Power grid customer service intelligent interaction method and system based on big data analysis

Through big data analysis, we identify the consulting intentions and electricity consumption behaviors of power users, build an intelligent interactive system, solve the problem of low interaction efficiency of power grid customer service, realize rapid response to power users' needs and personalized services, and improve the quality of power grid service.

CN119003738APending Publication Date: 2024-11-22国家电网有限公司客户服务中心
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411238848.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-05
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

The existing power grid customer service interactive system relies on manual customer service, which leads to untimely handling of customer power problems and affects service efficiency.

Method used

The intelligent interaction method of power grid customer service based on big data analysis is adopted. By collecting multi-modal consulting data, identifying user consulting intentions, building a consumer behavior matrix, setting up dialogue management methods, realizing intelligent question-and-answer matching and user feedback, and improving interaction efficiency.

Benefits of technology

It realizes timely and effective processing of electricity needs for power users, improves the efficiency of grid customer service, provides personalized services, reduces user waiting time and dissatisfaction, and improves service quality and user loyalty.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119003738A_ABST
    Figure CN119003738A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of artificial intelligence, and discloses a power grid customer service intelligent interaction method and system based on big data analysis, and the method comprises the steps: collecting the multi-mode consultation data of a power user, so as to set a demand recognition mode of the power user; extracting a data detection feature corresponding to the demand identification mode to screen out a feature consumption behavior corresponding to the power consumer under the consultation intention; analyzing the time sequence characteristics of the power consumption behaviors of the power users and the power consumption demand characteristics of the power users so as to set a dialogue management mode of the power users; identifying the emotional state of the power user, and setting the language understanding degree of the power grid customer service to set a question and answer matching mechanism; evaluating the satisfaction degree of the power customer on the question and answer matching mechanism, and constructing a user feedback interface of the satisfaction degree; and in combination with a question-answer matching mechanism and a user feedback interface, power grid customer service intelligent interaction is realized. According to the invention, the electricity demand of the power consumer can be timely and effectively processed, and the working efficiency of the power grid customer service is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a power grid customer service intelligent interaction method and system based on big data analysis. Background Art

[0002] Intelligent interaction of power grid customer service refers to the use of artificial intelligence technology in the field of user services in the power industry to enable computer systems, equipment or applications to communicate and interact with users in a way that is closer to natural human communication. It can improve the automation and intelligence level of the power customer service system, thereby improving user experience and service efficiency. With the gradual evolution of the power grid system, the demand for automation and efficiency in the power industry has increased. In order to meet the user's electricity demand to the greatest extent, optimizing the interaction method between power grid customer service and users has become an important means to improve the quality of power services.

[0003] The existing power grid customer service interaction system mainly adopts the manual customer service method. This method helps customers handle business or solve problems through face-to-face communication between customer service personnel and customers to improve customer satisfaction. However, due to the limited number of manual customer service, this method results in customers' electricity usage problems not being solved in a timely and effective manner, thus affecting the efficiency of power service. Summary of the invention

[0004] In order to solve the above problems, the present invention provides a power grid customer service intelligent interaction method and system based on big data analysis, which can timely and effectively handle the electricity demand of power users and improve the efficiency of power grid customer service.

[0005] In a first aspect, the present invention provides a power grid customer service intelligent interaction method based on big data analysis, comprising:

[0006] Collecting multimodal consultation data of power users, normalizing the multimodal consultation data to obtain target consultation data, analyzing the data type of the target consultation data, and setting a demand identification method for the power users according to the data type;

[0007] Extracting data detection features corresponding to the demand identification method, and performing feature fusion processing on the data detection features to obtain multimodal fusion features, identifying the consultation intention corresponding to the power user according to the multimodal fusion features, analyzing the power consumption behavior corresponding to the power user under the consultation intention, constructing a consumption behavior matrix corresponding to the power user according to the consultation intention and the power behavior consumption data, and screening out characteristic consumption behaviors corresponding to the power user from the power consumption behaviors based on the consumption behavior matrix;

[0008] Set the classification intention of the consultation intention, analyze the electricity consumption behavior corresponding to the classification intention, extract the time series characteristics corresponding to the electricity consumption behavior, analyze the time series change trend of the characteristic consumption behavior according to the time series characteristics, analyze the electricity demand characteristics corresponding to the electricity user under the time series change trend, match the consultation intention list of the electricity user according to the electricity demand characteristics, and set the dialogue management method corresponding to the electricity user in the consultation intention list;

[0009] Identify the emotional state of the power user in the dialogue management mode, and set the language comprehension level of the power grid customer service and the power user in the dialogue management mode, and set a question-answer matching mechanism between the power grid customer service and the power user in the consultation process based on the language comprehension level and the emotional state;

[0010] Evaluate the output effect of the question-answer matching mechanism in the consultation process of the power user, analyze the power user's satisfaction with the power grid customer service based on the output effect, and construct a user feedback interface for the power user to the power grid customer service according to the satisfaction degree;

[0011] In combination with the question-answer matching mechanism and the user feedback interface, an intelligent interaction mechanism between the power user and the power grid customer service during the consultation process is constructed.

[0012] In a possible implementation of the first aspect, identifying the consultation intention corresponding to the power user according to the multimodal fusion feature includes:

[0013] Setting a keyword detection device corresponding to the multimodal fusion feature;

[0014] Calculating the consultation urgency of the electricity user;

[0015] Extracting customer dimension information corresponding to the consultation urgency, and extracting operation data of the electric equipment from the customer dimension information;

[0016] Analyzing the power consumption habits of the power user and the abnormal factors of the power-consuming equipment according to the operation data and the keyword detection device;

[0017] Based on the electricity usage habit and the abnormal factor, the consultation intention corresponding to the electricity user is identified.

[0018] In a possible implementation of the first aspect, the analyzing the power consumption behavior corresponding to the power user under the consultation intention includes:

[0019] Analyzing the electricity usage time sequence characteristics of the power user according to the consultation intention;

[0020] Identifying the power consumption type corresponding to the power usage time sequence feature;

[0021] Evaluate the normal power usage threshold corresponding to the power consumption type;

[0022] Analyze the power consumption of the power user under the normal power consumption threshold;

[0023] Based on the used power quantity, the power consumption behavior of the power user corresponding to the consultation intention is analyzed.

[0024] In a possible implementation of the first aspect, constructing a consumption behavior matrix corresponding to the power user according to the consultation intention and the power consumption behavior data includes:

[0025] Analyzing the user type of the power user according to the consultation intention and the power consumption behavior data;

[0026] Identify the power usage time period corresponding to the user type and the power usage within the power usage time period;

[0027] Analyzing the power load and power consumption differences of the power user in different time periods according to the power consumption;

[0028] Based on the user type, the electricity consumption time period, the electricity load and the electricity consumption difference, a consumption behavior matrix corresponding to the electricity user is constructed.

[0029] In a possible implementation of the first aspect, analyzing the time series change trend of the characteristic consumption behavior according to the time series feature includes:

[0030] Analyzing the seasonal variation pattern of the characteristic consumption behavior according to the time series characteristics;

[0031] identifying the electricity consumption corresponding to the seasonal variation pattern;

[0032] Analyze the long-term trend of the electricity consumption during the electricity service process;

[0033] Identify external factors that influence the long-term trend;

[0034] Based on the long-term trend and the external influencing factors, the time series change trend of the characteristic consumption behavior is analyzed.

[0035] In a possible implementation of the first aspect, matching the consultation intention list of the power user according to the power demand characteristics includes:

[0036] According to the electricity demand characteristics, setting the consultation type of the electricity user;

[0037] Identifying a user profile of the power user;

[0038] Analyze the demand influencing factors corresponding to the user portrait;

[0039] Based on the consultation type and the demand influencing factors, the consultation intention list of the power user is matched.

[0040] In a possible implementation of the first aspect, setting the dialogue management mode corresponding to the power user in the consultation intention list includes:

[0041] Identify the expression type of the power user during the consultation process, and set a semantic recognition mechanism for the expression type;

[0042] Setting a dialogue tracking mode for the power user during the consultation process;

[0043] Setting the dialogue logic corresponding to the consultation intention list, and setting the user feedback corresponding to the dialogue logic;

[0044] According to the user feedback, a conversation end button is set;

[0045] In combination with the semantic recognition mechanism, the dialogue tracking mode and the dialogue end button, a dialogue management mode corresponding to the power user in the consultation intention list is set.

[0046] In a possible implementation of the first aspect, setting a question-answer matching mechanism between the grid customer service and the power user during the consultation process based on the language comprehension level and the emotional state includes:

[0047] Extracting a list of corresponding consultation intentions between the power grid customer service and the power user during the consultation process;

[0048] Setting a direct output result corresponding to the consultation intention list, and calculating the semantic similarity between the direct output result and the consultation intention list;

[0049] According to the semantic similarity, setting an associated search tool for the direct output result;

[0050] Based on the language comprehension level and the emotional state, setting a multi-round dialogue window between the power grid customer service and the power user;

[0051] Based on the direct output results, the associated search tool and the multi-round dialogue window, a question-answer matching mechanism is set between the power grid customer service and the power user during the consultation process.

[0052] In a possible implementation of the first aspect, the evaluating the output effect of the question-answer matching mechanism in the power user consultation process includes:

[0053] Collecting consultation result data of the question-answer matching mechanism during the consultation process of the power user;

[0054] Extracting the consultation type corresponding to the consultation result data, and calculating the precision rate corresponding to the consultation type;

[0055] Analyzing the degree of success of the question-answer matching mechanism in retrieving the consultation intention corresponding to the power user;

[0056] The output effect of the question-answer matching mechanism in the power user consultation process is evaluated based on the precision rate and the successful retrieval degree.

[0057] In a second aspect, the present invention provides a power grid customer service intelligent interaction system based on big data analysis, characterized in that the system comprises:

[0058] A data collection module, used for collecting multimodal consultation data of power users, normalizing the multimodal consultation data to obtain target consultation data, analyzing the data type of the target consultation data, and setting the demand identification method of the power user according to the data type;

[0059] a feature extraction module, used to extract data detection features corresponding to the demand identification method, and perform feature fusion processing on the data detection features to obtain multimodal fusion features, identify the consultation intention corresponding to the power user according to the multimodal fusion features, analyze the power consumption behavior corresponding to the power user under the consultation intention, construct a consumption behavior matrix corresponding to the power user according to the consultation intention and the power behavior consumption data, and screen out characteristic consumption behaviors corresponding to the power user from the power consumption behaviors based on the consumption behavior matrix;

[0060] An intention classification module is used to set the classification intention of the consultation intention, analyze the electricity consumption behavior corresponding to the classification intention, extract the time series characteristics corresponding to the electricity consumption behavior, analyze the time series change trend of the characteristic consumption behavior according to the time series characteristics, analyze the electricity demand characteristics corresponding to the electricity user under the time series change trend, match the consultation intention list of the electricity user according to the electricity demand characteristics, and set the dialogue management mode corresponding to the electricity user in the consultation intention list;

[0061] A question-answer matching module, used to identify the emotional state of the power user in the dialogue management mode, and set the language understanding degree between the power grid customer service and the power user in the dialogue management mode, and set a question-answer matching mechanism between the power grid customer service and the power user in the consultation process based on the language understanding degree and the emotional state;

[0062] A user feedback module, used to evaluate the output effect of the question-answer matching mechanism in the consultation process of the power user, analyze the power user's satisfaction with the power grid customer service based on the output effect, and construct a user feedback interface for the power user to the power grid customer service according to the satisfaction degree;

[0063] An intelligent interaction module is used to combine the question-answer matching mechanism and the user feedback interface to build an intelligent interaction mechanism between the power user and the power grid customer service during the consultation process.

[0064] Compared with the prior art, the technical principle and beneficial effects of this solution are:

[0065] The embodiment of the present invention can more comprehensively understand the consultation intentions and needs of users by collecting multimodal consultation data of power users, thereby providing more accurate intelligent interactive services; further, the embodiment of the present invention can identify the behavioral characteristics of different users by extracting the data detection features corresponding to the demand identification method, which is convenient for quickly locating the consultation intentions of power customers, and increasing the complementarity between the various features, so as to understand the needs of users from multiple angles and provide users with accurate and personalized services; the embodiment of the present invention can accurately locate the consultation intentions of customers by setting the classification intentions of the consultation intentions, shorten the waiting time of customers, and reduce multiple inquiries or wrong answers caused by unclear intention positioning; further, the embodiment of the present invention can accurately capture the consultation intentions and specific needs of power users by setting a question-and-answer matching mechanism between the power grid customer service and the power user in the consultation process based on the language understanding level and the emotional state, and alleviate the dissatisfaction or anxiety that users may have due to power problems, thereby improving user satisfaction. degree; the embodiment of the present invention can provide accurate answers to users by evaluating the precision and completeness of the question-answer matching mechanism in the consultation process of the power user, shorten the user's waiting time, and thus improve service efficiency; further, the embodiment of the present invention can identify the deficiencies in the service by analyzing the power user's satisfaction with the power grid customer service based on the output effect, improve the service quality of the power system, increase user loyalty, and reduce user churn; the embodiment of the present invention combines the question-answer matching mechanism and the user feedback interface to construct an intelligent interaction mechanism between the power user and the power grid customer service in the consultation process, which can ensure that the power grid customer service system can quickly identify the needs of power users and provide accurate answers, combine user feedback, provide more personalized services and support, and can solve user consultation problems in a timely and accurate manner, reducing user waiting time. In addition, through the problems fed back by users, common fault modes in the power service process can be identified, the fault response speed can be improved, and the service efficiency can be improved. Therefore, the intelligent interaction method of power grid customer service based on big data analysis proposed in the embodiment of the present invention can timely and effectively handle the power demand of power users and improve the efficiency of power grid customer service. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0067] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0068] Figure 1 A schematic diagram of a flow chart of a method for intelligent interaction of power grid customer service based on big data analysis provided by an embodiment of the present invention;

[0069] Figure 2 A schematic diagram of a module of a power grid customer service intelligent interaction system based on big data analysis provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0070] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0071] The embodiment of the present invention provides a method for intelligent interaction of power grid customer service based on big data analysis, and the execution subject of the method for intelligent interaction of power grid customer service based on big data analysis includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present invention. In other words, the method for intelligent interaction of power grid customer service based on big data analysis can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content distribution networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms.

[0072] See also Figure 1 FIG. 1 is a flow chart of a method for intelligent interaction of power grid customer service based on big data analysis provided by an embodiment of the present invention. Figure 1 The intelligent interaction method for power grid customer service based on big data analysis described in the paper includes:

[0073] S1. Collect multimodal consultation data of power users, normalize the multimodal consultation data to obtain target consultation data, analyze the data type of the target consultation data, and set the demand identification method of the power user according to the data type.

[0074] By collecting multimodal consultation data of power users, the embodiments of the present invention can more comprehensively understand the users' consultation intentions and needs, thereby providing more accurate intelligent interactive services. The power users refer to people who use power services, and the multimodal consultation data refer to different types of data generated by users when communicating with the power grid customer service intelligent interactive system, such as text data, image data, and video data.

[0075] As an embodiment of the present invention, the method of collecting multimodal consultation data of power users includes: setting a consultation channel for the power user, and configuring a data recorder corresponding to the consultation channel; identifying a language understanding method corresponding to the data recorder, and setting a semantic recognition partition corresponding to the language understanding method; configuring a data automatic classification component corresponding to the semantic recognition partition; and collecting the multimodal consultation data of the power user based on the data automatic classification component.

[0076] Among them, the consultation channel refers to the medium or way for users to communicate with the power grid customer service system, the data recorder refers to the system component used to record and store the consultation data submitted by users through the consultation channel, the language understanding method refers to the method used to parse and understand the user input language, the semantic recognition partition refers to the module partition for specific type of language understanding set according to different language characteristics or consultation content, and the data automatic classification component refers to the component that automatically recognizes and classifies user consultation content.

[0077] Optionally, the consultation channel setting of the electricity user can be implemented using a mobile application, such as the official APP of the power company. The data logger configuration corresponding to the consultation channel can be obtained through the application log of the power company. The language understanding method identification corresponding to the data logger can be implemented using natural language processing tools, such as the Stanford Core NLP tool. The data automatic classification component configuration corresponding to the semantic recognition partition can be obtained through a deep learning framework, such as the TensorFlow learning framework.

[0078] Furthermore, the embodiment of the present invention obtains target consulting data by normalizing the multimodal consulting data, thereby removing or correcting noise, missing values ​​and outliers in the consulting data, ensuring the consistency of consulting data from different sources and formats, and improving data quality. The target consulting data refers to consulting data of multiple types of electricity users after data cleaning and format unification.

[0079] Optionally, the normalization processing of the multimodal consultation data can be implemented using PyTorch-Wi deDeep.

[0080] By analyzing the data type of the target consultation data, the embodiment of the present invention can understand the user's emotional state and tone information, help identify the specific circumstances of the problem described by the user, and thus more accurately understand the user's intention. The data type refers to the classification of various types of data processed in the power grid customer service intelligent interaction system, such as video data, sensor data, and user behavior data.

[0081] Optionally, the data type analysis of the target consulting data can be obtained through a data analysis tool, such as a text data analysis tool and a video data analysis tool.

[0082] Furthermore, the embodiment of the present invention can quickly identify the specific problems consulted by the user by setting the demand identification method of the power user according to the data type, take corresponding solutions, and improve the response speed of the service. The demand identification method refers to a method for identifying the specific needs and intentions of the user, such as voice recognition and text scanning, which is set according to the data type consulted by the power user.

[0083] As an embodiment of the present invention, the method for identifying the power user's needs is set according to the data type, including: setting a search guide for the power user's consulting service according to the data type; extracting consulting keywords corresponding to the search guide, and configuring an attachment upload path for the consulting keywords; setting an attachment data format corresponding to the attachment upload path; and setting the power user's needs identification method according to the attachment data format and the consulting keywords.

[0084] Among them, the consulting business refers to various consulting services provided by power companies or power grid companies to customers, the search guide refers to the guiding instructions provided to power users for business consultation, the consulting keywords refer to the keywords or phrases that users may use during the consultation process, the attachment upload path refers to the specific path or upload interface provided by the system when the user uploads an attachment, and the attachment data format refers to the format of supporting materials related to the consulting business uploaded by power users, such as the JPEG format of images.

[0085] Optionally, according to the data type, the search guide setting of the power user consulting service can be obtained through an intelligent search tool, such as Elasticsearch, and the attachment upload path configuration of the consulting keyword can be implemented using a file management system, such as SharePoint.

[0086] S2. Extract data detection features corresponding to the demand identification method, and perform feature fusion processing on the data detection features to obtain multimodal fusion features, identify the consultation intention corresponding to the power user based on the multimodal fusion features, analyze the power consumption behavior corresponding to the power user under the consultation intention, construct a consumption behavior matrix corresponding to the power user based on the consultation intention and the power behavior consumption data, and screen out characteristic consumption behaviors corresponding to the power user from the power consumption behaviors based on the consumption behavior matrix.

[0087] The embodiment of the present invention can identify the behavioral characteristics of different users by extracting the data detection features corresponding to the demand identification method, so as to quickly locate the consulting intentions of power customers. The data detection features refer to the key features extracted from different types of consulting data in order to identify the specific needs and intentions of users in the power user consulting business, such as language features and facial expression features.

[0088] As an embodiment of the present invention, the extraction of data detection features corresponding to the demand identification method includes: analyzing the consultation data type corresponding to the demand identification method; formatting the consultation data type to obtain a target data format; constructing a vocabulary corresponding to the target data format content, and converting the vocabulary in the vocabulary into a vocabulary numerical vector; based on the vocabulary numerical vector, extracting the data detection features corresponding to the demand identification method.

[0089] Among them, the consultation data type refers to the consultation data format used by power users when consulting, such as text and audio, the target data format refers to the data format after cleaning and unifying the formats of different consultation data, the vocabulary table refers to a set of words or short sentences related to the user's consultation content, and the vocabulary numerical vector refers to a vector converted from a vocabulary that can capture vocabulary semantic information.

[0090] Optionally, the vocabulary construction corresponding to the target data format content can be obtained through a bag-of-words model, and the conversion of words in the vocabulary into vocabulary numerical vectors can be achieved using a word embedding model.

[0091] Furthermore, the embodiments of the present invention obtain multimodal fusion features by performing feature fusion processing on the data detection features, thereby reducing misunderstandings and misjudgments caused by single modality data, and improving the performance of machine learning models by understanding user needs from multiple perspectives. The multimodal fusion features refer to comprehensive features obtained by fusing data features from different modalities, such as text features, image features, and voice features.

[0092] Optionally, the feature fusion processing of the data detection features can be obtained through a multimodal attention mechanism of a deep learning model.

[0093] The embodiment of the present invention can increase the complementarity between the features by identifying the consultation intention corresponding to the power user based on the multimodal fusion features, understand the user's needs from multiple angles, and provide the user with accurate and personalized services. The consultation intention refers to the purpose or demand expressed by the power user when communicating with the power grid customer service intelligent interaction system.

[0094] As an embodiment of the present invention, identifying the consultation intention corresponding to the power user based on the multimodal fusion feature includes: setting a keyword detection device corresponding to the multimodal fusion feature; calculating the consultation urgency of the power user; extracting customer dimension information corresponding to the consultation urgency, and extracting operation data of the power equipment from the customer dimension information; analyzing the power user's power usage habits and abnormal factors of the power equipment based on the operation data and the keyword detection device; and identifying the consultation intention corresponding to the power user based on the power usage habits and the abnormal factors.

[0095] Among them, the keyword detection device refers to a keyword recognition system for extracting keyword recognition services related to power user consultation services, the consultation urgency refers to the urgency of the needs expressed by power users during consultation, the customer dimension information refers to various data and information related to power users, such as customer basic information, electricity usage information, payment information and service interaction information, the power-consuming equipment refers to various electrical appliances or equipment that consume electrical energy to work, the operating data refers to various data records generated by the power-consuming equipment during use, the power usage habits refer to the power usage patterns and preferences formed by power users in the long-term use of electricity, and the abnormal factors refer to data points or patterns that are significantly different from normal power usage patterns, such as equipment failures.

[0096] In an optional embodiment of the present invention, the consultation urgency of the power user is calculated using the following formula:

[0097]

[0098] Among them, C represents the consultation urgency of the power user, m represents the total number of service channels corresponding to the power user, e represents the variable index in the service channel, S e represents the number of consultations of all electricity users in service channel e, E e represents the urgency score of the consultation content of all power users on service channel e, g represents the total number of power users, t represents the variable index of power users, S te represents the consultation volume of the t-th power user in the service channel, E te represents the urgency score of the consultation content of the t-th power user on the service channel e;

[0099] Furthermore, the embodiment of the present invention can analyze the power consumption behavior corresponding to the power user under the consultation intention, predict the user's electricity demand through the power user's corresponding consumption behavior, provide customers with personalized service recommendations, and improve service quality. The power consumption behavior refers to the consumption behavior of the power user in the process of using power services.

[0100] As an embodiment of the present invention, the analysis of the power consumption behavior corresponding to the power user under the consultation intention includes: analyzing the power usage timing characteristics of the power user according to the consultation intention; identifying the power consumption type corresponding to the power usage timing characteristics; evaluating the normal power usage threshold corresponding to the power consumption type; analyzing the power usage of the power user under the normal power usage threshold; and based on the power usage, analyzing the power consumption behavior corresponding to the power user under the consultation intention.

[0101] Among them, the electricity consumption time series characteristics refer to the regular characteristics of the user's electricity consumption behavior over time, the electricity consumption type refers to the type of equipment that uses electricity services to operate, the normal electricity usage threshold refers to the normal power consumption range of the electrical equipment within a specified time, and the power consumption refers to the total amount of electricity consumed by the power user within a certain period of time.

[0102] Optionally, the analysis of the electricity user's electricity consumption timing characteristics can be implemented using the timing prediction module of the DTEmpower data modeling platform, the identification of the electricity consumption type corresponding to the electricity consumption timing characteristics can be obtained through the electricity consumption type corresponding to the electricity consumption type, and the analysis of the electricity usage of the electricity user under the normal electricity usage threshold can be implemented using an electricity metering device, such as an electricity meter.

[0103] The embodiment of the present invention constructs a consumption behavior matrix corresponding to the power user according to the consultation intention and the power consumption behavior data, thereby identifying the user's power usage habits and patterns, dividing the users into different groups, and providing more personalized services for different users. The consumption behavior matrix refers to a matrix constructed based on the power user's power consumption behavior data, which is used to represent the user's power consumption at different time points and other related characteristics.

[0104] As an embodiment of the present invention, constructing a consumption behavior matrix corresponding to the power user according to the consultation intention and the power behavior consumption data includes: analyzing the user type of the power user according to the consultation intention and the power behavior consumption data; identifying the power consumption time period corresponding to the user type and the power consumption within the power consumption time period; analyzing the power load and power consumption difference of the power user in different time periods according to the power consumption; and constructing the consumption behavior matrix corresponding to the power user based on the user type, the power consumption time period, the power load and the power consumption difference.

[0105] Among them, the user type refers to different categories divided according to the user's electricity consumption characteristics, social attributes, geographical location and other factors; the electricity consumption time period refers to the user's electricity consumption stage at different times of the day or the year, such as peak time, normal time and off-peak time; the electricity consumption refers to the total amount of electricity consumed by the user in a specific electricity consumption time period; the electricity load refers to the electricity demand pressure borne by the power system or user under specific time or conditions; the electricity consumption difference refers to the difference in electricity consumption or electricity consumption patterns of electricity users in different time periods or on different dates.

[0106] Optionally, the power load analysis of the power users in different time periods can be obtained through power load curve analysis, and the power consumption difference analysis of the power users in different time periods can be achieved by dividing the time periods by utilizing the power supply and demand conditions in various regions and the power consumption habits of the users.

[0107] Furthermore, the embodiment of the present invention can respond to the user's consultation and service needs more quickly and improve service efficiency by screening out the characteristic consumption behavior corresponding to the power user from the electricity consumption behavior based on the consumption behavior matrix. The characteristic consumption behavior refers to the user's representative and typical behavior pattern analyzed based on the power user's electricity consumption behavior data.

[0108] S3. Set the classification intention of the consultation intention, analyze the electricity consumption behavior corresponding to the classification intention, extract the time series characteristics corresponding to the electricity consumption behavior, analyze the time series change trend of the characteristic consumption behavior according to the time series characteristics, analyze the electricity demand characteristics corresponding to the electricity user under the time series change trend, match the consultation intention list of the electricity user according to the electricity demand characteristics, and set the dialogue management method corresponding to the electricity user in the consultation intention list.

[0109] The embodiment of the present invention can accurately locate the customer's consultation intention, shorten the customer's waiting time, and reduce multiple inquiries or wrong answers due to unclear intention positioning by setting the classification intention of the consultation intention. The classification intention refers to the classification demands obtained after classifying and dividing the consultation content of the power user.

[0110] Optionally, the classification intention setting of the consultation intention can be implemented by utilizing the consultation content of the power user.

[0111] Furthermore, the embodiments of the present invention can identify common failure modes, predict possible user needs, and provide users with more targeted consulting services by analyzing the electricity usage behavior corresponding to the classification intention. The electricity usage behavior refers to the behavior pattern of electricity users in the process of using electricity services.

[0112] Optionally, the electricity consumption behavior analysis corresponding to the classification intention can be obtained through a consumption behavior matrix corresponding to the electricity user.

[0113] The embodiment of the present invention can timely discover the user's abnormal electricity usage behavior by extracting the time series characteristics corresponding to the electricity usage behavior, which is helpful for early prediction and prevention of faults. The time series characteristics refer to the characteristics extracted from the electricity user's electricity usage behavior data and change with time.

[0114] Optionally, the time series feature extraction corresponding to the power consumption behavior can be obtained through a time series analysis tool, such as Statsmode ls.

[0115] Furthermore, the embodiments of the present invention can predict the electricity demand of electricity users by analyzing the time series change trend of the characteristic consumption behavior according to the time series characteristics, and recommend electricity products and services that better meet the needs of users based on the users' historical electricity consumption data and current prediction results. The time series change trend refers to the trend of the electricity consumption characteristics of electricity users changing over time.

[0116] As an embodiment of the present invention, analyzing the time series change trend of the characteristic consumption behavior according to the time series characteristics includes: analyzing the seasonal change pattern of the characteristic consumption behavior according to the time series characteristics; identifying the electricity consumption corresponding to the seasonal change pattern; analyzing the long-term trend of the electricity consumption in the power service process; identifying the external influencing factors corresponding to the long-term trend; and analyzing the time series change trend of the characteristic consumption behavior based on the long-term trend and the external influencing factors.

[0117] Among them, the seasonal change pattern refers to the regular pattern of electricity consumption changing with the seasons, the electricity consumption refers to the amount of electric energy consumed by electricity users or power systems within a specific time period, the electricity service refers to a series of service activities provided by power companies or power supply companies to electricity users, the long-term trend refers to the changing trend of electricity consumption over a longer period of time, and the external influencing factors refer to the external conditions that affect the changes in electricity consumption.

[0118] Among them, the seasonal change pattern analysis of the characteristic consumption behavior can be obtained through seasonal decomposition, such as STL decomposition, the long-term trend analysis of electricity consumption in the power service process can be achieved using trend line fitting, and the identification of external influencing factors corresponding to the long-term trend can be obtained through statistical tests.

[0119] The embodiment of the present invention can accurately predict the electricity demand of the electricity user in different seasons and time periods by analyzing the electricity demand characteristics corresponding to the electricity user under the time series change trend, so as to reasonably allocate power generation, transmission and distribution resources to avoid resource waste and shortage. The electricity demand characteristics refer to the demand characteristics of the electricity user caused by various factors within a certain period of time.

[0120] Optionally, the analysis of the electricity demand characteristics corresponding to the electricity user under the time series change trend can be implemented using a time series analysis tool, such as a Prophet tool.

[0121] Furthermore, the embodiment of the present invention can quickly identify the user's consulting intention by matching the power user's consulting intention list according to the power demand characteristics, and provide the user with a convenient and efficient service experience. The consulting intention list refers to a series of consulting questions and corresponding intention classifications that the user may ask in the power user and customer service interaction system.

[0122] As an embodiment of the present invention, matching the consultation intention list of the power user according to the power demand characteristics includes: setting the consultation type of the power user according to the power demand characteristics; identifying the user portrait of the power user; analyzing the demand influencing factors corresponding to the user portrait; and matching the consultation intention list of the power user based on the consultation type and the demand influencing factors.

[0123] Among them, the consultation type refers to the consultation category divided according to the user's electricity demand characteristics, the user portrait refers to the virtual image constructed based on the user's electricity consumption data, social attributes, behavioral habits and other information, and the demand influencing factors refer to various factors that affect the user's electricity demand, such as seasonal changes.

[0124] Optionally, the user portrait identification of the power user can be obtained through the most representative power consumption characteristics of the power user.

[0125] The embodiment of the present invention can interact with users more naturally by setting the dialogue management method corresponding to the power user in the consultation intention list, not only answering the user's questions but also guiding the user to conduct more in-depth consultation or operation, and realizing automatic processing through a preset dialogue management method without human intervention. The dialogue management method refers to a method for managing the dialogue process between users and systems in an interactive system.

[0126] As an embodiment of the present invention, the setting of the dialogue management method corresponding to the power user in the consultation intention list includes: identifying the expression type of the power user in the consultation process, and setting a semantic recognition mechanism for the expression type; setting the dialogue tracking mode of the power user in the consultation process; setting the dialogue logic corresponding to the consultation intention list, and setting user feedback corresponding to the dialogue logic; setting a dialogue end button based on the user feedback; and setting the dialogue management method corresponding to the power user in the consultation intention list in combination with the semantic recognition mechanism, the dialogue tracking mode and the dialogue end button.

[0127] Among them, the expression type refers to the expression method or classification of consultation content adopted by the user in the consultation process, the semantic recognition mechanism refers to the system or algorithm mechanism that can parse and understand the user's expression content, the dialogue tracking mode refers to the method used by the system to track the progress of the dialogue, user questions and system responses during the user consultation process, the dialogue logic refers to the dialogue process designed based on specific rules or algorithms, and the algorithm used to guide the answering of questions and information exchange in the consultation process, the user feedback refers to the user's evaluation and suggestions on the power service or consultation process, and the dialogue end button refers to the choice provided to the user on the dialogue interface, which is a device used to indicate that the user's problem has been solved or the user wants to end the current dialogue.

[0128] Optionally, the semantic recognition mechanism setting of the expression type can be obtained through named entity recognition, the conversation tracking mode setting of the power user during the consultation process can be implemented using task-oriented conversation tracking, the conversation logic setting corresponding to the consultation intention list can be obtained through the consultation intention of the power user, and the conversation end button setting can be implemented using the power user's response status and problem-solving degree.

[0129] S4. Identify the emotional state of the power user under the dialogue management mode, and set the language understanding level of the power grid customer service and the power user under the dialogue management mode. Based on the language understanding level and the emotional state, set a question-answer matching mechanism between the power grid customer service and the power user during the consultation process.

[0130] The embodiment of the present invention can understand the urgency of the user's pending demands, speed up the progress of consulting services, and make a quick response by identifying the emotional state of the power user under the dialogue management method. The emotional state refers to the emotional or emotional state shown by the power user during the dialogue process with the intelligent interactive system.

[0131] Optionally, the emotional state recognition of the power user in the dialogue management mode may be obtained through a monitoring device, such as a camera built into a mobile device.

[0132] Furthermore, the embodiment of the present invention can accurately understand user needs and dynamically adjust dialogue strategies according to user language habits by setting the language understanding level of the power grid customer service and the power user under the dialogue management mode to ensure that the dialogue can proceed smoothly. The power grid customer service refers to the customer service department or system provided by the power company that is responsible for handling various service requests from power users, such as inquiries, complaints, and repairs. The language understanding level refers to the power grid customer service system's ability and accuracy to understand the natural language (such as text or voice) input by the user.

[0133] Optionally, the language understanding level setting of the power grid customer service and the power user in the dialogue management mode can be obtained through a natural language processing mechanism, such as TensorFlow.

[0134] The embodiment of the present invention sets a question-answer matching mechanism between the power grid customer service and the power user during the consultation process based on the language comprehension level and the emotional state, thereby accurately capturing the consultation intention and specific needs of the power user, alleviating the user's dissatisfaction or anxiety that may be caused by power problems, and improving the user's satisfaction. The question-answer matching mechanism refers to a system mechanism that finds the most relevant and accurate answers to user questions from a pre-established knowledge base or a real-time query database for feedback.

[0135] As an embodiment of the present invention, the question-answer matching mechanism between the power grid customer service and the power user in the consultation process is set based on the language comprehension level and the emotional state, including: extracting the corresponding consultation intention list between the power grid customer service and the power user in the consultation process; setting the direct output result corresponding to the consultation intention list, and calculating the semantic similarity between the direct output result and the consultation intention list; according to the semantic similarity, setting an associated retrieval tool for the direct output result; based on the language comprehension level and the emotional state, setting a multi-round dialogue window between the power grid customer service and the power user; based on the direct output result, the associated retrieval tool and the multi-round dialogue window, setting the question-answer matching mechanism between the power grid customer service and the power user in the consultation process.

[0136] Among them, the direct output result refers to the clear solution provided for each consulting intention in the consulting intention list, the semantic similarity refers to the degree of similarity between two words, phrases or sentences at the semantic level, the associated retrieval tool refers to the search tool used by the system to retrieve additional relevant information when the direct output result cannot meet the user's needs or more information is required, and the multi-round dialogue window refers to the interface in the customer service system that supports multiple back-and-forth exchanges with the user.

[0137] Optionally, the multi-round dialogue window setting between the power grid customer service and the power user can be implemented using a dialogue management platform built into the system.

[0138] In an optional embodiment of the present invention, the semantic similarity between the direct output result and the consultation intention list is calculated using the following formula:

[0139]

[0140] Among them, A(D,B) represents the semantic similarity between the direct output result and the consultation intention list, d j represents the component of the direct output result on the jth semantic index, b j represents the component of the consultation intention list at the jth semantic index, n represents the total number of semantic indicators of the direct output result and the corresponding vector of the consultation intention list, and j represents the semantic indicator index of the direct output result and the corresponding vector of the consultation intention list.

[0141] S5. Evaluate the output effect of the question-answer matching mechanism in the power user consultation process, analyze the power user's satisfaction with the power grid customer service based on the output effect, and construct a user feedback interface for the power user to the power grid customer service based on the satisfaction level.

[0142] The embodiment of the present invention can provide users with accurate answers and shorten user waiting time by evaluating the precision and completeness of the question-answer matching mechanism in the power user consultation process, thereby improving service efficiency. The output effect refers to the effect of the solution provided by the question-answer matching mechanism after identifying the user's consultation intention on solving user needs.

[0143] As an embodiment of the present invention, the evaluation of the output effect of the question-answer matching mechanism in the power user consultation process includes: collecting consultation result data of the question-answer matching mechanism in the power user consultation process; extracting the consultation type corresponding to the consultation result data, and calculating the precision rate corresponding to the consultation type; analyzing the degree of successful retrieval of the question-answer matching mechanism for the consultation intention corresponding to the power user; and evaluating the output effect of the question-answer matching mechanism in the power user consultation process based on the precision rate and the degree of successful retrieval.

[0144] Among them, the consultation result data refers to the answers or solution data provided for the consultation questions of power users, the precision rate refers to the degree to which the question-answer matching mechanism can accurately identify the user's consultation intention and provide correct answers or solutions, and the successful retrieval degree refers to the degree to which the question-answer matching mechanism can successfully retrieve and solve the user's consultation intention and provide a complete answer or solution.

[0145] In an optional embodiment of the present invention, the precision rate corresponding to the consultation type is calculated using the following formula:

[0146]

[0147] Among them, R represents the precision rate corresponding to the consultation type, q represents the category variable of the consultation type, Q represents the total number of categories of the consultation type, and TP q represents the number of instances correctly predicted as positive for the consultation category q, FN q represents the number of instances that are incorrectly predicted as negative for the consultation category q, TP q +FN q represents the number of instances that are actually positive in the advisory category q.

[0148] Optionally, the analysis of the degree of success of the retrieval of the inquiry intention corresponding to the power user by the question-answer matching mechanism can be obtained through a recall formula.

[0149] Furthermore, the embodiment of the present invention can identify deficiencies in services, improve the service quality of the power system, increase user loyalty, and reduce user churn by analyzing the power users' satisfaction with the power grid customer service based on the output effect. The satisfaction level refers to the power users' overall evaluation of the services provided by the power grid customer service.

[0150] Optionally, based on the output effect, the analysis of the degree of satisfaction of the power users with the power grid customer service can be obtained by constructing feedback from power users.

[0151] The embodiment of the present invention can collect user evaluations, opinions and suggestions by constructing a user feedback interface for the power user to the power grid customer service according to the satisfaction level, so as to understand the user's true feelings about the service, identify deficiencies in the service according to the user's feedback, and continuously improve the service quality. The user feedback interface refers to a system or tool between the power user and the power grid customer service for collecting user feedback on the service experience.

[0152] Optionally, the user feedback interface construction of the power user to the power grid customer service according to the satisfaction level can be obtained by filling out an online questionnaire.

[0153] S6. Combining the question-answer matching mechanism and the user feedback interface, construct an intelligent interaction mechanism between the power user and the power grid customer service during the consultation process.

[0154] The embodiment of the present invention combines the question-answer matching mechanism and the user feedback interface to construct an intelligent interaction mechanism between the power user and the power grid customer service during the consultation process, thereby ensuring that the power grid customer service system can quickly identify the needs of power users and provide accurate answers, and provide more personalized services and support in combination with user feedback, and can promptly and accurately solve user consultation problems and reduce user waiting time. In addition, through the problems reported by users, common failure modes in the power service process can be identified, the failure response speed can be improved, and the service efficiency can be improved. The intelligent interaction mechanism refers to an intelligent interaction method between power users and power grid customer service in the power service system that is realized by using artificial skill technology.

[0155] It can be seen that the embodiments of the present invention can more comprehensively understand the consultation intentions and needs of users by collecting multimodal consultation data of power users, thereby providing more accurate intelligent interactive services; further, the embodiments of the present invention can identify the behavioral characteristics of different users by extracting the data detection features corresponding to the demand identification method, so as to facilitate the rapid positioning of the consultation intentions of power customers, and increase the complementarity between the various features, understand the needs of users from multiple angles, and provide users with accurate and personalized services; the embodiments of the present invention can accurately locate the consultation intentions of customers by setting the classification intentions of the consultation intentions, shorten the waiting time of customers, and reduce multiple inquiries or wrong answers caused by unclear intention positioning; further, the embodiments of the present invention can accurately capture the consultation intentions and specific needs of power users by setting a question-and-answer matching mechanism between the power grid customer service and the power user in the consultation process based on the language understanding level and the emotional state, and alleviate the dissatisfaction or anxiety that users may have due to power problems, thereby improving users' Satisfaction; The embodiment of the present invention can provide accurate answers to users and shorten user waiting time by evaluating the precision and completeness of the question-answer matching mechanism in the consultation process of the power user, thereby improving service efficiency; Further, the embodiment of the present invention can identify deficiencies in the service, improve the service quality of the power system, increase user loyalty, and reduce user churn by analyzing the satisfaction of the power user with the power grid customer service based on the output effect; The embodiment of the present invention combines the question-answer matching mechanism and the user feedback interface to construct an intelligent interaction mechanism between the power user and the power grid customer service in the consultation process, which can ensure that the power grid customer service system can quickly identify the needs of power users and provide accurate answers, combine user feedback, provide more personalized services and support, and can solve user consultation problems in a timely and accurate manner, reducing user waiting time. In addition, through the problems reported by users, common fault modes in the power service process can be identified, the fault response speed can be improved, and the service efficiency can be improved. Therefore, the intelligent interaction method of power grid customer service based on big data analysis proposed in the embodiment of the present invention can timely and effectively handle the power demand of power users and improve the efficiency of power grid customer service.

[0156] like Figure 2 As shown, it is a system functional module diagram of the power grid customer service intelligent interaction based on big data analysis of the present invention.

[0157] The power grid customer service intelligent interactive system 200 based on big data analysis of the present invention can be installed in an electronic device. According to the functions implemented, the power grid customer service intelligent interactive system based on big data analysis can include a data acquisition module 201, a feature extraction module 202, an intention classification module 203, a question-answer matching module 204, a user feedback module 205, and an intelligent interaction module 206. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.

[0158] In the embodiment of the present invention, the functions of each module / unit are as follows:

[0159] The data collection module 201 is used to collect multimodal consultation data of power users, normalize the multimodal consultation data to obtain target consultation data, analyze the data type of the target consultation data, and set the demand identification method of the power user according to the data type;

[0160] The feature extraction module 202 is used to extract the data detection features corresponding to the demand identification method, and perform feature fusion processing on the data detection features to obtain multimodal fusion features, identify the consultation intention corresponding to the power user according to the multimodal fusion features, analyze the power consumption behavior corresponding to the power user under the consultation intention, construct the consumption behavior matrix corresponding to the power user according to the consultation intention and the power behavior consumption data, and screen out the characteristic consumption behavior corresponding to the power user from the power consumption behavior based on the consumption behavior matrix;

[0161] The intention classification module 203 is used to set the classification intention of the consultation intention, analyze the electricity consumption behavior corresponding to the classification intention, extract the time series characteristics corresponding to the electricity consumption behavior, analyze the time series change trend of the characteristic consumption behavior according to the time series characteristics, analyze the electricity demand characteristics corresponding to the electricity user under the time series change trend, match the consultation intention list of the electricity user according to the electricity demand characteristics, and set the dialogue management mode corresponding to the electricity user in the consultation intention list;

[0162] A question-answer matching module 204 is used to identify the emotional state of the power user in the dialogue management mode, and set the language understanding degree between the power grid customer service and the power user in the dialogue management mode, and set a question-answer matching mechanism between the power grid customer service and the power user in the consultation process based on the language understanding degree and the emotional state;

[0163] A user feedback module 205 is used to evaluate the output effect of the question-answer matching mechanism in the consultation process of the power user, analyze the power user's satisfaction with the power grid customer service based on the output effect, and construct a user feedback interface for the power user to the power grid customer service according to the satisfaction;

[0164] The intelligent interaction module 206 is used to combine the question-answer matching mechanism and the user feedback interface to build an intelligent interaction mechanism between the power user and the power grid customer service during the consultation process.

[0165] In detail, each module in the power grid customer service intelligent interaction system 200 based on big data analysis in the embodiment of the present invention is used in the same manner as above. Figure 1 The method of intelligent interaction of power grid customer service based on big data analysis described in the text is the same technical means and can produce the same technical effects, so I will not go into details here.

[0166] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A power grid customer service intelligent interaction method based on big data analysis, characterized in that: The method comprises: Collecting multimodal consultation data of power users, normalizing the multimodal consultation data to obtain target consultation data, analyzing the data type of the target consultation data, and setting a demand identification method for the power users according to the data type; Extracting data detection features corresponding to the demand identification method, and performing feature fusion processing on the data detection features to obtain multimodal fusion features, identifying the consultation intention corresponding to the power user according to the multimodal fusion features, analyzing the power consumption behavior corresponding to the power user under the consultation intention, constructing a consumption behavior matrix corresponding to the power user according to the consultation intention and the power behavior consumption data, and screening out characteristic consumption behaviors corresponding to the power user from the power consumption behaviors based on the consumption behavior matrix; Set the classification intention of the consultation intention, analyze the electricity consumption behavior corresponding to the classification intention, extract the time series characteristics corresponding to the electricity consumption behavior, analyze the time series change trend of the characteristic consumption behavior according to the time series characteristics, analyze the electricity demand characteristics corresponding to the electricity user under the time series change trend, match the consultation intention list of the electricity user according to the electricity demand characteristics, and set the dialogue management method corresponding to the electricity user in the consultation intention list; Identify the emotional state of the power user in the dialogue management mode, and set the language understanding level of the power grid customer service and the power user in the dialogue management mode, and set a question-answer matching mechanism between the power grid customer service and the power user in the consultation process based on the language understanding level and the emotional state; Evaluate the output effect of the question-answer matching mechanism in the consultation process of the power user, analyze the power user's satisfaction with the power grid customer service based on the output effect, and construct a user feedback interface for the power user to the power grid customer service according to the satisfaction degree; In combination with the question-answer matching mechanism and the user feedback interface, an intelligent interaction mechanism between the power user and the power grid customer service during the consultation process is constructed.

2. The method according to claim 1, characterized in that The identifying the consultation intention corresponding to the power user according to the multimodal fusion feature includes: Setting a keyword detection device corresponding to the multimodal fusion feature; Calculating the consultation urgency of the electricity user; Extracting customer dimension information corresponding to the consultation urgency, and extracting operation data of the electric equipment from the customer dimension information; Analyzing the power consumption habits of the power user and the abnormal factors of the power-consuming equipment according to the operation data and the keyword detection device; Based on the electricity usage habit and the abnormal factor, the consultation intention corresponding to the electricity user is identified.

3. The method according to claim 1, characterized in that: The analyzing the power consumption behavior of the power user corresponding to the consultation intention includes: Analyzing the electricity usage time sequence characteristics of the power user according to the consultation intention; Identifying the power consumption type corresponding to the power usage time sequence feature; Evaluate the normal power usage threshold corresponding to the power consumption type; Analyze the power consumption of the power user under the normal power consumption threshold; Based on the used power quantity, the power consumption behavior of the power user corresponding to the consultation intention is analyzed.

4. The method according to claim 1, characterized in that: The step of constructing a consumption behavior matrix corresponding to the power user according to the consultation intention and the power consumption behavior data includes: Analyzing the user type of the power user according to the consultation intention and the power consumption behavior data; Identify the power usage time period corresponding to the user type and the power usage within the power usage time period; Analyzing the power load and power consumption differences of the power user in different time periods according to the power consumption; Based on the user type, the electricity consumption time period, the electricity load and the electricity consumption difference, a consumption behavior matrix corresponding to the electricity user is constructed.

5. The method according to claim 1, characterized in that Analyzing the time series change trend of the characteristic consumption behavior according to the time series characteristics includes: Analyzing the seasonal variation pattern of the characteristic consumption behavior according to the time series characteristics; identifying the electricity consumption corresponding to the seasonal variation pattern; Analyze the long-term trend of the electricity consumption during the electricity service process; Identify external factors that influence the long-term trend; Based on the long-term trend and the external influencing factors, the time series change trend of the characteristic consumption behavior is analyzed.

6. The method according to claim 1, characterized in that The matching the consultation intention list of the power user according to the power demand characteristics includes: According to the electricity demand characteristics, setting the consultation type of the electricity user; Identifying a user profile of the power user; Analyze the demand influencing factors corresponding to the user portrait; Based on the consultation type and the demand influencing factors, the consultation intention list of the power user is matched.

7. The method according to claim 1, characterized in that The setting of the dialogue management mode corresponding to the power user in the consultation intention list includes: Identify the expression type of the power user during the consultation process, and set a semantic recognition mechanism for the expression type; Setting a dialogue tracking mode for the power user during the consultation process; Setting the dialogue logic corresponding to the consultation intention list, and setting the user feedback corresponding to the dialogue logic; According to the user feedback, a conversation end button is set; In combination with the semantic recognition mechanism, the dialogue tracking mode and the dialogue end button, a dialogue management mode corresponding to the power user in the consultation intention list is set.

8. The method according to claim 1, characterized in that The setting of a question-answer matching mechanism between the power grid customer service and the power user during the consultation process based on the language comprehension level and the emotional state includes: Extracting a list of corresponding consultation intentions between the power grid customer service and the power user during the consultation process; Setting a direct output result corresponding to the consultation intention list, and calculating the semantic similarity between the direct output result and the consultation intention list; According to the semantic similarity, setting an associated search tool for the direct output result; Based on the language comprehension level and the emotional state, setting a multi-round dialogue window between the power grid customer service and the power user; Based on the direct output results, the associated search tool and the multi-round dialogue window, a question-answer matching mechanism is set between the power grid customer service and the power user during the consultation process.

9. The method according to claim 1, characterized in that: The evaluating of the output effect of the question-answer matching mechanism in the power user consultation process includes: Collecting consultation result data of the question-answer matching mechanism during the consultation process of the power user; Extracting the consultation type corresponding to the consultation result data, and calculating the precision rate corresponding to the consultation type; Analyzing the degree of success of the question-answer matching mechanism in retrieving the consultation intention corresponding to the power user; The output effect of the question-answer matching mechanism in the power user consultation process is evaluated based on the precision rate and the successful retrieval degree.

10. The power grid customer service intelligent interactive system based on big data analysis is characterized by: The system comprises: A data collection module, used for collecting multimodal consultation data of power users, normalizing the multimodal consultation data to obtain target consultation data, analyzing the data type of the target consultation data, and setting the demand identification method of the power user according to the data type; a feature extraction module, used to extract data detection features corresponding to the demand identification method, and perform feature fusion processing on the data detection features to obtain multimodal fusion features, identify the consultation intention corresponding to the power user according to the multimodal fusion features, analyze the power consumption behavior corresponding to the power user under the consultation intention, construct a consumption behavior matrix corresponding to the power user according to the consultation intention and the power behavior consumption data, and screen out characteristic consumption behaviors corresponding to the power user from the power consumption behaviors based on the consumption behavior matrix; An intention classification module is used to set the classification intention of the consultation intention, analyze the electricity consumption behavior corresponding to the classification intention, extract the time series characteristics corresponding to the electricity consumption behavior, analyze the time series change trend of the characteristic consumption behavior according to the time series characteristics, analyze the electricity demand characteristics corresponding to the electricity user under the time series change trend, match the consultation intention list of the electricity user according to the electricity demand characteristics, and set the dialogue management mode corresponding to the electricity user in the consultation intention list; A question-answer matching module, used to identify the emotional state of the power user in the dialogue management mode, and set the language understanding degree between the power grid customer service and the power user in the dialogue management mode, and set a question-answer matching mechanism between the power grid customer service and the power user in the consultation process based on the language understanding degree and the emotional state; A user feedback module, used to evaluate the output effect of the question-answer matching mechanism in the consultation process of the power user, analyze the power user's satisfaction with the power grid customer service based on the output effect, and construct a user feedback interface for the power user to the power grid customer service according to the satisfaction degree; An intelligent interaction module is used to combine the question-answer matching mechanism and the user feedback interface to build an intelligent interaction mechanism between the power user and the power grid customer service during the consultation process.

Citation Information

Cited By

  • Work order-based sensitive customer management method and device

    CN119226490A