hard Tanh function extracts nonlinear features to obtain the final features of the preset training voice information
  • the model generation module is configured to generate the preset emotion recognition model according to the final feature of the preset training voice information and the corresponding emotion category information.
  • emotion determination module 10 is also used for:
  • the target voice information is input into a preset voice recognition model, and the target voice recognition is performed through the preset voice recognition model to obtain text information corresponding to the target voice information.
  • emotion determination module 10 is also used for:
  • the initial speech recognition model is trained by an iterative algorithm, and the speech recognition accuracy of the trained speech recognition model is obtained, and the speech recognition model with the recognition accuracy higher than a preset threshold is used as the preset speech recognition model.
  • processing guidance module 30 specifically includes:
  • the feedback information acquiring unit is configured to broadcast the target service speech, and receive the feedback voice information of the target user based on the target service speech feedback, and input the feedback speech information into a preset speech recognition model, And obtain the feedback text information corresponding to the feedback voice information through the preset voice recognition model;
  • the business process verification unit is configured to determine whether the business process is the business process requested by the target user according to the feedback text information
  • the business process determining unit is configured to, if it is not the business process process requested by the target user, re-determine the business process process corresponding to the target business to be processed and the corresponding service technique according to the feedback voice information.
  • emotion determination module 10 specifically includes:
  • a user judging unit configured to obtain face information of the target user when receiving a service handling request sent by a terminal, and judge whether the age of the target user is greater than a preset threshold according to the face information
  • the assistance request unit is configured to generate assistance request information according to the terminal identification of the current service terminal if the age of the target user is greater than the preset threshold, and send the assistance request information to the staff, so as to remind relevant staff to The target user provides manual guidance;
  • emotion determination module 10 is also used for:
  • the target voice information of the target user If the age of the target user is not greater than the preset threshold, obtain target voice information of the target user, and convert the target voice information into corresponding text information, based on the preset emotion recognition model, the target voice information and the text information To determine the emotional category of the target user.
  • emotion determination module 10 is also used for:
  • emergency request information is generated according to the terminal identification of the current service terminal, and the emergency request information is sent to the staff, so as to remind the relevant staff to provide the target user with Quick assistance.
  • each module in the above-mentioned emotion analysis-based service processing device corresponds to each step in the above-mentioned embodiment of the above-mentioned emotion analysis-based service processing method, and its functions and implementation processes will not be repeated here.
  • the embodiment of the present application also provides a computer-readable storage medium.
  • a business processing program based on sentiment analysis is stored on the computer-readable storage medium of the present application.
  • the business processing program based on sentiment analysis is executed by a processor, the steps of the business processing method based on sentiment analysis as described above are implemented.
  • the computer-readable storage medium may be non-volatile or volatile.
  • the method implemented when the business processing program based on sentiment analysis is executed can refer to the various embodiments of the business processing method based on sentiment analysis of this application, which will not be repeated here.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • General Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Human Computer Interaction (AREA)
  • Acoustics & Sound (AREA)
  • Mathematical Physics (AREA)
  • Multimedia (AREA)
  • Child & Adolescent Psychology (AREA)
  • Hospice & Palliative Care (AREA)
  • Psychiatry (AREA)
  • Signal Processing (AREA)
  • Machine Translation (AREA)
  • Telephonic Communication Services (AREA)

Abstract

一种基于情绪分析的业务处理方法、装置、设备及存储介质,方法包括在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将目标语音信息转换为对应的文本信息,基于预设情绪识别模型、目标语音信息和文本信息,确定目标用户的情绪类别(S10);基于文本信息和情绪类别,确定目标用户对应的业务处理流程以及目标服务话术(S20);播报目标服务话术,并根据业务处理流程,指导目标用户完成对应的目标业务(S30)。根据用户语音信息确定用户的情绪类别,然后确定对应的业务处理流程和目标服务话术,然后指导所述目标用户完成对应的目标业务,提升业务办理效率,提升用户体验。

Description

基于情绪分析的业务处理方法、装置、设备及存储介质
本申请要求于2019年6月20日提交中国专利局、申请号为201910537349.1,发明名称为“基于情绪分析的业务处理方法、装置、设备及存储介质”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及人工智能领域,尤其涉及一种基于情绪分析的业务处理方法、装置、设备及计算机可读存储介质。
背景技术
随着互联网的迅速发展,人工智能应用领域不断扩大,人机对话应用也越来越广泛。传统的人机对话中,针对客户问题播报应答话术时,只是简单地进行语音语义的识别,即只是按照语音识别结果进行产品搜索,因此,发明人意识到产品搜索结果不一定准确,不仅降低业务办理效率,而且降低了用户体验。
因此,如何解决现有自助服务终端业务办理效率低下以及用户体验差的问题,是目前亟需解决的问题。
发明概述
技术问题
问题的解决方案
技术解决方案
本申请的主要目的在于提供一种基于情绪分析的业务处理方法、装置、设备及计算机可读存储介质,旨在解决自助服务终端业务办理效率低下以及用户体验差的技术问题。
为实现上述目的,本申请提供一种基于情绪分析的业务处理方法,所述基于情绪分析的业务处理方法包括以下步骤:
在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别;
基于所述文本信息和所述情绪类别,确定所述目标用户对应的业务处理流程以及目标服务话术;
播报所述目标服务话术,并根据所述业务处理流程,指导所述目标用户完成对应的目标业务。
此外,为实现上述目的,本申请还提供一种基于情绪分析的业务处理装置,所述基于情绪分析的业务处理装置包括:
情绪确定模块,用于在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别;
话术确定模块,用于基于所述文本信息和所述情绪类别,确定所述目标用户对应的业务处理流程以及目标服务话术;
处理指导模块,用于播报所述目标服务话术,并根据所述业务处理流程,指导所述目标用户完成对应的目标业务。
此外,为实现上述目的,本申请还提供一种基于情绪分析的业务处理设备,所述基于情绪分析的业务处理设备包括处理器、存储器、以及存储在所述存储器上并可被所述处理器执行的基于情绪分析的业务处理程序,其中所述基于情绪分析的业务处理程序被所述处理器执行时,实现如上述的基于情绪分析的业务处理方法的步骤。
此外,为实现上述目的,本申请还提供一种计算机可读存储介质,所述计算机可读存储介质上存储有基于情绪分析的业务处理程序,其中所述基于情绪分析的业务处理程序被处理器执行时,实现如上述的基于情绪分析的业务处理方法的步骤。
本申请提供一种基于情绪分析的业务处理方法,即在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别;基于所述文本信息和所述情绪类别,确定所述目标用户对应的业务处理流程以及目标服务话术;播报所述目标服务话术,并根据所述业务处理流程,指导所述目标用户完成对应的目标业务。通过上述方式,本申 请根据用户语音信息确定用户的情绪类别,然后根据情绪类别确定对应的业务处理流程和对应的目标服务话术,然后根据所述业务处理流程和目标服务话术指导所述目标用户完成对应的目标业务,提升业务办理效率,提升用户体验,解决了现有自助服务终端业务办理效率低下以及用户体验差的技术问题。
发明的有益效果
对附图的简要说明
附图说明
图1为本申请实施例方案中涉及的基于情绪分析的业务处理设备的硬件结构示意图;
图2为本申请基于情绪分析的业务处理方法第一实施例的流程示意图;
图3为本申请基于情绪分析的业务处理方法第二实施例的流程示意图;
图4为本申请基于情绪分析的业务处理方法第三实施例的流程示意图;
图5为本申请基于情绪分析的业务处理装置第一实施例的功能模块示意图。
本申请目的的实现、功能特点及优点将结合实施例,参照附图做进一步说明。
发明实施例
本发明的实施方式
应当理解,此处所描述的具体实施例仅仅用以解释本申请,并不用于限定本申请。
本申请实施例涉及的基于情绪分析的业务处理方法主要应用于基于情绪分析的业务处理设备,该基于情绪分析的业务处理设备可以是PC、便携计算机、移动终端等具有显示和处理功能的设备。
参照图1,图1为本申请实施例方案中涉及的基于情绪分析的业务处理设备的硬件结构示意图。本申请实施例中,基于情绪分析的业务处理设备可以包括处理器1001(例如CPU),通信总线1002,用户接口1003,网络接口1004,存储器1005。其中,通信总线1002用于实现这些组件之间的连接通信;用户接口1003可以包括显示屏(Display)、输入单元比如键盘(Keyboard);网络接口1004可选的可以包括标准的有线接口、无线接口(如WI-FI接口);存储器1005可以是 高速RAM存储器,也可以是稳定的存储器(non-volatile memory),例如磁盘存储器,存储器1005可选的还可以是独立于前述处理器1001的存储装置。
本领域技术人员可以理解,图1中示出的硬件结构并不构成对基于情绪分析的业务处理设备的限定,可以包括比图示更多或更少的部件,或者组合某些部件,或者不同的部件布置。
继续参照图1,图1中作为一种计算机可读存储介质的存储器1005可以包括操作系统、网络通信模块以及基于情绪分析的业务处理程序。
在图1中,网络通信模块主要用于连接服务器,与服务器进行数据通信;而处理器1001可以调用存储器1005中存储的基于情绪分析的业务处理程序,并执行本申请实施例提供的基于情绪分析的业务处理方法。
本申请实施例提供了一种基于情绪分析的业务处理方法。
参照图2,图2为本申请基于情绪分析的业务处理方法第一实施例的流程示意图。
本实施例中,所述基于情绪分析的业务处理方法包括以下步骤:
步骤S10,在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别;
本实施例中,客户在终端上触发业务办理请求,终端将业务办理请求发送至服务器,服务器对业务办理请求进行响应,即,服务器接收业务办理请求时,服务器向该终端发送预设初始语音,其中,预设初始语音是指预先设置的推荐语音,例如,预设初始语音为:尊敬的客户欢迎来到xxx平台,有什么我可以为您服务吗?终端接收服务器发送的预设初始语音并播放,在预设初始语音播报完成之后,终端通过预设语音采集装置采集客户的语音信息,作为目标语音信息。其中,预设语音采集装置是指预先设置的语音采集装置,例如,预设语音采集装置可以是终端上的录音装置,终端将采集到的针对预设初始语音的目标语音信息发送至服务器,服务器接收终端发送的目标语音信息,以根据目标语音信息进行业务办理。
具体地,所述将所述目标语音信息转换为对应的文本信息的步骤为:将所述目 标语音信息输入至预设语音识别模型,通过所述预设语音识别模型进行目标语音识别,得到所述目标语音信息对应的文本信息。更多实施例中,还可以对用户输入的目标语音信息进行预处理,所述预处理包括:分词、词性标注、命名实体识别、指代消歧、相似词语扩展。进一步地,所述分词、词性标注、命名实体识别是指使用深度神经网络训练的自然语言处理工具进行分词、词性标注、命名实体识别。所述分词是指将汉字序列切分成词序列。所述词性标注是指根据词语的词性进行判别及标注,词性包括:名词、副词、形容词、动词、代词等,例如,NT代表时间名词、V代表动词、NN代表口头名词、PU代表标定符号、AD代表副词、PN代表代词等。所述命名实体识别是指识别语句中人名、地名、组织机构名等命名实体,命名实体包括3大类,如实体类、时间类及数字类,和7小类,如人名、地名、机构名、时间、日期、货币及百分比。所述指代消歧是指消除人称代词的指代歧义,通过使用依存句法分析(Dependency Parsing,DP)识别语言单位内成分之间的依存关系揭示其句法结构,来完成指代消歧。所述相似语句扩展是指利用Word2vec技术进行相似词语的扩展。
将目标语音信息以及文本信息输入预先训练好的预设情绪识别模型,识别出目标用户的情绪类别。所述情绪类别包括:愤怒、焦急、生气、开心、失望、惊讶、好奇等。具体实施例中,还可以基于词典的情感分析方式和预先训练好的基于深度学习的情感分类模型,识别出用户的情绪类别。所述基于词典的情感分析方式是指通过具有语法敏感性的专业人士构建情感分析词典,根据构建的情感分析词典:正性情感词典、负性情感词典及中性情感词典,将某语句中用于表达情感的词汇分为三个类别,然后对比语句中正性、负性及中性情感词的个数,评估语句的情绪类别。所述基于深度学习的情感分类模型包括但不限于长短期记忆网络(Long Short-Term Memory,LSTM)模型、支持向量机(Support Vector Machine,SVM)模型、随机森林(Random Forests,RF)模型及朴素贝叶斯(Naive Bayesian Model,NBM)模型中一种预先训练好的模型。该模型是由已人工识别正、负及中性的文本通过机器学习等方式训练而成,在此不再赘述。更多实施例中,所述预设情绪识别模型的训练步骤为:获取预设训练语音信息以及对应的所属情绪类别信息,通过神经网络模型skip -gram或者连续词袋模型训练预设训练语音信息,得到低维度的字向量;通过查找字向量表的方式,将低维度的字向量转换成相应的向量,通过卷积、池化操作将所述向量表示式转换成特征向量,对所述特征向量使用Hard Tanh函数进行非线性特征的抽取,获取预设训练语音信息的最终特征;
根据所述预设训练语音信息的最终特征和对应的所属情绪类别信息生成所述预设情绪识别模型。
步骤S20,基于所述文本信息和所述情绪类别,确定所述目标用户对应的业务处理流程以及目标服务话术;
本实施例中,可预先建立不同情绪类别对应的服务话术库,如愤怒情绪类别对应的服务话术为简洁亲切地话术,如尊敬的客户X,您所需办理的业务为X,总共需要两个流程,第一个流程为X,第二个流程X,请核对上述信息是否正确等。焦急情绪类别对应的服务话术为安抚引导型,为了防止网络诈骗或者电话诈骗,还可以播放防诈骗宣传信息等。具体实施例中,还可以根据不同情绪类别设定不同的业务办理流程,如为焦急、愤怒情绪类别的用户推荐耗时最少或业务流程节点最少的高效率业务办理流程,为开心、愉悦的用户推荐加有理财产品推荐或者其他关联业务介绍的业务办理流程,以增加理财产品的销售量等。具体地,在确定文本信息和情绪类别之后,服务器根据文本信息确定所述目标用户需要办理的目标业务,如转账、贷款、办理银行卡或者信用卡等。然后根据所述目标用户的情绪类别确定适于服务所述目标用户的目标服务话术以及业务处理流程。
步骤S30,播报所述目标服务话术,并根据所述业务处理流程,指导所述目标用户完成对应的目标业务。
本实施例中,根据所述业务处理流程对应的各个操作节点显示对应的待操作界面,并同时播放对应的目标服务话术,如“将身份证放置于前方身份证识别区”。通过符合所述目标用户当前情绪类别的目标服务话术,来指导所述目标用户进行目标业务的办理。即确定适于服务所述目标用户对应的目标服务话术以及业务处理流程,并将所述业务处理流程对应的各个流程步骤的指导信息对应添加至所述目标服务话术中。其中,具有相同业务处理流程的业务类型,即具有相 同的流程引导信息的同种业务类型,可以同一套服务话术,然后针对具体业务名称或者业务编号等进行替换即可。如为焦急用户指导办理转账业务时,首先获取适用于焦急用户的欢迎语句,即尊敬的客户X,早上好,请不要着急,请留意引导语音,以便快速完成目标业务的办理。首先,请在身份证识别去放置您的身份证以读取您的身份信息,然后X,其次X......。具体实施例中,还可以将目标用户输入的当前收款人信息与所述目标用户对应的历史收款人信息进行比对,判断当前收款人是否为目标用户有交易往来的常用联系人。若不是,则播报防诈骗的警示信息,如最近诈骗犯罪案件频发,且当前收款人不是常用联系人,请再次确定是否进行当前转账操作等。从而在有效指导目标用户完成请求办理的目标业务的同时,提升用户体验。更多实施例中,还可以根据所述目标用户的情绪类别,显示当前操作界面的背景颜色,进一步提升用户体验。
本实施例提供一种基于情绪分析的业务处理方法,即在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别;基于所述文本信息和所述情绪类别,确定所述目标用户对应的业务处理流程以及目标服务话术;播报所述目标服务话术,并根据所述业务处理流程,指导所述目标用户完成对应的目标业务。通过上述方式,本申请根据用户语音信息确定用户的情绪类别,然后根据情绪类别确定对应的业务处理流程和对应的目标服务话术,然后根据所述业务处理流程和目标服务话术指导所述目标用户完成对应的目标业务,提升业务办理效率,提升用户体验,解决了现有自助服务终端业务办理效率低下以及用户体验差的技术问题。
参照图3,图3为本申请基于情绪分析的业务处理方法第二实施例的流程示意图。
基于上述图2所示实施例,本实施例中,所述步骤S30包括:
步骤S31,播报所述目标服务话术,并接收到所述目标用户基于所述目标服务话术反馈的反馈语音信息,并将所述反馈语音信息输入至预设语音识别模型,并通过所述预设语音识别模型得到所述反馈语音信息对应的反馈文本信息;
本实施例中,在播放所述目标服务话术的同时,实时获取目标用户基于所述目 标服务话术的反馈语音信息,通过预设语音识别模型,将所述反馈语音信息转换为对应的反馈文本信息。其中,所述预设语音设别模型生成步骤为:分别采集预设方言语系和普通话语系朗读语音数据,并提取各所述语音数据对应的语音特征参数,将各所述语音特征参数组成语音特征集合;抽取所述语音特征集合中预设比例的各所述语音特征参数,并通过各所述语音特征参数构建初始语音识别模型;通过迭代算法训练所述初始语音识别模型,并获取训练得到的语音识别模型的语音识别准确率,将识别准确率高于预设阈值的语音识别模型作为预设语音识别模型。
步骤S32,根据所述反馈文本信息,判断所述业务处理流程是否为所述目标用户请求办理的业务处理流程;
本实施例中,根据所述反馈文本信息,判断当前确定的目标服务话术以及业务处理流程是否是目标用户所请求办理的目标业务对应的目标服务话术以及目标业务对应的业务处理流程。若所述反馈文本信息为“不是”、“退出”等,即所述目标服务话术播报的业务处理流程的引导信息并不是目标用户所需办理的目标业务对应的业务流程。若所述反馈文本信息为“确认”、“谢谢”等,即所述目标服务话术播报的业务处理流程为目标用户需要办理的目标业务对应的业务流程。
步骤S33,若不是所述目标用户请求办理的业务处理流程,则根据所述反馈语音信息重新确定待办理目标业务对应的业务处理流程以及对应的服务话术。
本实施例中,若判定当前播报的业务处理流程错误,即不是所述目标用户请求办理的业务处理流程,可重新根据目标用户录入的反馈语音信息,如不是办理转账是办理储蓄的反馈语音信息,重新执行目标服务话术以及所述储蓄业务对应的业务处理流程的确定操作。具体实施例中,若所述目标用户的反馈语音信息为退出指令时,则重新录入目标用户的语音信息,以进行目标业务对应的业务处理流程以及目标服务话术的确定操作。
参照图4,图4为本申请基于情绪分析的业务处理方法第三实施例的流程示意图。
基于上述图2所示实施例,本实施例中,所述步骤S10具体包括:
步骤S11,在接收到终端发送的业务办理请求时,获取所述目标用户的人脸信 息,并根据所述人脸信息判断所述目标用户的年龄是否大于预设阈值;
本实施例中,为了防止中老年人由于不会使用自助终端,而造成业务办理障碍,可在检测到目标用户的年龄超过预设阈值时,如50岁,为所述目标用户提供人工协助。具体地,通过摄像头获取所述目标用户的人脸信息,并判断所述人脸信息对应的目标用户是否为需协助用户,即所述需协助人员为年龄超过预设阈值的用户。
步骤S12,若所述目标用户的年龄大于预设阈值,则根据当前业务终端的终端标识生成协助请求信息,并将所述协助请求信息发送至工作人员,以便提醒相关工作人员对所述目标用户提供人工指导;
本实施例中,若判定所述目标用户的年龄大于预设阈值,即所述目标用户符合需协助用户的条件,根据所述目标用户所办理业务的当前业务终端的终端标识,生成携带所述终端标识的协助请求信息,如3号自助终端当前有待协助用户在进行业务办理,请尽快前往协助处理,并将所述协助请求信息发送至工作人员的相关终端,如大堂经理对应的终端。从而及时提醒相关工作人员及时帮助一些年龄较大的中老年人进行自主终端的业务办理,提高业务办理效率。
步骤S13,若所述目标用户的年龄不大于预设阈值,则获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别。
本实施例中,若所述目标用户不是待协助用户,则根据用户语音信息确定用户的情绪类别,然后根据情绪类别确定对应的业务处理流程和对应的目标服务话术,然后根据所述业务处理流程和目标服务话术指导所述目标用户完成对应的目标业务。
进一步地,所述步骤S10之后,还包括:
判断所述目标用户的情绪类别是否为焦急情绪类别;
若判定所述目标用户的情绪类别为焦急情绪类别时,根据当前业务终端的终端标识生成紧急请求信息,并将所述紧急请求信息发送至工作人员,以便提醒相关工作人员对所述目标用户提供快捷协助。
本实施例中,还可以为具有紧急情况的用户提供快捷协助,如为银行卡被盗刷 的用户提供紧急服务窗口,以便快速为所述目标用户办理冻结银行卡的业务。即在判定所述目标用户的情绪类别为焦急情绪类别时,即根据所述目标用户办理业务的当前业务终端的终端标识向相关工作人员发送紧急请求信息,以便相关工作人员前往确认以及人工核查,并在确定目标用户确实所需办理业务确实为紧急业务时,为所述目标用户提供快捷协助,帮助所述目标用户快速完成目标业务。
此外,本申请实施例还提供一种基于情绪分析的业务处理装置。
参照图5,图5为本申请基于情绪分析的业务处理装置第一实施例的功能模块示意图。
本实施例中,所述基于情绪分析的业务处理装置包括:
情绪确定模块10,用于在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别;
话术确定模块20,用于基于所述文本信息和所述情绪类别,确定所述目标用户对应的业务处理流程以及目标服务话术;
处理指导模块30,用于播报所述目标服务话术,并根据所述业务处理流程,指导所述目标用户完成对应的目标业务。
进一步地,所述基于情绪分析的业务处理装置还包括:
信息训练模块,用于获取预设训练语音信息以及对应的所属情绪类别信息,通过神经网络模型skip-gram或者连续词袋模型训练预设训练语音信息,得到低维度的字向量;
特征提取模块,用于通过查找字向量表的方式,将低维度的字向量转换成相应的向量,通过卷积、池化操作将所述向量表示式转换成特征向量,对所述特征向量使用Hard Tanh函数进行非线性特征的抽取,获取预设训练语音信息的最终特征;
模型生成模块,用于根据所述预设训练语音信息的最终特征和对应的所属情绪类别信息生成所述预设情绪识别模型。
进一步地,所述情绪确定模块10还用于:
将所述目标语音信息输入至预设语音识别模型,通过所述预设语音识别模型进行目标语音识别,得到所述目标语音信息对应的文本信息。
进一步地,所述情绪确定模块10还用于:
分别采集预设方言语系和普通话语系朗读语音数据,并提取各所述语音数据对应的语音特征参数,将各所述语音特征参数组成语音特征集合;
抽取所述语音特征集合中预设比例的各所述语音特征参数,并通过各所述语音特征参数构建初始语音识别模型;
通过迭代算法训练所述初始语音识别模型,并获取训练得到的语音识别模型的语音识别准确率,将识别准确率高于预设阈值的语音识别模型作为预设语音识别模型。
进一步地,所述处理指导模块30具体包括:
反馈信息获取单元,用于播报所述目标服务话术,并接收到所述目标用户基于所述目标服务话术反馈的反馈语音信息,并将所述反馈语音信息输入至预设语音识别模型,并通过所述预设语音识别模型得到所述反馈语音信息对应的反馈文本信息;
业务流程核实单元,用于根据所述反馈文本信息,判断所述业务处理流程是否为所述目标用户请求办理的业务处理流程;
业务流程确定单元,用于若不是所述目标用户请求办理的业务处理流程,则根据所述反馈语音信息重新确定待办理目标业务对应的业务处理流程以及对应的服务话术。
进一步地,所述情绪确定模块10具体包括:
用户判断单元,用于在接收到终端发送的业务办理请求时,获取所述目标用户的人脸信息,并根据所述人脸信息判断所述目标用户的年龄是否大于预设阈值;
协助请求单元,用于若所述目标用户的年龄大于预设阈值,则根据当前业务终端的终端标识生成协助请求信息,并将所述协助请求信息发送至工作人员,以便提醒相关工作人员对所述目标用户提供人工指导;
进一步地,所述情绪确定模块10还用于:
若所述目标用户的年龄不大于预设阈值,则获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别。
进一步地,所述情绪确定模块10还用于:
判断所述目标用户的情绪类别是否为焦急情绪类别;
若判定所述目标用户的情绪类别为焦急情绪类别时,根据当前业务终端的终端标识生成紧急请求信息,并将所述紧急请求信息发送至工作人员,以便提醒相关工作人员对所述目标用户提供快捷协助。
其中,上述基于情绪分析的业务处理装置中各个模块与上述基于情绪分析的业务处理方法实施例中各步骤相对应,其功能和实现过程在此处不再一一赘述。
此外,本申请实施例还提供一种计算机可读存储介质。
本申请计算机可读存储介质上存储有基于情绪分析的业务处理程序,其中所述基于情绪分析的业务处理程序被处理器执行时,实现如上述的基于情绪分析的业务处理方法的步骤。所述计算机可读存储介质可以是非易失性,也可以是易失性。
其中,基于情绪分析的业务处理程序被执行时所实现的方法可参照本申请基于情绪分析的业务处理方法的各个实施例,此处不再赘述。
需要说明的是,在本文中,术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含,从而使得包括一系列要素的过程、方法、物品或者系统不仅包括那些要素,而且还包括没有明确列出的其他要素,或者是还包括为这种过程、方法、物品或者系统所固有的要素。在没有更多限制的情况下,由语句“包括一个......”限定的要素,并不排除在包括该要素的过程、方法、物品或者系统中还存在另外的相同要素。
上述本申请实施例序号仅仅为了描述,不代表实施例的优劣。
通过以上的实施方式的描述,本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现,当然也可以通过硬件,但很多情况下前者是更佳的实施方式。基于这样的理解,本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来,该计 算机软件产品存储在如上所述的一个存储介质(如ROM/RAM、磁碟、光盘)中,包括若干指令用以使得一台终端设备(可以是手机,计算机,服务器,空调器,或者网络设备等)执行本申请各个实施例所述的方法。
以上仅为本申请的优选实施例,并非因此限制本申请的专利范围,凡是利用本申请说明书及附图内容所作的等效结构或等效流程变换,或直接或间接运用在其他相关的技术领域,均同理包括在本申请的专利保护范围内。

Claims (20)

  1. 一种基于情绪分析的业务处理方法,其中,所述基于情绪分析的业务处理方法包括以下步骤:
    在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别;
    基于所述文本信息和所述情绪类别,确定所述目标用户对应的业务处理流程以及目标服务话术;
    播报所述目标服务话术,并根据所述业务处理流程,指导所述目标用户完成对应的目标业务。
  2. 如权利要求1所述的基于情绪分析的业务处理方法,其中,所述在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别的步骤之前,还包括:
    获取预设训练语音信息以及对应的所属情绪类别信息,通过神经网络模型skip-gram或者连续词袋模型训练预设训练语音信息,得到低维度的字向量;
    通过查找字向量表的方式,将低维度的字向量转换成相应的向量,通过卷积、池化操作将所述向量表示式转换成特征向量,对所述特征向量使用Hard Tanh函数进行非线性特征的抽取,获取预设训练语音信息的最终特征;
    根据所述预设训练语音信息的最终特征和对应的所属情绪类别信息生成所述预设情绪识别模型。
  3. 如权利要求1所述的基于情绪分析的业务处理方法,其中,所述将所述目标语音信息转换为对应的文本信息的步骤包括:
    将所述目标语音信息输入至预设语音识别模型,通过所述预设语 音识别模型进行目标语音识别,得到所述目标语音信息对应的文本信息。
  4. 如权利要求3所述的基于情绪分析的业务处理方法,其中,所述将所述目标语音信息输入至预设语音识别模型,通过所述预设语音识别模型进行目标语音识别,得到所述目标语音信息对应的文本信息的步骤之前,还包括:
    分别采集预设方言语系和普通话语系朗读语音数据,并提取各所述语音数据对应的语音特征参数,将各所述语音特征参数组成语音特征集合;
    抽取所述语音特征集合中预设比例的各所述语音特征参数,并通过各所述语音特征参数构建初始语音识别模型;
    通过迭代算法训练所述初始语音识别模型,并获取训练得到的语音识别模型的语音识别准确率,将识别准确率高于预设阈值的语音识别模型作为预设语音识别模型。
  5. 如权利要求4所述的基于情绪分析的业务处理方法,其中,所述播报所述目标服务话术,并根据所述业务处理流程,指导所述目标用户完成对应的目标业务的步骤具体包括:
    播报所述目标服务话术,并接收到所述目标用户基于所述目标服务话术反馈的反馈语音信息,并将所述反馈语音信息输入至预设语音识别模型,并通过所述预设语音识别模型得到所述反馈语音信息对应的反馈文本信息;
    根据所述反馈文本信息,判断所述业务处理流程是否为所述目标用户请求办理的业务处理流程;
    若不是所述目标用户请求办理的业务处理流程,则根据所述反馈语音信息重新确定待办理目标业务对应的业务处理流程以及对应的服务话术。
  6. 如权利要求1所述的基于情绪分析的业务处理方法,其中,所述在接收到终端发送的业务办理请求时,获取目标用户目标语音信息 ,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别的步骤之后,还包括:
    判断所述目标用户的情绪类别是否为焦急情绪类别;
    若判定所述目标用户的情绪类别为焦急情绪类别时,根据当前业务终端的终端标识生成紧急请求信息,并将所述紧急请求信息发送至工作人员,以便提醒相关工作人员对所述目标用户提供快捷协助。
  7. 如权利要求1至6任意一项所述的基于情绪分析的业务处理方法,其中,所述在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别的步骤包括:
    在接收到终端发送的业务办理请求时,获取所述目标用户的人脸信息,并根据所述人脸信息判断所述目标用户的年龄是否大于预设阈值;
    若所述目标用户的年龄大于预设阈值,则根据当前业务终端的终端标识生成协助请求信息,并将所述协助请求信息发送至工作人员,以便提醒相关工作人员对所述目标用户提供人工指导;
    若所述目标用户的年龄不大于预设阈值,则获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别。
  8. 一种基于情绪分析的业务处理设备,其中,所述基于情绪分析的业务处理设备包括处理器、存储器、以及存储在所述存储器上并可被所述处理器执行的基于情绪分析的业务处理程序,其中所述基于情绪分析的业务处理程序被所述处理器执行时,实现如下步骤:
    在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别;
    基于所述文本信息和所述情绪类别,确定所述目标用户对应的业务处理流程以及目标服务话术;
    播报所述目标服务话术,并根据所述业务处理流程,指导所述目标用户完成对应的目标业务。
  9. 如权利要求8所述的基于情绪分析的业务处理设备,其中,所述在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别的步骤之前,所述基于情绪分析的业务处理程序还被所述处理器执行实现:
    获取预设训练语音信息以及对应的所属情绪类别信息,通过神经网络模型skip-gram或者连续词袋模型训练预设训练语音信息,得到低维度的字向量;
    通过查找字向量表的方式,将低维度的字向量转换成相应的向量,通过卷积、池化操作将所述向量表示式转换成特征向量,对所述特征向量使用Hard Tanh函数进行非线性特征的抽取,获取预设训练语音信息的最终特征;
    根据所述预设训练语音信息的最终特征和对应的所属情绪类别信息生成所述预设情绪识别模型。
  10. 如权利要求8所述的基于情绪分析的业务处理设备,其中,所述基于情绪分析的业务处理程序被所述处理器执行实现所述将所述目标语音信息转换为对应的文本信息的步骤,包括:
    将所述目标语音信息输入至预设语音识别模型,通过所述预设语音识别模型进行目标语音识别,得到所述目标语音信息对应的文 本信息。
  11. 如权利要求10所述的基于情绪分析的业务处理设备,其中,所述将所述目标语音信息输入至预设语音识别模型,通过所述预设语音识别模型进行目标语音识别,得到所述目标语音信息对应的文本信息的步骤之前,所述基于情绪分析的业务处理程序还被所述处理器执行实现:
    分别采集预设方言语系和普通话语系朗读语音数据,并提取各所述语音数据对应的语音特征参数,将各所述语音特征参数组成语音特征集合;
    抽取所述语音特征集合中预设比例的各所述语音特征参数,并通过各所述语音特征参数构建初始语音识别模型;
    通过迭代算法训练所述初始语音识别模型,并获取训练得到的语音识别模型的语音识别准确率,将识别准确率高于预设阈值的语音识别模型作为预设语音识别模型。
  12. 如权利要求11所述的基于情绪分析的业务处理设备,其中,所述基于情绪分析的业务处理程序被所述处理器执行实现所述播报所述目标服务话术,并根据所述业务处理流程,指导所述目标用户完成对应的目标业务的步骤具体包括:
    播报所述目标服务话术,并接收到所述目标用户基于所述目标服务话术反馈的反馈语音信息,并将所述反馈语音信息输入至预设语音识别模型,并通过所述预设语音识别模型得到所述反馈语音信息对应的反馈文本信息;
    根据所述反馈文本信息,判断所述业务处理流程是否为所述目标用户请求办理的业务处理流程;
    若不是所述目标用户请求办理的业务处理流程,则根据所述反馈语音信息重新确定待办理目标业务对应的业务处理流程以及对应的服务话术。
  13. 如权利要求8所述的基于情绪分析的业务处理设备,其中,所述在 接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别的步骤之后,所述基于情绪分析的业务处理程序还被所述处理器执行实现:
    判断所述目标用户的情绪类别是否为焦急情绪类别;
    若判定所述目标用户的情绪类别为焦急情绪类别时,根据当前业务终端的终端标识生成紧急请求信息,并将所述紧急请求信息发送至工作人员,以便提醒相关工作人员对所述目标用户提供快捷协助。
  14. 如权利要求8至13任意一项所述的基于情绪分析的业务处理设备,其中,所述基于情绪分析的业务处理程序被所述处理器执行实现所述在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别的步骤包括:
    在接收到终端发送的业务办理请求时,获取所述目标用户的人脸信息,并根据所述人脸信息判断所述目标用户的年龄是否大于预设阈值;
    若所述目标用户的年龄大于预设阈值,则根据当前业务终端的终端标识生成协助请求信息,并将所述协助请求信息发送至工作人员,以便提醒相关工作人员对所述目标用户提供人工指导;
    若所述目标用户的年龄不大于预设阈值,则获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别。
  15. 一种计算机可读存储介质,其中,所述计算机可读存储介质上存储有基于情绪分析的业务处理程序,其中所述基于情绪分析的业 务处理程序被处理器执行时,实现如下步骤:
    在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别;
    基于所述文本信息和所述情绪类别,确定所述目标用户对应的业务处理流程以及目标服务话术;
    播报所述目标服务话术,并根据所述业务处理流程,指导所述目标用户完成对应的目标业务。
  16. 如权利要求15所述的计算机可读存储介质,其中,所述在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别的步骤之前,所述基于情绪分析的业务处理程序还被所述处理器执行实现:
    获取预设训练语音信息以及对应的所属情绪类别信息,通过神经网络模型skip-gram或者连续词袋模型训练预设训练语音信息,得到低维度的字向量;
    通过查找字向量表的方式,将低维度的字向量转换成相应的向量,通过卷积、池化操作将所述向量表示式转换成特征向量,对所述特征向量使用Hard Tanh函数进行非线性特征的抽取,获取预设训练语音信息的最终特征;
    根据所述预设训练语音信息的最终特征和对应的所属情绪类别信息生成所述预设情绪识别模型。
  17. 如权利要求15所述的计算机可读存储介质,其中,所述基于情绪分析的业务处理程序被所述处理器执行实现所述将所述目标语音信息转换为对应的文本信息的步骤包括:
    将所述目标语音信息输入至预设语音识别模型,通过所述预设语 音识别模型进行目标语音识别,得到所述目标语音信息对应的文本信息。
  18. 如权利要求17所述的计算机可读存储介质,其中,所述将所述目标语音信息输入至预设语音识别模型,通过所述预设语音识别模型进行目标语音识别,得到所述目标语音信息对应的文本信息的步骤之前,所述基于情绪分析的业务处理程序还被所述处理器执行实现:
    分别采集预设方言语系和普通话语系朗读语音数据,并提取各所述语音数据对应的语音特征参数,将各所述语音特征参数组成语音特征集合;
    抽取所述语音特征集合中预设比例的各所述语音特征参数,并通过各所述语音特征参数构建初始语音识别模型;
    通过迭代算法训练所述初始语音识别模型,并获取训练得到的语音识别模型的语音识别准确率,将识别准确率高于预设阈值的语音识别模型作为预设语音识别模型。
  19. 如权利要求18所述的计算机可读存储介质,其中,所述基于情绪分析的业务处理程序被所述处理器执行实现所述播报所述目标服务话术,并根据所述业务处理流程,指导所述目标用户完成对应的目标业务的步骤具体包括:
    播报所述目标服务话术,并接收到所述目标用户基于所述目标服务话术反馈的反馈语音信息,并将所述反馈语音信息输入至预设语音识别模型,并通过所述预设语音识别模型得到所述反馈语音信息对应的反馈文本信息;
    根据所述反馈文本信息,判断所述业务处理流程是否为所述目标用户请求办理的业务处理流程;
    若不是所述目标用户请求办理的业务处理流程,则根据所述反馈语音信息重新确定待办理目标业务对应的业务处理流程以及对应的服务话术。
  20. 如权利要求15所述的计算机可读存储介质,其中,所述在接收到终端发送的业务办理请求时,获取目标用户目标语音信息,并将所述目标语音信息转换为对应的文本信息,基于预设情绪识别模型、所述目标语音信息和文本信息,确定所述目标用户的情绪类别的步骤之后,所述基于情绪分析的业务处理程序还被所述处理器执行实现:
    判断所述目标用户的情绪类别是否为焦急情绪类别;
    若判定所述目标用户的情绪类别为焦急情绪类别时,根据当前业务终端的终端标识生成紧急请求信息,并将所述紧急请求信息发送至工作人员,以便提醒相关工作人员对所述目标用户提供快捷协助。
PCT/CN2020/086161 2019-06-20 2020-04-22 基于情绪分析的业务处理方法、装置、设备及存储介质 WO2020253362A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910537349.1 2019-06-20
CN201910537349.1A CN110379445A (zh) 2019-06-20 2019-06-20 基于情绪分析的业务处理方法、装置、设备及存储介质

Publications (1)

Publication Number Publication Date
WO2020253362A1 true WO2020253362A1 (zh) 2020-12-24

Family

ID=68249062

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/086161 WO2020253362A1 (zh) 2019-06-20 2020-04-22 基于情绪分析的业务处理方法、装置、设备及存储介质

Country Status (2)

Country Link
CN (1) CN110379445A (zh)
WO (1) WO2020253362A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115471141A (zh) * 2022-11-02 2022-12-13 成都飞机工业(集团)有限责任公司 一种业务流程周期管控方法、装置、设备及介质

Families Citing this family (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110379445A (zh) * 2019-06-20 2019-10-25 深圳壹账通智能科技有限公司 基于情绪分析的业务处理方法、装置、设备及存储介质
CN111026867A (zh) * 2019-11-28 2020-04-17 杭州飞步科技有限公司 客诉处理方法和装置、电子设备、存储介质
CN111177308B (zh) * 2019-12-05 2023-07-18 上海云洽信息技术有限公司 一种文本内容的识别情绪方法
CN111062332A (zh) * 2019-12-18 2020-04-24 秒针信息技术有限公司 信息推送方法和装置
CN111179903A (zh) * 2019-12-30 2020-05-19 珠海格力电器股份有限公司 一种语音识别方法、装置、存储介质及电器
CN113539275A (zh) * 2020-04-22 2021-10-22 北京有限元科技有限公司 确定话术的方法、装置以及存储介质
CN111540358B (zh) * 2020-04-26 2023-05-26 云知声智能科技股份有限公司 人机交互方法、装置、设备和存储介质
WO2021217769A1 (zh) * 2020-04-27 2021-11-04 平安科技(深圳)有限公司 基于情绪识别的答复方法、装置、计算机设备及存储介质
CN112201277B (zh) * 2020-09-29 2024-03-22 中国银行股份有限公司 语音应答的方法、装置、及设备及计算机可读存储介质
CN112612878A (zh) * 2020-12-17 2021-04-06 大唐融合通信股份有限公司 一种客服信息提供方法、电子设备及装置
CN112949708B (zh) * 2021-02-26 2023-10-24 平安科技(深圳)有限公司 情绪识别方法、装置、计算机设备和存储介质
CN112951429A (zh) * 2021-03-25 2021-06-11 浙江连信科技有限公司 用于中小学生心理危机筛查的信息处理方法及装置
CN113345419B (zh) * 2021-06-30 2022-05-27 广西电网有限责任公司 基于方言口音的语音转译方法、系统和可读存储介质
CN113516183B (zh) * 2021-07-05 2024-04-16 深圳小湃科技有限公司 故障响应方法、系统、设备及存储介质
CN113609273A (zh) * 2021-08-12 2021-11-05 云知声(上海)智能科技有限公司 一种机器话术配置方法、装置、电子设备和存储介质
CN113743126B (zh) * 2021-11-08 2022-06-14 北京博瑞彤芸科技股份有限公司 一种基于用户情绪的智能交互方法和装置
CN115022395B (zh) * 2022-05-27 2023-08-08 艾普科创(北京)控股有限公司 业务视频推送方法、装置、电子设备及存储介质
CN115171284B (zh) * 2022-07-01 2023-12-26 国网汇通金财(北京)信息科技有限公司 一种老年人关怀方法及装置

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080096533A1 (en) * 2006-10-24 2008-04-24 Kallideas Spa Virtual Assistant With Real-Time Emotions
CN102802114A (zh) * 2012-06-20 2012-11-28 北京语言大学 利用语音进行座席筛选的方法及系统
CN105895101A (zh) * 2016-06-08 2016-08-24 国网上海市电力公司 用于电力智能辅助服务系统的语音处理设备及处理方法
CN109767791A (zh) * 2019-03-21 2019-05-17 中国—东盟信息港股份有限公司 一种针对呼叫中心通话的语音情绪识别及应用系统
CN109815494A (zh) * 2019-01-16 2019-05-28 中民乡邻投资控股有限公司 一种基于客户情绪的问答服务方法
CN110379445A (zh) * 2019-06-20 2019-10-25 深圳壹账通智能科技有限公司 基于情绪分析的业务处理方法、装置、设备及存储介质

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10037767B1 (en) * 2017-02-01 2018-07-31 Wipro Limited Integrated system and a method of identifying and learning emotions in conversation utterances
CN107705807B (zh) * 2017-08-24 2019-08-27 平安科技(深圳)有限公司 基于情绪识别的语音质检方法、装置、设备及存储介质
CN108427722A (zh) * 2018-02-09 2018-08-21 卫盈联信息技术(深圳)有限公司 智能交互方法、电子装置及存储介质
CN109033257A (zh) * 2018-07-06 2018-12-18 中国平安人寿保险股份有限公司 话术推荐方法、装置、计算机设备和存储介质
CN109389971B (zh) * 2018-08-17 2022-06-17 深圳壹账通智能科技有限公司 基于语音识别的保险录音质检方法、装置、设备和介质

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080096533A1 (en) * 2006-10-24 2008-04-24 Kallideas Spa Virtual Assistant With Real-Time Emotions
CN102802114A (zh) * 2012-06-20 2012-11-28 北京语言大学 利用语音进行座席筛选的方法及系统
CN105895101A (zh) * 2016-06-08 2016-08-24 国网上海市电力公司 用于电力智能辅助服务系统的语音处理设备及处理方法
CN109815494A (zh) * 2019-01-16 2019-05-28 中民乡邻投资控股有限公司 一种基于客户情绪的问答服务方法
CN109767791A (zh) * 2019-03-21 2019-05-17 中国—东盟信息港股份有限公司 一种针对呼叫中心通话的语音情绪识别及应用系统
CN110379445A (zh) * 2019-06-20 2019-10-25 深圳壹账通智能科技有限公司 基于情绪分析的业务处理方法、装置、设备及存储介质

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115471141A (zh) * 2022-11-02 2022-12-13 成都飞机工业(集团)有限责任公司 一种业务流程周期管控方法、装置、设备及介质
CN115471141B (zh) * 2022-11-02 2023-03-24 成都飞机工业(集团)有限责任公司 一种业务流程周期管控方法、装置、设备及介质

Also Published As

Publication number Publication date
CN110379445A (zh) 2019-10-25

Similar Documents

Publication Publication Date Title
WO2020253362A1 (zh) 基于情绪分析的业务处理方法、装置、设备及存储介质
WO2019153522A1 (zh) 智能交互方法、电子装置及存储介质
US9740677B2 (en) Methods and systems for analyzing communication situation based on dialogue act information
US10977452B2 (en) Multi-lingual virtual personal assistant
US11500917B2 (en) Providing a summary of a multimedia document in a session
US7860705B2 (en) Methods and apparatus for context adaptation of speech-to-speech translation systems
US9753914B2 (en) Natural expression processing method, processing and response method, device, and system
US9792279B2 (en) Methods and systems for analyzing communication situation based on emotion information
WO2021042904A1 (zh) 会话意图识别方法、装置、计算机设备和存储介质
KR20190039572A (ko) 다음 사용자 프롬프트 유형 선택하기
CN111540353B (zh) 一种语义理解方法、装置、设备及存储介质
US9792909B2 (en) Methods and systems for recommending dialogue sticker based on similar situation detection
CN111241237A (zh) 一种基于运维业务的智能问答数据处理方法及装置
CN107967250B (zh) 一种信息处理方法及装置
Adel et al. Features for factored language models for code-Switching speech.
US11163961B2 (en) Detection of relational language in human-computer conversation
CN110223134B (zh) 基于语音识别的产品推荐方法及相关设备
CN112699686A (zh) 基于任务型对话系统的语义理解方法、装置、设备及介质
CN114817507A (zh) 基于意图识别的回复推荐方法、装置、设备和存储介质
TW202133027A (zh) 人機協作對話系統與方法
CN115602160A (zh) 基于语音识别的业务办理方法、装置及电子设备
CN110765242A (zh) 一种客服信息的提供方法,装置及系统
CN114418320A (zh) 客服服务质量评估方法、装置、设备、介质和程序产品
CN112800189A (zh) 人机交互方法、装置、智能机器人和存储介质
CN109359181B (zh) 负面情绪原因识别方法、设备及计算机可读存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20825616

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20825616

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 19.08.2022)

122 Ep: pct application non-entry in european phase

Ref document number: 20825616

Country of ref document: EP

Kind code of ref document: A1