Intelligent customer service sentiment analysis method and system based on machine learning and electronic equipment
Through the intelligent customer service emotion analysis system based on machine learning, combined with voice, text and picture data, the accurate recognition and response to customer emotions is achieved, and the problem that the existing intelligent customer service system cannot recognize customer emotional status is solved, improving user experience and Quality of service.
Patent Information
- Application Number
- CN202411444860.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-16
- Publication Date
- 2024-12-17
AI Technical Summary
The existing intelligent customer service system cannot accurately identify the customer's emotional state, such as anger, disappointment or satisfaction, and thus cannot provide personalized services, resulting in poor user experience.
Using an intelligent customer service emotion analysis system based on machine learning, the emotional thesaurus is constructed by the emotional thesaurus, user interaction module, emotion analysis module and self-learning module, combined with voice, text and picture data, and used deep learning models for emotion analysis, and based on the analysis results Generate response content and optimize the system in real time to improve service quality.
It realizes accurate recognition and response to customer emotions, improves user experience and satisfaction, and continuously optimizes the sentiment analysis model and response accuracy through automatic update of the emotional vocabulary.
Smart Images
Figure CN119150885A_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the field of computer technology, and in particular, relates to an intelligent customer service sentiment analysis method, system and electronic device based on machine learning. Background Art
[0002] With the development of Internet technology, the continuous expansion of application scenarios and the increase in customer personalized needs, the limitations of traditional text and telephone customer service have become increasingly prominent. There are problems such as time and space constraints, long waiting times, low efficiency, language barriers, limited emotional interaction, and limited personalized services. With the rapid development and increasing improvement of technologies such as AI technology, multimodal interaction capabilities, and cloud computing capabilities, intelligent customer service systems have been realized and continuously improved.
[0003] Existing operator intelligent customer service systems often lack a deep understanding of and ability to respond to user emotions, and are unable to provide truly personalized services; they fail to identify customers' emotional states, such as anger, disappointment, or satisfaction, and are thus unable to accurately serve users, resulting in a poor user experience.
[0004] In response to the above problems, the present application proposes an intelligent customer service sentiment analysis method, system and electronic device based on machine learning. Summary of the invention
[0005] In order to address the deficiencies of the prior art, the present application provides an intelligent customer service sentiment analysis system and electronic device based on machine learning to solve the problems that the prior art fails to identify the customer's emotional state, such as anger, disappointment or satisfaction, and thus cannot accurately serve users, cannot provide truly personalized services, and has a poor user experience.
[0006] The technical effects to be achieved by this application are achieved through the following solutions: In a first aspect, the present application provides an intelligent customer service sentiment analysis system based on machine learning, the system comprising: The sentiment word library construction module is used to obtain corpus information and construct a sentiment word library based on the corpus information; the corpus information includes voice, text and pictures; A user interaction module, used to interact with the user and collect the user's consultation data, wherein the consultation data includes text and voice; A sentiment analysis module, used to obtain the consultation data, perform sentiment analysis on the consultation data using a deep learning model, and determine a sentiment analysis result of the user; a response generation module, used to obtain the sentiment analysis result, generate corresponding response content according to the sentiment analysis result, and feed back the response content to the user; and The self-learning module is used to optimize the system in real time, learn user feedback information and consultation data of each user, and automatically update the sentiment word library.
[0007] In some embodiments, constructing a sentiment word library based on the corpus information includes: Performing data preprocessing on the corpus information to obtain training corpus; The training corpus is sentimentally annotated to determine the sentiment results of unknown sentiment words.
[0008] In some embodiments, the sentiment tagging of the training corpus includes: Unknown sentiment word is S * , where W i is a Chinese character in the unknown sentiment word, and the unknown sentiment word is represented by W1W2…W n , i and n are integers; Each Chinese character W in the unknown sentiment word i The probability is: ; Among them, Fre(S * ) represents unknown sentiment word S * The frequency in the training corpus, Fre(S * ,W i ) represents the Chinese character W i In the unknown sentiment word S * The frequency in The probability that the unknown sentiment word is a sentiment word is: ; Wherein, C represents the set of words in the training corpus, Z i represents a word in set C, Fre(C) represents the frequency of set C, Fre(Z i ) represents the word Z in the set C i frequency.
[0009] In some embodiments, determining the sentiment result of the unknown sentiment word includes: Based on each Chinese character W in the unknown sentiment word i The probability p(W i / S * ) and the probability p(S * ) Calculate the unknown sentiment word S * Emotional weight: ; Wherein, the sentiment weight of the unknown sentiment word includes a positive weight value and a negative weight value; The sentiment result of the unknown sentiment word is the difference between the positive weight value and the negative weight value. If the difference is a positive value, it indicates that the unknown sentiment word is a positive sentiment word. If the difference is not a positive value, it indicates that the unknown sentiment word is a negative sentiment word.
[0010] In some embodiments, the user interaction module has an interactive interface, and the interactive interface includes a text input box, a voice input option, emoticons, a video voice collector, and a user video expression collector; The user interaction module further includes an intention recognition unit and an entity recognition unit, wherein the intention recognition unit determines the user's query intention, wherein the query intention includes consultation, complaint, and suggestion; The entity recognition unit extracts key information from user input, where the key information includes time, location, and product model.
[0011] In some embodiments, the emotion analysis module includes a multimodal emotion recognition unit and a real-time analysis and feedback unit, wherein the multimodal emotion recognition unit uses deep learning technology to extract emotional features and rhythmic features in speech, and distinguishes positive emotions, neutral emotions, and negative emotions contained in the user's speech; the real-time analysis and feedback unit performs real-time emotion analysis and feeds back the emotion analysis results to the intelligent customer service system, so that the intelligent customer service system can provide corresponding services and responses based on the user's emotional state.
[0012] In some embodiments, the self-learning module includes a complaint or error analysis unit, which identifies errors made by the system in understanding a problem based on user feedback cases, corrects them, and provides feedback.
[0013] In some embodiments, the self-learning module further includes a knowledge base updating unit, and the knowledge base updating unit automatically updates the emotion vocabulary in real time.
[0014] In a second aspect, the present application provides an intelligent customer service sentiment analysis method based on machine learning, the method comprising: The emotional word library construction module obtains corpus information and constructs an emotional word library based on the corpus information; the corpus information includes voice, text and pictures; The user interaction module interacts with the user and collects the user's consultation data, wherein the consultation data includes text and voice; The sentiment analysis module obtains the consultation data, performs sentiment analysis on the consultation data using a deep learning model, and determines a sentiment analysis result of the user; The response generation module obtains the sentiment analysis result, generates corresponding response content according to the sentiment analysis result, and feeds back the response content to the user; and The self-learning module optimizes the system in real time, learns user feedback information and consultation data of each user, and automatically updates the sentiment word library.
[0015] In some embodiments, constructing a sentiment word library based on the corpus information includes: Performing data preprocessing on the corpus information to obtain training corpus; The training corpus is sentimentally annotated to determine the sentiment results of unknown sentiment words.
[0016] In some embodiments, the sentiment tagging of the training corpus includes: Unknown sentiment word is S * , where W i is a Chinese character in the unknown sentiment word, and the unknown sentiment word is represented by W1W2…W n , i and n are integers; Each Chinese character W in the unknown sentiment word i The probability is: ; Among them, Fre(S * ) represents the frequency of the position sentiment word in the training expectation, Fre(S * ,W i ) represents the Chinese character W i In the unknown sentiment word S * The frequency in The probability that the unknown sentiment word is a sentiment word is: ; Wherein, C represents the set of words in the training corpus, Z i represents a word in set C, Fre(C) represents the frequency of set C, Fre(Z i ) represents the word Z in the set C i frequency.
[0017] In some embodiments, determining the sentiment result of the unknown sentiment word includes: Based on each Chinese character W in the unknown sentiment word i The probability p(W i / S * ) and the probability p(S * ) Calculate the unknown sentiment word S * Emotional weight: ; Wherein, the sentiment weight of the unknown sentiment word includes a positive weight value and a negative weight value; The sentiment result of the unknown sentiment word is the difference between the positive weight value and the negative weight value. If the difference is a positive value, it indicates that the unknown sentiment word is a positive sentiment word. If the difference is not a positive value, it indicates that the unknown sentiment word is a negative sentiment word.
[0018] In a third aspect, the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any one of the methods described above when executing the computer program.
[0019] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement any of the methods described above.
[0020] Through the machine learning-based intelligent customer service sentiment analysis system and electronic equipment provided in this application, the system builds a vertical field sentiment corpus through advanced natural language processing and machine learning technologies, realizes accurate identification and response to customer emotions, and automatically updates the knowledge base based on customer feedback data, continuously optimizing the sentiment analysis model and response accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] In order to more clearly illustrate the embodiments of the present application or the existing technical solutions, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0022] Figure 1 This is a structural diagram of an intelligent customer service sentiment analysis system based on machine learning in one embodiment of the present application; Figure 2 This is a flow chart of an intelligent customer service sentiment analysis method based on machine learning in one embodiment of the present application; Figure 3 It is a schematic block diagram of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION
[0023] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.
[0024] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in one or more embodiments of the present application should be understood by people with ordinary skills in the field to which the present application belongs. The "first", "second" and similar words used in one or more embodiments of the present application do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.
[0025] The technological development of customer service systems is from traditional telephone call centers to Internet online customer service, and then to cloud call centers and cloud customer service software, and finally to modern intelligent customer service systems. Although intelligent customer service cannot completely replace traditional manual customer service, intelligence and personalization must be the main development direction of customer service systems in the future.
[0026] Currently, sentiment analysis has been widely used in all walks of life and has gradually shifted from general fields to vertical fields. The performance of sentiment analysis models is highly dependent on large amounts of high-quality labeled data, and the quality of data and the accuracy of annotations are one of the challenges. Currently, there are many corpora of sentimental terms, but there are few corpora for operator vertical fields.
[0027] There is a lack of emotional word database in the operator vertical field. In actual customer service scenarios, customers’ questions are diverse and mostly colloquial, which may involve the use of professional terms in the operator field.
[0028] In sentiment analysis, it is very important to analyze sentiment in context, and the lack of contextual data will greatly reduce the accuracy of the analysis.
[0029] Existing intelligent customer service focuses more on generating text that conforms to grammatical and semantic rules based on user questions to answer customer questions, but lacks a deep understanding of emotions and intentions.
[0030] The key to intelligent customer service is to continuously optimize sentiment analysis models and response strategies based on customer feedback in real time, and to provide personalized customer service experience based on the user's historical interaction records and emotional state. However, existing intelligent customer service systems are relatively weak in this regard.
[0031] Sentiment analysis technology enables intelligent customer service to identify the emotional state of customers, such as anger, disappointment or satisfaction, and provide personalized services accordingly, thereby improving customer experience and satisfaction. Therefore, it is necessary to propose an intelligent customer service sentiment analysis system based on machine learning in this application.
[0032] Various non-limiting implementations of the present application are described in detail below in conjunction with the accompanying drawings.
[0033] First, refer to Figure 1 , the machine learning-based intelligent customer service sentiment analysis system of this application is described in detail.
[0034] The present application provides an intelligent customer service sentiment analysis system based on machine learning, the system comprising: The sentiment word library construction module is used to obtain corpus information and construct a sentiment word library based on the corpus information; the corpus information includes voice, text and pictures; specifically, the expected information is obtained from the data collection module, and the corpus information includes but is not limited to information from customer service records, user feedback and evaluation information, user evaluation, etc.
[0035] A user interaction module, used to interact with the user and collect the user's consultation data, wherein the consultation data includes text and voice; A sentiment analysis module, used to obtain the consultation data, perform sentiment analysis on the consultation data using a deep learning model, and determine a sentiment analysis result of the user; a response generation module, used to obtain the sentiment analysis result, generate corresponding response content according to the sentiment analysis result, and feed back the response content to the user; and The self-learning module is used to optimize the system in real time, learn user feedback information and consultation data of each user, and automatically update the sentiment word library.
[0036] In some embodiments, constructing a sentiment word library based on the corpus information includes: Performing data preprocessing on the corpus information to obtain training corpus; The training corpus is sentimentally annotated to determine the sentiment results of unknown sentiment words.
[0037] Specifically, the sentiment corpus construction module includes the following units: Data collection unit: used to collect relevant text data, voice data and image data generated by the intelligent customer service system, including customer service records, user feedback and evaluation, customer comments and other data.
[0038] Data preprocessing unit: used to convert the collected voice and image data into text, unify the data format, and clean the unified text, including removing irrelevant characters, removing stop words, part-of-speech tagging and word segmentation, and using the cleaned data as training corpus for constructing the sentiment vocabulary.
[0039] The sentiment dictionary preparation unit is used to collect open source sentiment dictionaries as training sets.
[0040] The sentiment word tagging unit performs sentiment tagging on the training corpus, including: Unknown sentiment word is S * , where W i is a Chinese character in the unknown sentiment word, and the unknown sentiment word is represented by W1W2…W n , i and n are integers; Each Chinese character W in the unknown sentiment word i The probability is: ; Among them, Fre(S * ) represents unknown sentiment word S * The frequency in the training corpus, Fre(S * ,W i ) represents the Chinese character W i In the unknown sentiment word S * The frequency in The probability that the unknown sentiment word is a sentiment word is: ; Wherein, C represents the set of words in the training corpus, Z i represents a word in set C, Fre(C) represents the frequency of set C, Fre(Z i ) represents the word Z in the set C i The frequency of, i and n are integers.
[0041] In some embodiments, determining the sentiment result of the unknown sentiment word includes: According to the Bayesian formula, and based on each Chinese character W in the unknown sentiment word i The probability p(W i / S * ) and the probability p(S * ) Calculate the unknown sentiment word S * Emotional weight: ; Among them, p(S * / W1W2…W n ) represents unknown sentiment words W1W2…W nis the probability of sentiment words, since p(W1W2…W n / S * ) is unknown and uncalculated. Assuming that the probability of each Chinese character in the unknown sentiment word is conditionally independent, we can derive S # Approximately .
[0042] The calculated sentiment weight of the unknown sentiment word includes the positive weight value S # 正 and negative weight S # 负 ; If the difference is a positive value, it indicates that the unknown sentiment word is a positive sentiment word. If the difference is not a positive value, it indicates that the unknown sentiment word is a negative sentiment word. For example, the sentiment tendency of the sentiment word can be represented by S #1 =S # 正 -S # 负 , that is, if S # If it is a positive value, the word is a positive sentiment word, otherwise it is a negative sentiment word.
[0043] Based on the above logic, a sentiment corpus, for example, in the operator field, can be established based on the collected corpus information.
[0044] In some embodiments, the user interaction module has an interactive interface, and the interactive interface includes a text input box, a voice input option, emoticons, a video voice collector, and a user video expression collector; The user interaction module further includes an intention recognition unit and an entity recognition unit, wherein the intention recognition unit determines the user's query intention, wherein the query intention includes consultation, complaint, and suggestion; The entity recognition unit extracts key information from user input, where the key information includes time, location, and product model.
[0045] Exemplarily, the user interaction module has a multimodal input support unit: it implements input support for multiple forms such as text, voice, picture, video, etc.
[0046] Exemplarily, the user interaction module has a natural language understanding (NLU) unit: integrating natural language understanding technology enables the system to understand the user's natural language input and extract key information.
[0047] Exemplarily, the user interaction module has a context management unit: a context management system is designed to enable the system to keep track of user intentions and topics during the conversation.
[0048] In some embodiments, the emotion analysis module includes a multimodal emotion recognition unit and a real-time analysis and feedback unit, wherein the multimodal emotion recognition unit uses deep learning technology to extract emotional features and rhythmic features in speech, and distinguishes positive emotions, neutral emotions, and negative emotions contained in the user's speech; the real-time analysis and feedback unit performs real-time emotion analysis and feeds back the emotion analysis results to the intelligent customer service system, so that the intelligent customer service system can provide corresponding services and responses based on the user's emotional state.
[0049] In some embodiments, the self-learning module includes a complaint or error analysis unit, which identifies errors made by the system in understanding a problem based on user feedback cases, corrects them, and provides feedback.
[0050] In some embodiments, the self-learning module further includes a knowledge base updating unit, and the knowledge base updating unit automatically updates the emotion vocabulary in real time.
[0051] Exemplarily, the response generation module automatically generates corresponding response content based on the result of sentiment analysis and provides it to the user. The response generation module includes: Knowledge base retrieval unit: retrieves relevant information and preset response templates from the knowledge base based on user intention and emotional state.
[0052] Response content generation unit: Generates response content that is natural, fluent, and consistent with the user's emotional state. This may include text, voice, or multi-modal responses.
[0053] Personalized recommendation unit: Provides personalized service recommendations or solutions based on the user's emotional state and historical interaction records.
[0054] Exception handling unit: When the system cannot understand user input or encounters an error, it provides appropriate feedback and guides the user to further interaction when necessary.
[0055] In some embodiments, the self-learning module is responsible for continuously optimizing the system by continuously learning user feedback and consultation data to improve the accuracy of sentiment analysis and the quality of responses. The self-learning module includes: Complaint or Error Analysis Unit: Analyze cases where user feedback is poorly accurate or unsatisfactory, identify errors in the system's understanding of the problem, and continuously correct system feedback to improve response accuracy.
[0056] Model iteration unit: Regularly update and optimize sentiment analysis models based on new data and forecasts collected to improve the accuracy of intelligent customer service responses and sentiment analysis.
[0057] Model self-assessment unit: Based on complaints and relevant data such as performance collected, the model is self-assessed regularly to ensure continuous improvement of service and performance.
[0058] Knowledge base update unit: automatically updates the knowledge base and continuously improves the sentiment vocabulary.
[0059] Personalized learning unit: Customize personalized learning paths and response strategies based on the characteristics and preferences of different users.
[0060] Figure 2 Flowchart of the intelligent customer service sentiment analysis method based on machine learning in one embodiment of the present application; Figure 2 As shown, the present application provides an intelligent customer service sentiment analysis method based on machine learning, including: S1, the sentiment word library construction module obtains corpus information and constructs a sentiment word library based on the corpus information; the corpus information includes voice, text and pictures; S2, the user interaction module interacts with the user and collects the user's consultation data, the consultation data including text and voice; S3, the sentiment analysis module obtains the consultation data, performs sentiment analysis on the consultation data using a deep learning model, and determines the sentiment analysis result of the user; S4, a response generation module obtains the sentiment analysis result, generates corresponding response content according to the sentiment analysis result, and feeds back the response content to the user; and S5. The self-learning module optimizes the system in real time, learns user feedback information and consultation data of each user, and automatically updates the sentiment word library.
[0061] In some embodiments, constructing a sentiment word library based on the corpus information includes: Performing data preprocessing on the corpus information to obtain training corpus; The training corpus is sentimentally annotated to determine the sentiment results of unknown sentiment words.
[0062] In some embodiments, the sentiment tagging of the training corpus includes: Unknown sentiment word is S * , where W i is a Chinese character in the unknown sentiment word, and the unknown sentiment word is represented by W1W2…W n , i and n are integers; Each Chinese character W in the unknown sentiment word i The probability is: ; Among them, Fre(S *) represents the frequency of the position sentiment word in the training expectation, Fre(S * ,W i ) represents the Chinese character W i In the unknown sentiment word S * The frequency in The probability that the unknown sentiment word is a sentiment word is: ; Wherein, C represents the set of words in the training corpus, Z i represents a word in set C, Fre(C) represents the frequency of set C, Fre(Z i ) represents the word Z in the set C i frequency.
[0063] In some embodiments, determining the sentiment result of the unknown sentiment word includes: Based on each Chinese character W in the unknown sentiment word i The probability p(W i / S * ) and the probability p(S * ) Calculate the unknown sentiment word S * Emotional weight: ; Wherein, the sentiment weight of the unknown sentiment word includes a positive weight value and a negative weight value; The sentiment result of the unknown sentiment word is the difference between the positive weight value and the negative weight value. If the difference is a positive value, it indicates that the unknown sentiment word is a positive sentiment word. If the difference is not a positive value, it indicates that the unknown sentiment word is a negative sentiment word.
[0064] The machine learning-based intelligent customer service sentiment analysis method, system and electronic device provided in this application build a vertical field sentiment corpus through advanced natural language processing and machine learning technologies, achieve accurate identification and response to customer emotions, and automatically update the knowledge base based on customer feedback data, continuously optimize the sentiment analysis model and response accuracy, and improve the user experience.
[0065] It should be noted that the method of one or more embodiments of the present application can be performed by a single device, such as a computer or server. The method of this embodiment can also be applied to a distributed scenario and completed by multiple devices cooperating with each other. In the case of such a distributed scenario, one of the multiple devices can only perform one or more steps in the method of one or more embodiments of the present application, and the multiple devices will interact with each other to complete the described method.
[0066] It should be noted that the above describes specific embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0067] Based on the same inventive concept, corresponding to any of the above-mentioned embodiments and methods, the present application also discloses an electronic device; Specifically, Figure 3 The hardware structure diagram of an electronic device for an intelligent customer service sentiment analysis method based on machine learning provided in this embodiment is shown, and the device may include: a processor 410, a memory 420, an input / output interface 430, a communication interface 440, and a bus 450. The processor 410, the memory 420, the input / output interface 430, and the communication interface 440 are connected to each other in communication within the device through the bus 450.
[0068] The processor 410 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application.
[0069] The memory 420 may be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 420 may store an operating system and other application programs. When the technical solution provided in the embodiment of the present application is implemented by software or firmware, the relevant program code is stored in the memory 420 and is called and executed by the processor 410.
[0070] The input / output interface 430 is used to connect the input / output module to realize information input and output. The input / output module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.
[0071] The communication interface 440 is used to connect a communication module (not shown in the figure) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (for example, USB, network cable, etc.) or a wireless mode (for example, mobile network, WIFI, Bluetooth, etc.).
[0072] The bus 450 includes a path that transmits information between the various components of the device (eg, the processor 410 , the memory 420 , the input / output interface 430 , and the communication interface 440 ).
[0073] It should be noted that, although the above device only shows the processor 410, the memory 420, the input / output interface 430, the communication interface 440 and the bus 450, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include the components necessary for implementing the embodiment of the present application, and does not necessarily include all the components shown in the figure.
[0074] The electronic device of the above-mentioned embodiment is used to implement the corresponding method of the intelligent customer service sentiment analysis system based on machine learning in any of the aforementioned embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be repeated here.
[0075] Based on the same inventive concept, corresponding to any of the above-mentioned embodiment methods, one or more embodiments of the present application also provide a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the method involved in the intelligent customer service sentiment analysis system based on machine learning as described in any of the above embodiments.
[0076] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.
[0077] The computer instructions stored in the storage medium of the above embodiment are used to enable the computer to execute the method involved in the intelligent customer service sentiment analysis system based on machine learning as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.
[0078] A person skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present application (including the claims) is limited to these examples. In line with the concept of the present application, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of different aspects of one or more embodiments of the present application as described above, which are not provided in detail for the sake of simplicity.
[0079] In addition, to simplify the description and discussion, and in order not to make one or more embodiments of the present application difficult to understand, the well-known power / ground connections to the integrated circuit (IC) chip and other components may or may not be shown in the provided figures. In addition, the device may be shown in the form of a block diagram to avoid making one or more embodiments of the present application difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform on which one or more embodiments of the present application will be implemented (that is, these details should be fully within the scope of understanding of those skilled in the art). Where specific details (e.g., circuits) are set forth to describe exemplary embodiments of the present application, it is obvious to those skilled in the art that one or more embodiments of the present application can be implemented without these specific details or with changes in these specific details. Therefore, these descriptions should be considered illustrative rather than restrictive.
[0080] Although the present application has been described in conjunction with specific embodiments of the present application, many alternatives, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the discussed embodiments.
[0081] One or more embodiments of the present application are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of one or more embodiments of the present application should be included in the scope of protection of the present application.
Claims
1. An intelligent customer service sentiment analysis system based on machine learning, It is characterized in that The system comprises: The sentiment word library construction module is used to obtain corpus information and construct a sentiment word library based on the corpus information; the corpus information includes voice, text and pictures; A user interaction module, used to interact with the user and collect the user's consultation data, wherein the consultation data includes text and voice; A sentiment analysis module, used to obtain the consultation data, perform sentiment analysis on the consultation data using a deep learning model, and determine a sentiment analysis result of the user; a response generation module, used to obtain the sentiment analysis result, generate corresponding response content according to the sentiment analysis result, and feed back the response content to the user; and The self-learning module is used to optimize the system in real time, learn user feedback information and consultation data of each user, and automatically update the sentiment word library.
2. According to the machine learning-based intelligent customer service sentiment analysis system of claim 1, It is characterized in that The constructing of a sentiment word library based on the corpus information comprises: Performing data preprocessing on the corpus information to obtain training corpus; The training corpus is sentimentally annotated to determine the sentiment results of unknown sentiment words.
3. According to the machine learning-based intelligent customer service sentiment analysis system of claim 2, It is characterized in that The sentiment tagging of the training corpus includes: Unknown sentiment word is S * , where W i is a Chinese character in the unknown sentiment word, and the unknown sentiment word is represented by W 1 W 2 …W n , i and n are integers; Each Chinese character W in the unknown sentiment word i The probability is: ; Among them, Fre(S * ) represents unknown sentiment word S * The frequency in the training corpus, Fre(S * ,W i ) represents the Chinese character W i In the unknown sentiment word S * The frequency in The probability that the unknown sentiment word is a sentiment word is: ; Wherein, C represents the set of words in the training corpus, Z i represents a word in set C, Fre(C) represents the frequency of set C, Fre(Z i ) represents the word Z in the set C i frequency.
4. According to claim 3, the intelligent customer service sentiment analysis system based on machine learning, It is characterized in that Determine the emotional results of unknown emotional words, including: Based on each Chinese character W in the unknown sentiment word i The probability p(W i / S * ) and the probability p(S * ) Calculate the unknown sentiment word S * Emotional weight: ; Wherein, the sentiment weight of the unknown sentiment word includes a positive weight value and a negative weight value; The sentiment result of the unknown sentiment word is the difference between the positive weight value and the negative weight value. If the difference is a positive value, it indicates that the unknown sentiment word is a positive sentiment word. If the difference is not a positive value, it indicates that the unknown sentiment word is a negative sentiment word.
5. The intelligent customer service sentiment analysis system based on machine learning according to claim 1, It is characterized in that The user interaction module has an interactive interface, which includes a text input box, a voice input option, emoticons, a video voice collector, and a user video expression collector; The user interaction module further includes an intention recognition unit and an entity recognition unit, wherein the intention recognition unit determines the user's query intention, wherein the query intention includes consultation, complaint, and suggestion; The entity recognition unit extracts key information from user input, where the key information includes time, location, and product model.
6. The intelligent customer service sentiment analysis system based on machine learning according to claim 1, It is characterized in that The emotion analysis module includes a multimodal emotion recognition unit and a real-time analysis and feedback unit, wherein the multimodal emotion recognition unit uses deep learning technology to extract emotion features and rhythm features in speech, and distinguishes positive emotions, neutral emotions, and negative emotions contained in user speech; the real-time analysis and feedback unit performs real-time emotion analysis and feeds back the emotion analysis results to the intelligent customer service system so that the intelligent customer service system can provide corresponding services and responses according to the user's emotional state.
7. The intelligent customer service sentiment analysis system based on machine learning according to claim 1, It is characterized in that The self-learning module includes a complaint or error analysis unit, which identifies errors in the system's understanding of problems based on user feedback cases, corrects them and provides feedback.
8. According to claim 7, the intelligent customer service sentiment analysis system based on machine learning, It is characterized in that The self-learning module also includes a knowledge base updating unit, which automatically updates the emotional vocabulary in real time.
9. A method for intelligent customer service sentiment analysis based on machine learning. It is characterized in that The method comprises: The emotional word library construction module obtains corpus information and constructs an emotional word library based on the corpus information; the corpus information includes voice, text and pictures; The user interaction module interacts with the user and collects the user's consultation data, wherein the consultation data includes text and voice; The sentiment analysis module obtains the consultation data, performs sentiment analysis on the consultation data using a deep learning model, and determines a sentiment analysis result of the user; The response generation module obtains the sentiment analysis result, generates corresponding response content according to the sentiment analysis result, and feeds back the response content to the user; and The self-learning module optimizes the system in real time, learns user feedback information and consultation data of each user, and automatically updates the sentiment word library.
10. The intelligent customer service sentiment analysis method based on machine learning according to claim 9, It is characterized in that The constructing of a sentiment word library based on the corpus information comprises: Performing data preprocessing on the corpus information to obtain training corpus; The training corpus is sentimentally annotated to determine the sentiment results of unknown sentiment words.
11. The intelligent customer service sentiment analysis method based on machine learning according to claim 10, It is characterized in that The sentiment tagging of the training corpus includes: Unknown sentiment word is S * , where W i is a Chinese character in the unknown sentiment word, and the unknown sentiment word is represented by W 1 W 2 …W n , i and n are integers; Each Chinese character W in the unknown sentiment word i The probability is: ; Among them, Fre(S * ) represents the frequency of the position sentiment word in the training expectation, Fre(S * ,W i ) represents the Chinese character W i In the unknown sentiment word S * The frequency in The probability that the unknown sentiment word is a sentiment word is: ; Where C represents the set of words in the training corpus, Z i represents a word in the set C, Fre(C) represents the frequency of the set, and Fre(Z i ) represents the word Z in the set i frequency.
12. The intelligent customer service sentiment analysis method based on machine learning according to claim 11, It is characterized in that Determine the emotional results of unknown emotional words, including: Based on each Chinese character W in the unknown sentiment word i The probability p(W i / S * ) and the probability p(S * ) Calculate the unknown sentiment word S * Emotional weight: ; Wherein, the sentiment weight of the unknown sentiment word includes a positive weight value and a negative weight value; The sentiment result of the unknown sentiment word is the difference between the positive weight value and the negative weight value. If the difference is a positive value, it indicates that the unknown sentiment word is a positive sentiment word. If the difference is not a positive value, it indicates that the unknown sentiment word is a negative sentiment word.
13. An electronic device, the electronic device include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 9 to 12 when executing the computer program.