Cigarette marketing customer service system facilitating rapid communication of users

Abstract

The invention discloses a cigarette marketing customer service system facilitating rapid communication of users, and relates to the technical field of cigarette marketing, the system comprises a customer service request acquisition unit, a customer service request analysis unit, an associated business query unit, a customer service request response analysis unit and a customer service request response unit, the client is used for establishing a data channel between a cigarette marketing customer service side and a user side; the cigarette marketing customer service end receives a customer service request about cigarette marketing sent by the user end through the data channel; the customer service request response analysis unit is used for analyzing a response mode corresponding to the customer service request according to the emotion information, the demand information and the business module information of the user to obtain a customer service request response result; and the customer service request response unit is used for the cigarette marketing customer service terminal to send the customer service request response result to the user through the data channel. The communication efficiency between the cigarette enterprise and the user is improved.

Description

Technical Field

[0001] The present application relates to the technical field of cigarette marketing, and in particular to a cigarette marketing customer service system that facilitates quick communication among users. Background Art

[0002] In recent years, with the rapid development of artificial intelligence technology, domestic research on the application of virtual assistants in the field of customer service has become increasingly in-depth. Many companies have begun to try to develop their own virtual assistants to improve the efficiency and quality of customer service. In the field of customer service, the application of virtual assistants will become more extensive and in-depth. With the continuous innovation of technology, virtual assistants have surpassed traditional telephone and email communication methods and gradually expanded to channels such as instant messaging.

[0003] In the field of cigarette marketing, tobacco companies usually visit users regularly to communicate with them in order to solve their doubts in terms of operation and ordering. However, this method will lead to technical problems such as low communication efficiency between tobacco companies and users.

[0004] Based on this, there is an urgent need for a cigarette marketing customer service system that facilitates quick communication with users, so that tobacco companies can communicate quickly with users. Summary of the invention

[0005] The embodiment of the present application provides a cigarette marketing customer service system that facilitates users to communicate quickly to solve the above-mentioned technical problems.

[0006] The present application provides a cigarette marketing customer service system that facilitates quick communication between users, including a customer service request acquisition unit, a customer service request parsing unit, a related business query unit, a customer service request response analysis unit, and a customer service request response unit:

[0007] The customer service request acquisition unit is used to establish a data channel between the cigarette marketing customer service end and the user end; the cigarette marketing customer service end receives the customer service request for cigarette marketing sent by the user end through the data channel;

[0008] The customer service request parsing unit is used to determine the user's emotion information and demand information according to the instruction information carried in the received customer service request; wherein the instruction information includes voice instructions and / or gesture operation instructions;

[0009] The associated business query unit is used to query and obtain corresponding business module information according to the demand information corresponding to the customer service request based on the global navigation of the cigarette marketing customer service terminal;

[0010] The customer service request response analysis unit is used to analyze the response method corresponding to the customer service request according to the user's emotion information, the demand information and the business module information to obtain a customer service request response result;

[0011] The customer service request response unit is used by the cigarette marketing customer service end to send the customer service request response result to the user through the data channel.

[0012] Further, determining the user's emotion information according to the instruction information carried in the received customer service request includes:

[0013] Applying a multi-head attention mechanism to the instruction information to obtain a contextual semantic representation set of the instruction information;

[0014] Using a graph convolutional network to extract features from the contextual semantic representation set to obtain a sentiment graph;

[0015] Generate a set of emotion feature parameters based on the dependency information of each node and its adjacent nodes in the emotion graph;

[0016] Acquire a preset emotion determination library; wherein the emotion determination library includes reference feature parameters, and the reference feature parameters correspond to preset emotion information;

[0017] Calculating the similarity between the emotion feature parameter set and the reference feature parameters included in the emotion determination library;

[0018] The emotion information corresponding to the reference feature parameters in the emotion determination library whose similarity with the emotion feature parameter set is greater than a similarity threshold is determined as the emotion information of the user.

[0019] Furthermore, each node in the emotion graph represents a semantic emotion entity, and each edge in the emotion graph represents a dependency relationship between semantic emotion entities.

[0020] Furthermore, the similarity between the emotion feature parameter set and the reference feature parameters included in the emotion determination library is calculated.

[0021] Further, the response method corresponding to the customer service request is analyzed according to the emotion information, the demand information and the business module information of the user to obtain the customer service request response result, including:

[0022] According to the business module information, filtering out historical data rows related to the business module information from a cigarette marketing database;

[0023] Performing adjacent mean processing on the data in the historical data row to form adjacent data rows;

[0024] Performing importance analysis on each row of data in the adjacent data rows by using a wrapping feature extraction algorithm to determine importance data corresponding to each row of data;

[0025] Performing data drift detection on the importance data corresponding to each row of data to generate a data drift detection result;

[0026] Based on the importance data corresponding to each row of data and the data drift detection result, feature extraction processing is performed on the adjacent data rows to obtain key reference features;

[0027] After performing residual processing of different dimensions on the key reference features, convolution processing is performed together with the demand information and the emotion information to obtain initial reference features;

[0028] The initial reference feature is processed by dimensional feature mapping to obtain an initial node feature vector;

[0029] The better solutions in the initial node feature vector are screened based on a preset screening coefficient to analyze the response method corresponding to the customer service request to obtain the customer service request response result.

[0030] Furthermore, after the key reference features are subjected to residual processing of different dimensions, convolution processing is performed together with the demand information and the emotion information to obtain initial reference features, including:

[0031] The key reference features are split based on three spatial dimensions to obtain horizontal attention weight features, vertical attention weight features and deep attention weight features;

[0032] The horizontal attention weight feature, the vertical attention weight feature and the depth attention weight feature are respectively fused with the demand feature to obtain a first fused reference feature; wherein the demand feature is extracted from the demand information;

[0033] The horizontal attention weight feature, the vertical attention weight feature and the deep attention weight feature are respectively fused with the emotion feature parameter set to obtain a second fused reference feature;

[0034] The first fused reference feature and the second fused reference feature are processed using a convolutional layer of a deep network structure to obtain the initial reference feature.

[0035] Furthermore, the cigarette marketing customer service system that facilitates quick communication with users also includes:

[0036] A cigarette marketing customer service end guidance providing unit is used for the cigarette marketing customer service end to obtain the terminal construction guidance request sent by the user through the data channel, and determine the store business district positioning information and business conditions based on the user's location information carried in the terminal construction guidance request; use the big data technology to determine the surrounding store facility data, surrounding store business data, surrounding store store display data and business district population flow data according to the store business district positioning information; analyze the user's store characteristics based on the surrounding store facility data, the surrounding store business data, the surrounding store store display data and the business district population flow data, and provide information support and guidance for the user's store shelf display.

[0037] Furthermore, the cigarette marketing database includes multiple types of the following data: cigarette brand, cigarette packaging form, cigarette packaging color, cigarette type, filter type, tar content, nicotine content, filter length, cigarette circumference, cigarette name, sales volume, inventory volume, purchase price and selling price.

[0038] Furthermore, the cigarette marketing customer service system that facilitates quick communication among users also includes: a social platform communication unit, which is used to obtain real-time dynamic information posted by the user on the social media platform, and to communicate with the user in real time based on the real-time dynamic information.

[0039] Based on the embodiment provided by the present application, the cigarette marketing customer service system that facilitates users to communicate quickly includes a customer service request acquisition unit, a customer service request parsing unit, an associated business query unit, a customer service request response analysis unit, and a customer service request response unit: the customer service request acquisition unit is used to establish a data channel between the cigarette marketing customer service end and the user end; the cigarette marketing customer service end receives the customer service request about cigarette marketing sent by the user end through the data channel; the customer service request parsing unit is used to determine the user's emotional information and demand information according to the instruction information carried in the received customer service request; wherein the instruction information includes voice instructions and / or gesture operation instructions; the associated business query unit is used to query the corresponding business module information based on the global navigation of the cigarette marketing customer service end according to the demand information corresponding to the customer service request; the customer service request response analysis unit is used to analyze the response method corresponding to the customer service request according to the user's emotional information, demand information and business module information, and obtain the customer service request response result; the customer service request response unit is used for the cigarette marketing customer service end to send the customer service request response result to the user through the data channel. Thus, tobacco enterprises can communicate with users automatically and quickly, and improve the communication efficiency between tobacco enterprises and users.

[0040] Specifically, it has the following beneficial effects: the customer service request acquisition unit establishes a stable data channel to ensure unimpeded information transmission between the customer service end and the user end, laying the foundation for subsequent customer service request processing; the customer service request parsing unit accurately captures the user's emotions and needs by parsing the user's instruction information, so that the customer service system can provide more personalized and emotional services; the associated business query unit uses the global navigation system to quickly locate the business modules corresponding to the user's needs, improves the query efficiency and accuracy, and ensures that the user can obtain the required business information in a timely manner; the customer service request response analysis unit: comprehensively analyzes the user's emotional information, demand information and business module information, and ensures that the generated response results both meet user needs and comply with business specifications; the customer service request response unit sends the analyzed response results to the user through the data channel in a timely manner, ensuring the timeliness of the customer service response and improving user satisfaction and the service efficiency of the customer service system. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The drawings described herein are used to provide a further understanding of the embodiments of the present invention and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0042] Figure 1 A structural diagram of an optional cigarette marketing customer service system that facilitates quick communication with users according to an embodiment of the present application;

[0043] Figure 2 The present invention is an optional flowchart for determining a user's emotion information according to an embodiment of the present application.

[0044] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0045] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0046] Alternatively, if Figure 1 As shown, the present application provides a cigarette marketing customer service system that facilitates users to communicate quickly, including a customer service request acquisition unit 101, a customer service request parsing unit 102, a related business query unit 103, a customer service request response analysis unit 104 and a customer service request response unit 105:

[0047] The customer service request acquisition unit 101 is used to establish a data channel between the cigarette marketing customer service end and the user end; the cigarette marketing customer service end receives the customer service request for cigarette marketing sent by the user end through the data channel;

[0048] The customer service request parsing unit 102 is used to determine the user's emotion information and demand information according to the instruction information carried in the received customer service request; wherein the instruction information includes voice instructions and / or gesture operation instructions;

[0049] The associated business query unit 103 is used to query and obtain corresponding business module information according to the demand information corresponding to the customer service request based on the global navigation of the cigarette marketing customer service terminal;

[0050] The customer service request response analysis unit 104 is used to analyze the response method corresponding to the customer service request according to the user's emotion information, demand information and business module information to obtain the customer service request response result;

[0051] The customer service request response unit 105 is used for the cigarette marketing customer service end to send the customer service request response result to the user through the data channel.

[0052] Based on the embodiment provided by the present application, the cigarette marketing customer service system that facilitates users to communicate quickly includes a customer service request acquisition unit, a customer service request parsing unit, an associated business query unit, a customer service request response analysis unit, and a customer service request response unit: the customer service request acquisition unit is used to establish a data channel between the cigarette marketing customer service end and the user end; the cigarette marketing customer service end receives the customer service request about cigarette marketing sent by the user end through the data channel; the customer service request parsing unit is used to determine the user's emotional information and demand information according to the instruction information carried in the received customer service request; wherein the instruction information includes voice instructions and / or gesture operation instructions; the associated business query unit is used to query the corresponding business module information based on the global navigation of the cigarette marketing customer service end according to the demand information corresponding to the customer service request; the customer service request response analysis unit is used to analyze the response method corresponding to the customer service request according to the user's emotional information, demand information and business module information, and obtain the customer service request response result; the customer service request response unit is used for the cigarette marketing customer service end to send the customer service request response result to the user through the data channel. Thus, tobacco enterprises can communicate with users automatically and quickly, and improve the communication efficiency between tobacco enterprises and users.

[0053] Specifically, it has the following beneficial effects: the customer service request acquisition unit establishes a stable data channel to ensure unimpeded information transmission between the customer service end and the user end, laying the foundation for subsequent customer service request processing; the customer service request parsing unit accurately captures the user's emotions and needs by parsing the user's instruction information, so that the customer service system can provide more personalized and emotional services; the associated business query unit uses the global navigation system to quickly locate the business modules corresponding to the user's needs, improves the query efficiency and accuracy, and ensures that the user can obtain the required business information in a timely manner; the customer service request response analysis unit: comprehensively analyzes the user's emotional information, demand information and business module information, and ensures that the generated response results both meet user needs and comply with business specifications; the customer service request response unit sends the analyzed response results to the user through the data channel in a timely manner, ensuring the timeliness of the customer service response and improving user satisfaction and the service efficiency of the customer service system.

[0054] Alternatively, if Figure 2 As shown, according to the instruction information carried in the received customer service request, the user's emotional information is determined, including:

[0055] S201, applying a multi-head attention mechanism to the instruction information to obtain a contextual semantic representation set of the instruction information;

[0056] S202, using a graph convolutional network to extract features from the context semantic representation set to obtain a sentiment graph;

[0057] S203, generating a set of emotion feature parameters based on the dependency information of each node and its adjacent nodes in the emotion graph;

[0058] S204, obtaining a preset emotion determination library; wherein the emotion determination library includes reference feature parameters, and the reference feature parameters correspond to preset emotion information;

[0059] S205, calculating the similarity between the emotion feature parameter set and the reference feature parameters included in the emotion determination library;

[0060] S206: Determine the emotion information corresponding to the reference feature parameters in the emotion determination library whose similarity with the emotion feature parameter set is greater than a similarity threshold as the emotion information of the user.

[0061] Furthermore, each node in the sentiment graph represents a semantic sentiment entity, and each edge in the sentiment graph represents the dependency relationship between semantic sentiment entities.

[0062] Further, the similarity between the emotion feature parameter set and the reference feature parameters included in the emotion determination library is calculated based on the following formula;

[0063]

[0064] Wherein, similar is the similarity between the emotion feature parameter set and the reference feature parameters included in the emotion determination library; x i is the parameter value of the i-th feature parameter in the emotional feature parameter set; y i is the parameter value of the jth reference feature parameter included in the emotion determination library; the number of feature parameters in the emotion feature parameter set and the number of feature parameters in the emotion determination library are both N.

[0065] Furthermore, according to the user's emotional information, demand information and business module information, the response method corresponding to the customer service request is analyzed to obtain the customer service request response result, including:

[0066] According to the business module information, historical data rows related to the business module information are filtered out from the cigarette marketing database;

[0067] Perform adjacent mean processing on the data in the historical data row to form adjacent data rows;

[0068] The importance of each row of data in the adjacent data row is analyzed by using a wrapping feature extraction algorithm to determine the importance data corresponding to each row of data;

[0069] Perform data drift detection on the importance data corresponding to each row of data to generate data drift detection results;

[0070] Based on the importance data corresponding to each row of data and the data drift detection results, feature extraction is performed on the adjacent data rows to obtain key reference features;

[0071] After performing residual processing of different dimensions on the key reference features, they are convolved with the demand information and sentiment information to obtain the initial reference features.

[0072] After processing the initial reference features through dimensional feature mapping, the initial node feature vector is obtained;

[0073] The better solutions in the initial node feature vector are screened based on the preset screening coefficient to analyze the response method corresponding to the customer service request and obtain the customer service request response result.

[0074] Furthermore, the preset screening coefficient is obtained based on the following formula:

[0075]

[0076] Where t represents the current iteration number; j represents the number of the optimal solution under the current iteration number; Indicates the preset screening coefficient set for the better solution numbered j; Indicates the optimization screening factor set for the better solution numbered j; Vec represents the iterative influence factor set for the optimal solution numbered j; initial Represents the initial node feature vector.

[0077] Furthermore, the key reference features are processed with residuals of different dimensions and then convolved with the demand information and sentiment information to obtain the initial reference features, including:

[0078] The key reference features are split based on three spatial dimensions to obtain horizontal attention weight features, vertical attention weight features, and deep attention weight features;

[0079] The horizontal attention weight feature, the vertical attention weight feature and the deep attention weight feature are respectively fused with the demand feature to obtain a first fused reference feature; wherein the demand feature is extracted from the demand information;

[0080] The horizontal attention weight feature, the vertical attention weight feature and the deep attention weight feature are respectively fused with the sentiment feature parameter set to obtain a second fusion reference feature;

[0081] The first fused reference feature and the second fused reference feature are processed using a convolutional layer of a deep network structure to obtain an initial reference feature;

[0082] F initial =σ[W f ×{Upsample(F1)+Downsample(F2)}+b f ]

[0083] Among them, F initial is the initial reference feature; σ is the activation function; W f is the weight of the convolutional layer; b f is the bias of the convolutional layer; F1 is the first fusion reference feature; F2 is the second fusion reference feature; Upsample means upsampling; Downsample means downsampling.

[0084] Furthermore, the cigarette marketing customer service system that facilitates quick communication with users also includes:

[0085] The cigarette marketing customer service end guidance providing unit is used for the cigarette marketing customer service end to obtain the terminal construction guidance request sent by the user through the data channel, and determine the store business district positioning information and business conditions based on the user's location information carried in the terminal construction guidance request; use big data technology to determine the surrounding store facility data, surrounding store operation data, surrounding store store display data and business district population flow data based on the store business district positioning information; analyze the user's store characteristics based on the surrounding store facility data, surrounding store operation data, surrounding store store display data and business district population flow data, and provide information support and guidance for the user's store shelf display.

[0086] Furthermore, the cigarette marketing database includes multiple of the following data: cigarette brand, cigarette packaging form, cigarette packaging color, cigarette type, filter type, tar content, nicotine content, filter length, cigarette circumference, cigarette name, sales volume, inventory volume, purchase price and selling price.

[0087] Furthermore, the cigarette marketing customer service system that facilitates quick communication with users also includes: a social platform communication unit, which is used to obtain real-time dynamic information posted by users on social media platforms, and to communicate with users in real time based on the real-time dynamic information.

[0088] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A cigarette marketing customer service system that facilitates quick communication between users, characterized in that: It includes a customer service request acquisition unit, a customer service request parsing unit, a related business query unit, a customer service request response analysis unit and a customer service request response unit: The customer service request acquisition unit is used to establish a data channel between the cigarette marketing customer service end and the user end; the cigarette marketing customer service end receives the customer service request for cigarette marketing sent by the user end through the data channel; The customer service request parsing unit is used to determine the user's emotion information and demand information according to the instruction information carried in the received customer service request; wherein the instruction information includes voice instructions and / or gesture operation instructions; The associated business query unit is used to query and obtain corresponding business module information according to the demand information corresponding to the customer service request based on the global navigation of the cigarette marketing customer service terminal; The customer service request response analysis unit is used to analyze the response method corresponding to the customer service request according to the user's emotion information, the demand information and the business module information to obtain a customer service request response result; The customer service request response unit is used by the cigarette marketing customer service end to send the customer service request response result to the user through the data channel.

2. The cigarette marketing customer service system that facilitates quick communication with users according to claim 1, characterized in that: Determining the user's emotion information according to the instruction information carried in the received customer service request includes: Applying a multi-head attention mechanism to the instruction information to obtain a contextual semantic representation set of the instruction information; Using a graph convolutional network to extract features from the contextual semantic representation set to obtain a sentiment graph; Generate a set of emotion feature parameters based on the dependency information of each node and its adjacent nodes in the emotion graph; Acquire a preset emotion determination library; wherein the emotion determination library includes reference feature parameters, and the reference feature parameters correspond to preset emotion information; Calculating the similarity between the emotion feature parameter set and the reference feature parameters included in the emotion determination library; The emotion information corresponding to the reference feature parameters in the emotion determination library whose similarity with the emotion feature parameter set is greater than a similarity threshold is determined as the emotion information of the user.

3. The cigarette marketing customer service system that facilitates quick communication with users according to claim 2, characterized in that: Each node in the emotion graph represents a semantic emotion entity, and each edge in the emotion graph represents a dependency relationship between semantic emotion entities.

4. The cigarette marketing customer service system that facilitates quick communication with users according to claim 2, characterized in that: The similarity between the emotion feature parameter set and the reference feature parameters included in the emotion determination library is calculated.

5. The cigarette marketing customer service system that facilitates quick communication with users according to claim 4, characterized in that: The step of analyzing the response method corresponding to the customer service request according to the emotion information, the demand information, and the business module information of the user to obtain a customer service request response result includes: According to the business module information, filtering out historical data rows related to the business module information from a cigarette marketing database; Performing adjacent mean processing on the data in the historical data row to form adjacent data rows; Performing importance analysis on each row of data in the adjacent data rows by using a wrapping feature extraction algorithm to determine importance data corresponding to each row of data; Performing data drift detection on the importance data corresponding to each row of data to generate a data drift detection result; Based on the importance data corresponding to each row of data and the data drift detection result, feature extraction processing is performed on the adjacent data rows to obtain key reference features; After performing residual processing of different dimensions on the key reference features, convolution processing is performed together with the demand information and the emotion information to obtain initial reference features; The initial reference feature is processed by dimensional feature mapping to obtain an initial node feature vector; The better solutions in the initial node feature vector are screened based on a preset screening coefficient to analyze the response method corresponding to the customer service request to obtain the customer service request response result.

6. The cigarette marketing customer service system that facilitates quick communication with users according to claim 5, characterized in that: After the key reference features are subjected to residual processing of different dimensions, convolution processing is performed together with the demand information and the emotion information to obtain initial reference features, including: The key reference features are split based on three spatial dimensions to obtain horizontal attention weight features, vertical attention weight features and deep attention weight features; The horizontal attention weight feature, the vertical attention weight feature and the depth attention weight feature are respectively fused with the demand feature to obtain a first fused reference feature; wherein the demand feature is extracted from the demand information; The horizontal attention weight feature, the vertical attention weight feature and the deep attention weight feature are respectively fused with the emotion feature parameter set to obtain a second fused reference feature; The first fused reference feature and the second fused reference feature are processed using a convolutional layer of a deep network structure to obtain the initial reference feature.

7. The cigarette marketing customer service system for facilitating quick communication with users according to claim 1, characterized in that: The cigarette marketing customer service system that facilitates quick communication with users also includes: A cigarette marketing customer service end guidance providing unit is used for the cigarette marketing customer service end to obtain the terminal construction guidance request sent by the user through the data channel, and determine the store business district positioning information and business conditions based on the user's location information carried in the terminal construction guidance request; use the big data technology to determine the surrounding store facility data, surrounding store operation data, surrounding store store display data and business district population flow data according to the store business district positioning information; analyze the user's store characteristics based on the surrounding store facility data, the surrounding store operation data, the surrounding store store display data and the business district population flow data, and provide information support and guidance for the user's store shelf display.

8. The cigarette marketing customer service system that facilitates quick communication with users according to claim 5, characterized in that: The cigarette marketing database includes multiple types of the following data: cigarette brand, cigarette packaging form, cigarette packaging color, cigarette type, filter type, tar content, nicotine content, filter length, cigarette circumference, cigarette name, sales volume, inventory volume, purchase price and selling price.

9. The cigarette marketing customer service system for facilitating quick communication with users according to claim 1, characterized in that: The cigarette marketing customer service system that facilitates quick communication with users also includes: The social platform communication unit is used to obtain the real-time dynamic information posted by the user on the social media platform, and to communicate with the user in real time based on the real-time dynamic information.

Images