Emotion perception-based internet of things information technology customer service method and system

Abstract

The application provides an emotion perception-based Internet of Things information technology customer service method and system, and relates to the technical field of information technology services.The method comprises the following steps: obtaining real-time message data of a user and a customer service, and transmitting the data to a cloud server in real time through an Internet of Things platform; performing text analysis on the obtained real-time message data, and analyzing the current emotional state of the user through an emotion analysis method; analyzing the intention of the user according to the real-time message data, and generating reply content through a reply database according to the current emotional state of the user; continuously monitoring the interaction data between the user and the customer service, and dynamically adjusting the reply content.The application can analyze the emotional state of the user through emotion perception technology, thereby better understanding the demand and emotional state of the user, adjusting the reply content according to the emotional state of the user, and providing more personalized, warm and considerate services that are closer to the psychological demand of the user.

Description

Technical Field

[0001] This invention relates to the field of information technology service technology, and more specifically, to an Internet of Things (IoT) information technology customer service method and system based on emotion perception. Background Technology

[0002] The information technology services industry, as a core component of the IT field, is committed to providing clients with comprehensive technical support and services, including software development, hardware installation, and system maintenance. These services not only focus on solving existing technical problems but also broadly cover multiple dimensions such as consulting, management, operations, and security. They aim to help users achieve their business goals and maximize value by improving the overall performance and efficiency of IT systems. The core value of the IT services industry lies in its ability to provide customized solutions for enterprises and organizations, enabling them to fully utilize existing resources and the latest technological trends, thereby standing out in market competition. With the rapid development of emerging technologies such as cloud computing, big data, artificial intelligence, and the Internet of Things, the IT services industry is facing unprecedented opportunities and challenges. These technologies have not only changed business operation models but also greatly increased enterprises' demand for IT services. Therefore, IT service providers need to continuously update their service content and technical capabilities to adapt to the pace of technological progress and changes in market demands.

[0003] Customer service is a crucial area aimed at enhancing customer satisfaction and loyalty through effective communication and problem-solving. This includes not only resolving customer problems and questions but also providing product information, assisting customers in making purchasing decisions, and handling orders and returns. While technological advancements have enabled customer service to exist and be provided to users in various forms, existing customer service methods still have some significant shortcomings: traditional methods often struggle to capture and analyze users' emotional states in real time, lacking effective mechanisms to identify emotional changes during interactions. This can lead to service responses that do not match the user's actual emotional state, negatively impacting the user experience.

[0004] There are currently no effective solutions to the problems in the relevant technologies. Summary of the Invention

[0005] In view of this, the present invention provides an IoT information technology customer service method and system based on emotion perception to solve the problem mentioned above of the lack of an effective mechanism to identify users' emotional changes during the interaction process.

[0006] To solve the above problems, the specific technical solution adopted by the present invention is as follows: According to one aspect of the present invention, an emotion-aware Internet of Things (IoT) information technology customer service method is provided, the emotion-aware IoT information technology customer service method comprising the following steps: S1. Use IoT devices to acquire real-time message data between users and customer service, and transmit it to the cloud server in real time through the IoT platform. The real-time message data is the text information exchanged between users and customer service. S2. Perform text analysis on the acquired real-time message data and analyze the user's current emotional state using sentiment analysis methods; S3. Analyze the user's intent and purpose based on real-time message data, and generate reply content through the reply database based on the user's current emotional state; S4. Utilize the IoT platform to push the generated response content back to the user's IoT device, and continuously monitor the interaction data between the user and customer service to dynamically adjust the response content.

[0007] Preferably, the process of performing text analysis on the acquired real-time message data and analyzing the user's current emotional state using sentiment analysis methods includes the following steps: S21. Preprocess the acquired real-time message data, including text cleaning, stop word removal and word segmentation. S22. Construct an emotion topic model to analyze the preprocessed real-time message data, extract the emotion topic distribution in the real-time message data, and establish a text-emotion topic matrix; S23. Based on the established text-emotion topic matrix, the importance of emotion topics is analyzed using the neighborhood rough set reduction algorithm, and the emotion topics are reduced according to the importance analysis results. S24. Based on the reduced emotion theme, use the emotion dictionary algorithm to analyze the user's current emotional state.

[0008] Preferably, constructing a sentiment topic model to analyze preprocessed real-time message data and extracting the sentiment topic distribution from the real-time message data to establish a text-sentiment topic matrix includes the following steps: S221. Collect historical sample message data, analyze the historical sample message data using syntactic analysis, and extract emotional feature words and emotional expression words; S222. Cluster the emotion feature words and emotion opinion words using the similarity measurement method, and filter the clustering results using the mutual information method to obtain the emotion feature word set and emotion opinion word set with semantic redundancy eliminated. S223. Construct an emotion corpus based on the filtered emotion feature word set and emotion expression word set, and train a pre-set LDA topic model based on the emotion corpus to obtain the emotion topic model; S224. Analyze the preprocessed real-time message data based on the obtained sentiment topic model, extract the sentiment topic distribution, and establish a text-sentiment topic matrix.

[0009] Preferably, the process of clustering emotion feature words and emotion opinion words using a similarity measurement method, and then filtering the clustering results using mutual information to obtain a set of emotion feature words and a set of emotion opinion words with semantic redundancy eliminated, includes the following steps: S2221. Calculate the similarity between emotion feature words and emotion opinion words based on semantic similarity and opinion similarity; S2222. Based on the similarity measurement results, a clustering algorithm is used to cluster emotional feature words and emotional opinion words respectively, and the frequency of occurrence of words in each cluster is counted. S2223. Calculate the correlation between words and sentiment expression in each cluster using the mutual information method, and select clusters that meet the preset correlation threshold. S2224. Merge and optimize the filtered clustering results to eliminate semantically repetitive emotion feature words and emotion opinion words, and obtain a set of emotion feature words and a set of emotion opinion words with semantically repetitive words eliminated.

[0010] Preferably, based on the established text-emotion topic matrix, the importance of emotion topics is analyzed using a neighborhood rough set reduction algorithm, and the emotion topics are reduced according to the importance analysis results, including the following steps: S231. The established text-emotion topic matrix is ​​converted into a topic decision engine, where the emotion topic is used as a conditional attribute, each text is used as a decision rule, and the category to which each text belongs is used as a decision attribute. S232. Calculate the positive domain of each emotion topic using the neighborhood rough set method, and calculate the importance of each emotion topic based on the calculation results of the positive domain. S233. Set an importance threshold, compare the importance of each emotion topic with the set importance threshold, and retain the emotion topics that meet the importance threshold to achieve the reduction of emotion topics.

[0011] Preferably, based on the reduced emotional topic, the analysis of the user's current emotional state using an emotional dictionary algorithm includes the following steps: S241. Construct an emotion dictionary using a basic emotion dictionary, a degree adverb dictionary, a negation word dictionary, an interjection word dictionary, a rhetorical question word dictionary, and a relational conjunction word dictionary; S242. Import the reduced emotion theme into the emotion dictionary and identify the emotion words in the reduced emotion theme. The emotion words include basic emotion words, degree adverbs, negation words, interjections, rhetorical questions and relational conjunctions. S243. Assign weights to the identified sentiment words based on semantic rules, and calculate the sentiment value of the sentiment topic. S244. Determine the user's emotional state based on the calculated emotion value.

[0012] Preferably, the formula for calculating the sentiment value of an emotional topic is as follows: ; In the formula, V represents the sentiment value of the emotional topic; n represents the number of basic sentiment words; m indicates the number of interjections; r indicates the number of negative words; W i This represents the weight of the i-th basic sentiment word; N i This represents the weight of the i-th interjection; D indicates the weight of the degree adverb; X represents the negation gate; it takes a value of 1 if a negation word is present, and a value of 0 otherwise. Y represents the gate for rhetorical rhetorical questions. If a rhetorical rhetorical question appears, the value is 1; otherwise, the value is 0. H represents the influence coefficient; W represents the weight of the current basic sentiment word.

[0013] Preferably, the process of analyzing user intent based on real-time message data and generating response content from a response database based on the user's current emotional state includes the following steps: S31. Collect message text data labeled with user intent and extract features from the message text data; S32. An intent recognition model based on a RoBERTa pre-trained model is used to train the extracted features and then use the trained intent recognition model to identify the user's intent based on the extracted features. S33. Based on the user's intent and purpose, select a response template from the response database that matches the user's intent and purpose; S34. Adjust the tone and content of the reply template according to the user's current emotional state to obtain the final reply content.

[0014] Preferably, selecting a response template from the response database that matches the user's intent or purpose includes the following steps: S331. Collect user intent questions and customer service responses through historical human customer service records; S332. Utilize natural language processing technology to perform text analysis and semantic understanding on intent questions and customer service responses, and extract key information and intent tags from the responses; S333. Based on the key information of the response, construct a response template for each intent tag, and establish a relationship between the response template and the intent tag to build a response database; S334. Match the user's intent with the intent tags in the response database based on similarity. S335. Based on the similarity matching results, select the response template associated with the intent tag with the highest similarity as the response template that matches the user's intent.

[0015] According to another aspect of the present invention, an IoT information technology customer service system based on emotion perception is provided. The IoT information technology customer service system based on emotion perception includes: a data acquisition module, an emotion analysis module, a response content generation module, and a response content push module, and the data acquisition module, the emotion analysis module, the response content generation module, and the response content push module are connected sequentially. The data acquisition module is used to acquire real-time message data between users and customer service using IoT devices, and transmit it to the cloud server in real time through the IoT platform. The real-time message data is the text information exchanged between users and customer service. The sentiment analysis module is used to perform text analysis on the acquired real-time message data and analyze the user's current emotional state through sentiment analysis methods. The reply content generation module is used to analyze the user's intent and purpose based on real-time message data, and generate reply content from the reply database based on the user's current emotional state. The response content push module is used to push the generated response content back to the user's IoT device using the IoT platform, and continuously monitor the interaction data between the user and customer service to dynamically adjust the response content.

[0016] The beneficial effects of this invention are as follows: 1. This invention utilizes emotion perception technology to analyze users' emotional states, thereby better understanding their needs and emotional states. Adjusting responses based on these emotional states allows for a closer alignment with users' psychological needs, providing more personalized, warm, and considerate services, thus improving customer satisfaction. By leveraging IoT devices and cloud servers, customer service is automated, enabling emotion perception and service responses without human intervention. This saves labor costs, improves service efficiency, and achieves automated and intelligent customer service.

[0017] 2. This invention, through text and sentiment analysis of real-time message data, can accurately identify users' emotional states and intentions, helping to better understand users' needs and emotional states. This provides an accurate foundation for generating subsequent response content. By establishing a text-emotion topic matrix and reducing the emotional topics, the invention can more accurately grasp users' emotional topics. Thus, based on the user's current emotional state, targeted response content can be generated using an emotional dictionary algorithm. Through the analysis of the importance of emotional topics and the calculation of sentiment values, the invention can discover the topics and emotional tendencies that users care about more, helping to continuously improve its intelligence level and service quality, and better meet users' needs and expectations.

[0018] 3. This invention analyzes historical customer service records and intent-based questions to create response templates that encompass various intents. By utilizing natural language processing (NLP) technology for text analysis and semantic understanding, it can extract key information and intent tags from the responses, thereby constructing response templates that match the user's intent. This helps to accurately select responses that match the user's intent from a large number of response templates. By selecting a matching response template based on the user's intent, personalized response content can be generated to meet the personalized requirements of different user needs, thus improving user experience and enhancing user satisfaction. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly described below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 This is a flowchart of an IoT information technology customer service method based on emotion perception according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a customer service system based on emotion perception in the Internet of Things (IoT) according to an embodiment of the present invention.

[0020] In the picture: 1. Data acquisition module; 2. Sentiment analysis module; 3. Response content generation module; 4. Response content push module. Detailed Implementation

[0021] To enable those skilled in the art to better understand the technical solutions in this application, the following will clearly and completely describe the technical solutions in the embodiments of this application in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of this application.

[0022] According to an embodiment of the present invention, there is provided an Internet of Things information technology customer service method and system based on emotion perception.

[0023] The present invention will be further described in conjunction with the accompanying drawings and specific embodiments. As Figure 1 shown, according to an embodiment of the present invention, there is provided an Internet of Things information technology customer service method based on emotion perception. The Internet of Things information technology customer service method based on emotion perception includes the following steps: S1. Use Internet of Things devices to obtain real-time message data between users and customer service, and transmit it to the cloud server in real time through the Internet of Things platform. The real-time message data is the text information exchanged between users and customer service; Specifically, the Internet of Things devices include smartphones, smart computers, etc.; when a user communicates with customer service, the Internet of Things devices can capture real-time message data and transmit it to the cloud server in real time through the Internet of Things platform. Among them, the Internet of Things platform provides a data transmission channel and protocol to ensure that data can be transmitted to the specified cloud server safely and stably; usually, the Internet of Things platform will provide various connection options, such as Wi-Fi, cellular network, LoRaWAN, etc., to meet the needs of different devices and scenarios.

[0024] S2. Perform text analysis on the obtained real-time message data, and analyze the current emotional state of the user through an emotion analysis method; As a preferred embodiment, performing text analysis on the obtained real-time message data and analyzing the current emotional state of the user through an emotion analysis method includes the following steps: S21. Preprocess the obtained real-time message data. The preprocessing includes text cleaning, stop word removal, and word segmentation; It should be noted that the real-time message data may contain various noises and unnecessary contents, such as HTML tags, special characters, punctuation marks, etc. In the text cleaning stage, these irrelevant contents need to be removed, and only the text information is retained; Stop words refer to words that frequently appear in the text but do not contribute substantially to text analysis, such as "of", "is", "in", etc.; in text analysis, these stop words are usually removed from the text to reduce the data volume and improve the analysis efficiency and accuracy.

[0025] Word segmentation is the process of dividing text data into words or phrases that have practical meaning. Word segmentation divides text data into a sequence of words or phrases, providing a foundation for subsequent analysis and other tasks.

[0026] S22. Construct an emotion topic model to analyze the preprocessed real-time message data, extract the emotion topic distribution in the real-time message data, and establish a text-emotion topic matrix; As a preferred implementation, constructing an emotion topic model to analyze preprocessed real-time message data and extracting the emotion topic distribution from the real-time message data to establish a text-emotion topic matrix includes the following steps: S221. Collect historical sample message data, analyze the historical sample message data using syntactic analysis, and extract emotional feature words and emotional expression words; Specifically, the analysis of historical sample message data using syntactic analysis to extract sentiment feature words and sentiment expression words includes the following steps: The historical sample message data is segmented into a sequence of individual sentences using a sentence segmenter; Each sentence is divided into a sequence of words or phrases, and each word is labeled with its part of speech, that is, the grammatical role of the word in the sentence. Part of speech labeling can help identify the subject, predicate, object and other components in the sentence. Dependency parsing is used to analyze the dependency relationships between words in a sentence, that is, the subordinate relationships between words. Based on the syntactic analysis, semantic analysis is performed to understand the meaning and connotation of the sentence. Based on semantic analysis, emotion feature words and emotion expression words are identified in the text. Emotion feature words are usually keywords that express feelings and emotions, such as happy, angry, sad, etc. Emotion expression words are words used to express feelings and emotions, such as like, angry, sad, etc.

[0027] S222. Cluster the emotion feature words and emotion opinion words using the similarity measurement method, and filter the clustering results using the mutual information method to obtain the emotion feature word set and emotion opinion word set with semantic redundancy eliminated. As a preferred implementation, clustering emotion feature words and emotion opinion words using a similarity measurement method, and filtering the clustering results using mutual information to obtain a set of emotion feature words and a set of emotion opinion words with semantic redundancy eliminated, includes the following steps: S2221. Calculate the similarity between emotion feature words and emotion opinion words based on semantic similarity and opinion similarity; It should be noted that by using knowledge bases such as Hownet, the semantic relationship and path distance between sentiment feature words and sentiment opinion words are calculated to obtain their semantic similarity. When calculating the similarity between emotion feature words and emotion opinion words, in addition to considering semantic similarity, it is also necessary to consider their opinion similarity. Opinion similarity can be calculated based on the degree of similarity of words in terms of emotional polarity, that is, whether the polarity of the emotion words is similar. For example, if two words both express positive emotions or both express negative emotions, their opinion similarity is high; if one word expresses positive emotions and the other expresses negative emotions, their opinion similarity is low.

[0028] Finally, by comprehensively considering semantic similarity and opinion similarity, a weighted sum is calculated on the semantic similarity and opinion similarity to obtain the final similarity value.

[0029] S2222. Based on the similarity measurement results, a clustering algorithm is used to cluster emotional feature words and emotional opinion words respectively, and the frequency of occurrence of words in each cluster is counted. It should be noted that a similarity matrix is ​​constructed based on the calculated similarity measurement results between emotional feature words and emotional opinion words. Then, hierarchical clustering is used to cluster emotional feature words and emotional opinion words, which can assign emotional feature words and emotional opinion words to different clusters. By traversing each cluster and calculating the occurrence frequency of each word in it, the degree of association between words in each cluster and emotional expression is determined by correlation analysis.

[0030] S2223. Calculate the correlation between words and sentiment expression in each cluster using the mutual information method, and select clusters that meet the preset correlation threshold. It should be noted that mutual information is an indicator used to measure the correlation between two random variables. It reflects the amount of information contained in one random variable about another.

[0031] For each word in each cluster, the mutual information between it and the sentiment expression is calculated. Then, the words are sorted according to their mutual information values ​​with the sentiment expression. Based on a preset relevance threshold, the clusters containing words with mutual information values ​​greater than or equal to the threshold are selected.

[0032] S2224. Merge and optimize the filtered clustering results to eliminate semantically repetitive emotion feature words and emotion opinion words, and obtain a set of emotion feature words and a set of emotion opinion words with semantically repetitive words eliminated.

[0033] It should be noted that for the merged clusters, we check whether there are duplicate emotion feature words and emotion opinion words. If duplicates are found, we retain only one of them to eliminate semantic redundancy.

[0034] S223. Construct an emotion corpus based on the filtered emotion feature word set and emotion expression word set, and train a pre-set LDA topic model based on the emotion corpus to obtain the emotion topic model; It should be noted that by using the filtered set of emotion feature words and the set of emotion expression words as a vocabulary, an emotion corpus is constructed. Then, each word sample is represented as a bag-of-words model, and the bag-of-words model is used as the input to the LDA model to train the LDA topic model, and finally, the emotion topic model is obtained.

[0035] S224. Analyze the preprocessed real-time message data based on the obtained sentiment topic model, extract the sentiment topic distribution, and establish a text-sentiment topic matrix.

[0036] Specifically, the preprocessed real-time message data is converted into a bag-of-words model and loaded into the sentiment topic model, which can capture different sentiment topics and extract the sentiment topic distribution of each message. The matrix is ​​initialized according to the number of topics in the sentiment topic model. The rows of the matrix represent different texts (real-time messages), and the columns represent different sentiment topics. The matrix is ​​filled according to the sentiment topic distribution of each message.

[0037] S23. Based on the established text-emotion topic matrix, the importance of emotion topics is analyzed using the neighborhood rough set reduction algorithm, and the emotion topics are reduced according to the importance analysis results. As a preferred implementation, based on the established text-emotion topic matrix, the importance of emotion topics is analyzed using a neighborhood rough set reduction algorithm, and the emotion topics are reduced according to the importance analysis results, including the following steps: S231. The established text-emotion topic matrix is ​​converted into a topic decision engine, where the emotion topic is used as a conditional attribute, each text is used as a decision rule, and the category to which each text belongs is used as a decision attribute. It should be noted that the topic decision engine, also known as the topic decision system, treats topics as conditional attributes, each text in the corpus as a decision rule, and the category to which each text belongs as a decision attribute. It can be seen that different topics have different levels of importance to texts of different categories. The relative reduction of knowledge in the topic decision system refers to removing redundant topics while keeping the dependency relationship between text topics and text categories unchanged in the original topic decision table.

[0038] Emotional themes refer to the primary emotions or sentiment categories identified in sentiment analysis, while text refers to the raw message data from users or documents. Emotional themes can be defined according to the needs of sentiment analysis, such as positive, negative, neutral, etc.

[0039] S232. Calculate the positive domain of each emotion topic using the neighborhood rough set method, and calculate the importance of each emotion topic based on the calculation results of the positive domain. It should be noted that in neighborhood rough set theory, the positive domain refers to the set of objects that can be accurately classified. For each emotion topic, its positive domain is calculated, that is, the set of texts that can be clearly classified into a specific emotion is found. The importance of an emotion topic refers to the degree of its role in the classification or decision-making process. By analyzing the influence of each emotion topic on text classification, the importance of each emotion topic can be calculated.

[0040] Specifically, the formula for calculating importance is: ; In the formula, G represents the importance of each emotional theme; This indicates that the emotional theme TC K The size of the positive domain following the emotion theme set RED; This represents the size of the positive domain under the initial set of emotion topics RED; D represents the decision attribute; S233. Set an importance threshold, compare the importance of each emotion topic with the set importance threshold, and retain the emotion topics that meet the importance threshold to achieve the reduction of emotion topics.

[0041] S24. Based on the reduced emotion theme, use the emotion dictionary algorithm to analyze the user's current emotional state.

[0042] As a preferred implementation, analyzing the user's current emotional state using an emotion dictionary algorithm based on the reduced emotion topic includes the following steps: S241. Construct an emotion dictionary using a basic emotion dictionary, a degree adverb dictionary, a negation word dictionary, an interjection word dictionary, a rhetorical question word dictionary, and a relational conjunction word dictionary; S242. Import the reduced emotion theme into the emotion dictionary and identify the emotion words in the reduced emotion theme. The emotion words include basic emotion words, degree adverbs, negation words, interjections, rhetorical questions and relational conjunctions. It should be noted that by traversing the words in the sentiment dictionary and matching them with the reduced sentiment themes, if a sentiment word appears in a sentiment theme, then the word is considered a sentiment word.

[0043] S243. Assign weights to the identified sentiment words based on semantic rules, and calculate the sentiment value of the sentiment topic. It should be noted that a set of semantic rules is formulated to determine the weight of each emotional word. These rules include the emotional polarity, emotional intensity, and association with other emotional words. According to the defined semantic rules, each emotional word is assigned a corresponding weight. These weights can be numerical and represent the intensity or degree of emotion. For example, positive emotional words have positive weights, while negative emotional words have negative weights.

[0044] Specifically, the formula for calculating the sentiment value of an emotional topic is as follows: ; In the formula, V represents the sentiment value of the emotional topic; n represents the number of basic sentiment words; m indicates the number of interjections; r indicates the number of negative words; W i This represents the weight of the i-th basic sentiment word; N i This represents the weight of the i-th interjection; D indicates the weight of the degree adverb; X represents the negation gate; it takes a value of 1 if a negation word is present, and a value of 0 otherwise. Y represents the gate for rhetorical rhetorical questions. If a rhetorical rhetorical question appears, the value is 1; otherwise, the value is 0. H represents the influence coefficient; W represents the weight of the current basic sentiment word.

[0045] S244. Determine the user's emotional state based on the calculated emotion value.

[0046] It should be noted that determining a user's emotional state based on the calculated sentiment score includes the following steps: The range of sentiment values ​​is divided into several intervals, for example, negative values ​​represent negative emotions, positive values ​​represent positive emotions, and values ​​close to zero represent neutral emotions; Set a threshold for the emotional value based on the actual situation. For example, a value greater than 0.5 indicates positive emotions, less than -0.5 indicates negative emotions, and a value between -0.5 and 0.5 indicates neutral emotions. S3. Analyze the user's intent and purpose based on real-time message data, and generate reply content through the reply database based on the user's current emotional state; As a preferred implementation, analyzing the user's intent based on real-time message data and generating response content from a response database based on the user's current emotional state includes the following steps: S31. Collect message text data labeled with user intent and extract features from the message text data; Specifically, feature extraction is the process of converting text data into feature representations that can be used for model training and analysis.

[0047] S32. An intent recognition model based on a RoBERTa pre-trained model is used to train the extracted features and then use the trained intent recognition model to identify the user's intent based on the extracted features. It should be noted that by using the aforementioned extracted features as input to the training set, a RoBERTa pre-trained model is loaded using a deep learning framework. An intent recognition model is then built on top of the RoBERTa model, typically involving adding additional fully connected layers or other appropriate structures to adapt to the specific intent recognition task. The output of this layer is usually a score used to classify user intent. The intent recognition model is then trained using the prepared training set data. During training, appropriate loss functions (such as cross-entropy loss) and optimizers (such as Adam) are used to optimize the model parameters. Finally, the user's real-time message data is input into the trained intent recognition model to obtain the corresponding intent label or probability distribution, thereby achieving the recognition of the user's intent purpose.

[0048] S33. Based on the user's intent and purpose, select a response template from the response database that matches the user's intent and purpose; As a preferred implementation, selecting a response template from the response database that matches the user's intent or purpose includes the following steps: S331. Collect user intent questions and customer service responses through historical human customer service records; Specifically, historical records of human customer service are obtained, and then AI assistants or customer service personnel are used to classify and label the intent of each conversation, categorizing the questions raised by users and the responses from customer service representatives into different intent categories.

[0049] S332. Utilize natural language processing technology to perform text analysis and semantic understanding on intent questions and customer service responses, and extract key information and intent tags from the responses; Specifically, the process of using natural language processing (NLP) technology to perform text analysis and semantic understanding on intent questions and customer service responses, and to extract key information and intent tags, includes the following steps: The intent questions and customer service responses are segmented into sequences of words or phrases, each word is labeled with its part of speech, and the grammatical structure of the sentences is analyzed. Identify named entities in text and the relationships between entities in text; The Naive Bayes classification model is used to classify intent questions to determine the intent category to which each question belongs; Perform semantic analysis on customer service responses to understand whether the customer service representative's answer matches the user's question; Identify keywords and phrases in the text and extract key information related to intent. Keywords and phrases can reflect the theme or content of the text, and key information includes service name, service requirements, etc., to help answer questions or perform subsequent actions.

[0050] S333. Based on the key information of the response, construct a response template for each intent tag, and establish a relationship between the response template and the intent tag to build a response database; It should be noted that, based on the key information of the response, a corresponding response template is designed for each intent tag. The response template contains general information that can answer a specific intent. Each response template is associated with the corresponding intent tag so that the appropriate response can be selected according to the user's intent. All associated response templates and intent tags are stored in the database to obtain the response database.

[0051] S334. Match the user's intent with the intent tags in the response database based on similarity. Specifically, the cosine similarity method is used to compare the similarity between the user's proposed intent and the intent tags in each response database, and a similarity score list is generated, which shows the degree of similarity between each intent tag and the user's intent.

[0052] S335. Based on the similarity matching results, select the response template associated with the intent tag with the highest similarity as the response template that matches the user's intent.

[0053] S34. Adjust the tone and content of the reply template according to the user's current emotional state to obtain the final reply content.

[0054] Specifically, the tone and content of the reply template are adjusted according to the user's current emotional state, resulting in the following final reply content: Adjust the tone of your reply based on the user's emotional state. For example, for an angry user, the tone of your reply may need to be more gentle and understanding; while for a happy user, the tone of your reply can be more lively and positive.

[0055] For users who are expressing positive emotions, reinforce those positive emotions by replying with words of praise or encouragement.

[0056] For users exhibiting negative emotions, strategies can be employed to alleviate these emotions, such as providing specific assistance, expressing sympathy, or offering discounts.

[0057] S4. Utilize the IoT platform to push the generated response content back to the user's IoT device, and continuously monitor the interaction data between the user and customer service to dynamically adjust the response content.

[0058] Specifically, the generated response content is pushed back to the user's IoT device using an IoT platform, and the interaction data between the user and customer service is continuously monitored to dynamically adjust the response content, including the following steps: We continuously collect interaction data between users and customer service representatives, and analyze users' reactions to responses by monitoring their emotional changes in real time. Based on the collected interaction data and user feedback, analyze the effectiveness of the response content and determine whether adjustments are needed; By using machine learning and automated rules, the tone and content of the response template are dynamically adjusted based on user feedback and emotional changes. Through continuous monitoring and adjustment, the response content is constantly optimized to improve user satisfaction and interaction efficiency.

[0059] like Figure 2 As shown, according to another embodiment of the present invention, an IoT information technology customer service system based on emotion perception is provided. The IoT information technology customer service system based on emotion perception includes: a data acquisition module 1, an emotion analysis module 2, a reply content generation module 3, and a reply content push module 4, and the data acquisition module 1, the emotion analysis module 2, the reply content generation module 3, and the reply content push module 4 are connected in sequence. Data acquisition module 1 is used to acquire real-time message data between users and customer service using IoT devices, and transmit it to the cloud server in real time through the IoT platform. The real-time message data is the text information exchanged between users and customer service. The sentiment analysis module 2 is used to perform text analysis on the acquired real-time message data and analyze the user's current emotional state through sentiment analysis methods. The reply content generation module 3 is used to analyze the user's intent and purpose based on real-time message data, and generate reply content through the reply database based on the user's current emotional state. The response content push module 4 is used to push the generated response content back to the user's IoT device using the IoT platform, and continuously monitor the interaction data between the user and customer service to dynamically adjust the response content.

[0060] In summary, by utilizing the above-mentioned technical solutions of this invention, this invention, through emotion perception technology, can analyze the user's emotional state, thereby better understanding the user's needs and emotional state. Adjusting the response content according to the user's emotional state allows for a closer alignment with the user's psychological needs, providing more personalized, warm, and considerate service, thus improving customer satisfaction. Utilizing IoT devices and cloud servers, it automates customer service, enabling emotion perception and service response without manual intervention. This saves labor costs, improves service efficiency, and achieves automation and intelligence in customer service. Furthermore, by performing text and emotion analysis on real-time message data, this invention can accurately identify the user's emotional state and intentions, helping to better understand the user's needs and emotional state, providing an accurate foundation for subsequent response content generation. Establishing a text-emotion topic matrix and performing emotion topic reduction processing allows for more precise... By accurately grasping the user's emotional theme, this invention can generate targeted responses based on the user's current emotional state using an emotional dictionary algorithm. Through importance analysis and sentiment value calculation of emotional themes, it can discover topics and emotional tendencies that users care about more, helping to continuously improve its intelligence level and service quality, and better meet user needs and expectations. This invention can also analyze historical human customer service records and intent questions to create response templates containing various intents. By using natural language processing technology for text analysis and semantic understanding, it can extract key information and intent tags for responses, thereby constructing response templates that match the user's intent. This helps to accurately select responses that match the user's intent from a large number of response templates. By selecting a matching response template based on the user's intent and purpose, personalized response content can be generated to meet the personalized requirements of different user needs, thus improving user experience and enhancing user satisfaction.

[0061] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, etc.) containing computer-usable program code.

[0062] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A customer service method based on emotion perception using Internet of Things (IoT) information technology, characterized in that, This emotion-aware IoT-based customer service method includes the following steps: S1. Use IoT devices to obtain real-time message data between users and customer service representatives, and transmit it to the cloud server in real time through the IoT platform. The real-time message data is the text information exchanged between users and customer service representatives. S2. Perform text analysis on the acquired real-time message data and analyze the user's current emotional state using sentiment analysis methods; S3. Analyze the user's intent and purpose based on real-time message data, and generate reply content through the reply database based on the user's current emotional state; S4. Utilize the IoT platform to push the generated response content back to the user's IoT device, and continuously monitor the interaction data between the user and customer service to dynamically adjust the response content. The process of performing text analysis on the acquired real-time message data and analyzing the user's current emotional state using sentiment analysis methods includes the following steps: S21. Preprocess the acquired real-time message data, including text cleaning, stop word removal and word segmentation. S22. Construct an emotion topic model to analyze the preprocessed real-time message data, extract the emotion topic distribution in the real-time message data, and establish a text-emotion topic matrix; S23. Based on the established text-emotion topic matrix, the importance of emotion topics is analyzed using the neighborhood rough set reduction algorithm, and the emotion topics are reduced according to the importance analysis results. S24. Based on the reduced emotional theme, use the emotional dictionary algorithm to analyze the user's current emotional state; The process of constructing a sentiment topic model involves analyzing preprocessed real-time message data, extracting the distribution of sentiment topics from the real-time message data, and establishing a text-sentiment topic matrix, including the following steps: S221. Collect historical sample message data, analyze the historical sample message data using syntactic analysis, and extract emotional feature words and emotional expression words; S222. Cluster the emotion feature words and emotion opinion words using the similarity measurement method, and filter the clustering results using the mutual information method to obtain the emotion feature word set and emotion opinion word set with semantic redundancy eliminated. S223. Construct an emotion corpus based on the filtered emotion feature word set and emotion expression word set, and train a pre-set LDA topic model based on the emotion corpus to obtain the emotion topic model; S224. Analyze the preprocessed real-time message data based on the obtained sentiment topic model, extract the sentiment topic distribution, and establish a text-sentiment topic matrix.

2. The IoT information technology customer service method based on emotion perception according to claim 1, characterized in that, The process of clustering emotion feature words and emotion opinion words using a similarity measurement method, and then filtering the clustering results using mutual information to obtain a set of emotion feature words and an emotion opinion word set with semantic redundancy eliminated, includes the following steps: S2221. Calculate the similarity between emotion feature words and emotion opinion words based on semantic similarity and opinion similarity; S2222. Based on the similarity measurement results, a clustering algorithm is used to cluster emotional feature words and emotional opinion words respectively, and the frequency of occurrence of words in each cluster is counted. S2223. Calculate the correlation between words and sentiment expression in each cluster using the mutual information method, and select clusters that meet the preset correlation threshold. S2224. Merge and optimize the filtered clustering results to eliminate semantically repetitive emotion feature words and emotion opinion words, and obtain a set of emotion feature words and a set of emotion opinion words with semantically repetitive words eliminated.

3. The customer service method based on emotion perception in the Internet of Things (IoT) information technology according to claim 2, characterized in that, The process of analyzing the importance of emotion topics based on the established text-emotion topic matrix, using a neighborhood rough set reduction algorithm, and then reducing the emotion topics according to the importance analysis results includes the following steps: S231. The established text-emotion topic matrix is ​​converted into a topic decision engine, where the emotion topic is used as a conditional attribute, each text is used as a decision rule, and the category to which each text belongs is used as a decision attribute. S232. Calculate the positive domain of each emotion topic using the neighborhood rough set method, and calculate the importance of each emotion topic based on the calculation results of the positive domain. S233. Set an importance threshold, compare the importance of each emotion topic with the set importance threshold, and retain the emotion topics that meet the importance threshold to achieve the reduction of emotion topics.

4. The IoT information technology customer service method based on emotion perception according to claim 2, characterized in that, The process of analyzing a user's current emotional state based on the reduced emotional topic and using an emotional dictionary algorithm includes the following steps: S241. Construct an emotion dictionary using a basic emotion dictionary, a degree adverb dictionary, a negation word dictionary, an interjection word dictionary, a rhetorical question word dictionary, and a relational conjunction word dictionary; S242. Import the reduced emotional theme into the emotional dictionary and identify the emotional words in the reduced emotional theme. The emotional words include basic emotional words, degree adverbs, negation words, interjections, rhetorical questions and relational conjunctions. S243. Assign weights to the identified sentiment words based on semantic rules, and calculate the sentiment value of the sentiment topic. S244. Determine the user's emotional state based on the calculated emotion value.

5. A customer service method based on emotion perception using Internet of Things information technology according to claim 4, characterized in that, The formula for calculating the sentiment value of the emotional topic is as follows: ； In the formula, V represents the sentiment value of the emotional topic; n represents the number of basic sentiment words; m indicates the number of interjections; r indicates the number of negative words; W i This represents the weight of the i-th basic sentiment word; N i This represents the weight of the i-th interjection; D indicates the weight of the degree adverb; X represents the negation gate; it takes a value of 1 if a negation word is present, and a value of 0 otherwise. Y represents the gate for rhetorical rhetorical questions. If a rhetorical rhetorical question appears, the value is 1; otherwise, the value is 0. H represents the influence coefficient; W represents the weight of the current basic sentiment word.

6. The customer service method based on emotion perception in the Internet of Things (IoT) information technology according to claim 1, characterized in that, The process of analyzing user intent based on real-time message data and generating response content from a response database based on the user's current emotional state includes the following steps: S31. Collect message text data labeled with user intent and extract features from the message text data; S32. An intent recognition model based on a RoBERTa pre-trained model is used to train the extracted features and then use the trained intent recognition model to identify the user's intent based on the extracted features. S33. Based on the user's intent and purpose, select a response template from the response database that matches the user's intent and purpose; S34. Adjust the tone and content of the reply template according to the user's current emotional state to obtain the final reply content.

7. A customer service method based on emotion perception using Internet of Things information technology according to claim 6, characterized in that, The step of selecting a response template from the response database that matches the user's intent and purpose includes the following steps: S331. Collect user intent questions and customer service responses through historical human customer service records; S332. Utilize natural language processing technology to perform text analysis and semantic understanding on intent questions and customer service responses, and extract key information and intent tags from the responses; S333. Based on the key information of the response, construct a response template for each intent tag, and establish a relationship between the response template and the intent tag to build a response database; S334. Match the user's intent with the intent tags in the response database based on similarity. S335. Based on the similarity matching results, select the response template associated with the intent tag with the highest similarity as the response template that matches the user's intent.

8. An emotion-aware Internet of Things (IoT) information technology customer service system, used to implement the emotion-aware IoT information technology customer service method according to any one of claims 1-7, characterized in that, The IoT information technology customer service system based on emotion perception includes: a data acquisition module, an emotion analysis module, a response content generation module, and a response content push module, and the data acquisition module, the emotion analysis module, the response content generation module, and the response content push module are connected in sequence. The data acquisition module is used to acquire real-time message data between users and customer service using IoT devices, and transmit it to the cloud server in real time through the IoT platform. The real-time message data is text information exchanged between users and customer service. The sentiment analysis module is used to perform text analysis on the acquired real-time message data and analyze the user's current emotional state through sentiment analysis methods. The reply content generation module is used to analyze the user's intent and purpose based on real-time message data, and generate reply content through the reply database based on the user's current emotional state. The response content push module is used to push the generated response content back to the user's IoT device using the IoT platform, and continuously monitor the interaction data between the user and customer service to dynamically adjust the response content.

Images