Intelligent customer service guidance method, system and storage medium based on customer service robot

Abstract

The present invention relates to the technical field of intelligent customer service guidance, and specifically to an intelligent customer service guidance method, system and storage medium based on a customer service robot, comprising: collecting customer service robot question data and answer data, and performing data processing on the collected data; achieving feature matching between questions and answers based on the processed data; and achieving intelligent customer service guidance based on the answer data matching results; the present invention ensures the accurate correspondence between questions and answers through a timestamp matching verification mechanism to avoid data quality problems caused by time sequence disorder, and introduces a dynamic re-collection mechanism to ensure the accuracy and reliability of collected data, providing a high-quality data basis for subsequent matching of questions and answers; by using a graph structure model to model the correlation between question features and answer features, the complex correlation relationship between question and answer features is captured intuitively and efficiently, significantly improving the matching accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent customer service guidance, and specifically to an intelligent customer service guidance method, system and storage medium based on a customer service robot. Background Art

[0002] In traditional customer service systems, users’ questions are usually answered through rule matching or keyword retrieval. However, these methods have the following defects:

[0003] Traditional systems have difficulty ensuring the temporal consistency or content relevance between the question data and the answer data entered by the user, resulting in low-quality matching results;

[0004] The existing keyword matching methods rely on static rules and cannot handle complex semantic association problems. Conventional technologies are difficult to deal with multi-level and multi-semantic problems.

[0005] The current customer service system lacks targeted guidance when answer matching fails, resulting in poor user experience. In particular, questions that are beyond the robot's answer range are difficult to divert or recommend in a timely manner. Summary of the invention

[0006] In view of the above-mentioned problems, the present invention is proposed.

[0007] To solve the above technical problems, the present invention provides the following technical solutions: an intelligent customer service guidance method based on a customer service robot, comprising the following steps:

[0008] Collect customer service robot question data and answer data, and process the collected data, specifically:

[0009] Collect the text data of questions entered by customers and the text data of answers output by the customer service robot, construct the text data into corresponding data sets, and use data processing algorithms to process the constructed data sets;

[0010] Based on the processed data, feature matching between questions and answers is achieved, specifically:

[0011] Construct answer feature vectors and question feature vectors respectively, and construct product feature vectors based on product data. Use the feature values ​​in the product feature vectors and the feature values ​​in the answer feature vectors as nodes, and use the associations between nodes as edges to construct a graph structure model. Use the similarity between the feature values ​​to determine whether the edges of the graph structure model can be established between the nodes. Use the number of graph structure edges established between the product feature values ​​and the answer feature values ​​to achieve matching between the answer and the product. At the same time, use the matching results between the question features and the product to achieve feature matching between the question and the answer.

[0012] And, based on the answer data matching results, intelligent customer service guidance is achieved, specifically:

[0013] An answer scoring mechanism is introduced to intelligently score the answer data corresponding to each question feature, and intelligent matching between question data and the best answer is achieved based on the intelligent scoring results. At the same time, different intelligent customer service guidance plans are formulated for different answer data matching results to achieve intelligent matching between question data and the best answer.

[0014] As a preferred solution of the intelligent customer service guidance method, system and storage medium based on a customer service robot described in the present invention, the data processing of the constructed data set using the data processing algorithm is specifically as follows:

[0015] Timestamp of the recorded problem data and the timestamp of the answer data , the time matching is completed through the difference of timestamps, specifically:

[0016]

[0017] in, Indicates The timestamp of the question data. Indicates The timestamp of the answer data. Represents the time difference between the question data and the corresponding answer data, which is used to match the collection time. Then,

[0018] Setting the time difference threshold , and complete the acquisition time matching based on the comparison result between the set threshold and the calculated time difference, specifically:

[0019] If the comparison result between the two satisfies the formula , it means that the collected answer data matches the question data;

[0020] If the comparison result between the two satisfies the formula , it means that the collected answer data does not match the question data. The answer data corresponding to the question data is collected again, and the time difference calculation is performed on the re-collected answer data until the collection time matching formula is satisfied.

[0021] As a preferred solution of the intelligent customer service guidance method, system and storage medium based on a customer service robot described in the present invention, the graph structure model is specifically as follows:

[0022] The eigenvalues ​​in the product feature vector are and the eigenvalues ​​in the answer eigenvector As nodes, the correlation between nodes is used as edges to construct a graph structure model. According to the constructed graph structure model, the correlation between products and answers is judged, then,

[0023] According to the similarity between the eigenvalues, we can determine whether a graph structure edge relationship can be established between two nodes. Then,

[0024]

[0025] in, represents the eigenvalue in the product feature vector, represents the eigenvalue in the answer eigenvector, Represents the similarity between two eigenvalues, which is used to determine whether a graph structure edge relationship can be established between nodes. Specifically:

[0026] Setting the similarity threshold , if the comparison result between the set similarity threshold and the calculated similarity satisfies the formula , indicating that the similarity between the two feature values ​​does not meet the feature matching requirements, and the relationship between the two nodes cannot be established. The relationship between the edges of the graph structure;

[0027] If the comparison result between the set similarity threshold and the calculated similarity satisfies the formula , indicating that the similarity between the two feature values ​​meets the feature matching requirements, and the relationship between the two nodes can establish a graph structure edge, and establish The relationship between the edges of the graph structure.

[0028] As a preferred solution of the intelligent customer service guidance method, system and storage medium based on a customer service robot described in the present invention, the matching of answers and products is specifically as follows:

[0029] For any eigenvalue in the answer eigenvector , for the edge of the graph structure constructed between it and the product feature value, if the formula is satisfied When , it means that the eigenvalue in the current answer eigenvector can establish more than two graph structure edge relationships with the eigenvalue in the product vector, which means that one eigenvalue in the current answer vector can match multiple eigenvalues ​​in the product vector, and the current answer data matches the product;

[0030] If the formula cannot be satisfied When , it means that the eigenvalue in the current answer eigenvector can establish no more than two variable relationships of the graph structure with the eigenvalue in the product vector. This means that one eigenvalue in the current answer vector can only match one eigenvalue in the product vector at most, and the current answer data does not match the product.

[0031] As a preferred solution of the intelligent customer service guidance method, system and storage medium based on a customer service robot described in the present invention, the feature matching between the question and the answer is specifically as follows:

[0032] For the feature values ​​corresponding to the answer data that can match the product , select the question feature value corresponding to the current answer feature , calculate the similarity between the current problem feature value and the product feature value , if the calculated similarity satisfies the formula , it means that the current question features match the product, and the current question features match the answer features;

[0033] If the calculated similarity satisfies the formula , it means that the current question characteristics do not match the product, the current question does not match the product, it is beyond the answer range of the customer service robot, and the current question characteristics do not match the answer characteristics;

[0034] For the feature values ​​corresponding to the answer data that cannot be matched with the product , select the question feature value corresponding to the current answer feature , calculate the similarity between the current problem feature value and the product feature value , if the calculated similarity satisfies the formula , it means that the current question features match the product, but the answer features do not match the product, and the current question features do not match the answer features;

[0035] If the calculated similarity satisfies the formula , it means that the current question features do not match the product, and the answer features do not match the product either. The current question data and answer data are beyond the answer range of the customer service robot.

[0036] As a preferred solution of the intelligent customer service guidance method, system and storage medium based on a customer service robot described in the present invention, the answer scoring mechanism introduced to intelligently score the answer data corresponding to each question feature is specifically as follows:

[0037] After each answer given by the customer service robot, collect the user's satisfaction score for the current answer At the same time, the time difference between the current question data and the corresponding answer data is collected , calculate the intelligent score of the current answer feature, then we have,

[0038]

[0039] in, Indicates the similarity between the current problem feature value and the product feature value, Indicates the similarity between the current answer features and the product features, Indicates the similarity between the current answer feature and the question feature, Indicates the user's satisfaction rating for the current answer. Indicates the time difference between the current question data and the corresponding answer data, which is the response time of the current answer feature to the question feature. , They represent weight coefficients respectively, Indicates Intelligent scoring of answer features is used to achieve intelligent matching between question data and answer features.

[0040] As a preferred solution of the intelligent customer service guidance method, system and storage medium based on a customer service robot described in the present invention, wherein: the intelligent matching between the question data and the best answer is specifically as follows:

[0041] For the intelligent scoring of the answer features corresponding to the calculated question data, the same question is asked multiple times, and the intelligent score of each answer is calculated, and the calculated intelligent score is constructed as a data set. ,in, Indicates The first intelligent scoring of answer features, Indicates The second intelligent scoring of answer features, Indicates The answer feature Smart scores and select the one with the highest smart score The corresponding answer is taken as the best answer to the current question;

[0042] For question data that cannot be matched with answer features , extract the answer features corresponding to the question data at this time , if the question data at this time matches the product, then the current question data is intelligently matched with the answer data in the historical database, and the intelligent score corresponding to the answer feature in the historical database is calculated, and the answer feature with the highest score is selected as the best answer to the current question data;

[0043] If the question data at this time does not match the product, it means that the current question is beyond the answer range of the customer service robot. At this time, the current customer’s historical question data is collected, and based on the matching degree between the historical question data and the product, the product with the highest matching degree is automatically recommended to achieve the guidance of intelligent customer service.

[0044] Intelligent customer service guidance system, including a data collection and processing module, a question-answer intelligent matching module and an intelligent customer service guidance module;

[0045] The data collection and processing module is used to collect the question text data input by the customer and the answer text data of the customer service robot, and construct the collected data into a corresponding data set;

[0046] The question-answer intelligent matching module compares and verifies the question features with the answer features according to the features in the data set, and realizes intelligent matching of the question and the answer according to the comparison and verification results;

[0047] The intelligent customer service guidance module realizes intelligent matching between question data and the best answer based on the matching results between questions and answers and through an intelligent guidance mechanism.

[0048] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, an intelligent customer service guidance method based on a customer service robot is implemented.

[0049] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of an intelligent customer service guidance method based on a customer service robot.

[0050] Beneficial effects of the present invention:

[0051] The present invention ensures the accurate correspondence between questions and answers through the matching verification mechanism of timestamps, avoids data quality problems caused by time sequence disorder, and introduces a dynamic re-collection mechanism to automatically adjust when time matching fails, thereby improving the integrity of data collection, ensuring the accuracy and reliability of collected data, and providing a high-quality data foundation for subsequent matching of questions and answers;

[0052] By using a graph structure model to model the correlation between question features and answer features, with feature values ​​as nodes and similarity as edge weights, the matching degree between questions and answers can be accurately measured, and the complex correlation between question and answer features can be captured intuitively and efficiently, significantly improving matching accuracy.

[0053] Through the dynamic edge construction mechanism of feature matching (determination of similarity threshold), intelligent dynamic matching of questions and answers is achieved, making the matching process more flexible and adaptable to various scenarios and question types;

[0054] By introducing an intelligent scoring mechanism, we comprehensively consider user satisfaction, response time, and the feature similarity between questions and answers to generate intelligent scoring results. Through multiple scoring and optimal selection strategies, we ensure accurate matching of questions and the best answers. This not only improves the credibility of the matching results, but also dynamically adjusts the recommendation results based on real-time feedback from users, and provides historical data-related recommendations for unmatched questions, thus realizing intelligent and diversified guidance methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0056] Figure 1 The present invention is a schematic diagram of the overall method step structure of an intelligent customer service guidance method, system and storage medium based on a customer service robot. DETAILED DESCRIPTION

[0057] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0058] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0059] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0060] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0061] Meanwhile, in the description of the present invention, it should be noted that the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0062] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0063] Example 1

[0064] Reference Figure 1 , as an embodiment of the present invention, provides an intelligent customer service guidance method based on a customer service robot, comprising the following steps:

[0065] S1: Collect customer service robot question data and answer data, and process the collected data.

[0066] Specifically, the collection of customer service robot question data and answer data is to collect the question text data received by the current robot within a certain time range, and the corresponding answer text data given after receiving the question data, and construct the collected text data into a corresponding data set, then,

[0067] Problem data,

[0068] Answer data,

[0069] in, Indicates the upper limit of the amount of data collected. Indicates the collected Question data, represents the constructed problem data set, Indicates the collected Answer data, Represents the constructed answer data set.

[0070] Furthermore, the data processing of the collected data is to process the constructed data set through a data processing algorithm to ensure the accuracy of the data, and the specific processing is as follows:

[0071] Match and verify the collection time of the collected question data and answer data, then we have:

[0072] Timestamp of the recorded problem data and the timestamp of the answer data , the time matching is completed through the difference of timestamps, specifically:

[0073]

[0074] in, Indicates The timestamp of the question data. Indicates The timestamp of the answer data. Represents the time difference between the question data and the corresponding answer data, which is used to match the collection time. Then,

[0075] Setting the time difference threshold , and complete the acquisition time matching based on the comparison result between the set threshold and the calculated time difference, specifically:

[0076] If the comparison result between the two satisfies the formula , it means that the collected answer data matches the question data;

[0077] If the comparison result between the two satisfies the formula , it means that the collected answer data does not match the question data. The answer data corresponding to the question data is collected again, and the time difference calculation is performed on the re-collected answer data until the collection time matching formula is satisfied.

[0078] S2: Based on the processed data, feature matching between questions and answers is achieved.

[0079] Specifically, the feature matching between the question and the answer is to compare and verify the question features with the answer features, and based on the comparison and verification results, achieve intelligent matching of the question and the answer.

[0080] Furthermore, the answer feature is based on the constructed answer data set , calculate the correlation between the product and answer data sets, and build the association relationship based on the constructed correlation. The specific implementation is as follows:

[0081] Extract product features from product data, including type, specification, and price, and construct product feature vectors , and satisfies the formula ,in, Indicates eigenvalues, The dimension that represents the product characteristics, which is also the upper limit of the types of product characteristic values;

[0082] For each answer in the answer data set , extract its features and construct the corresponding feature vector, ,in, Indicates The answer data in eigenvalues, The dimension that represents the answer feature, which is also the upper limit of the type of feature values ​​in the answer data, and extracts the question corresponding to each answer , and construct the corresponding feature vector ,in, Indicates The first question in the data eigenvalues;

[0083] According to the constructed feature vector, the correlation between the product and the answer is calculated as follows:

[0084] The eigenvalues ​​in the product feature vector are and the eigenvalues ​​in the answer eigenvector As nodes, the correlation between nodes is used as edges to construct a graph structure model. According to the constructed graph structure model, the correlation between products and answers is judged, then,

[0085] According to the similarity between the eigenvalues, we can determine whether a graph structure edge relationship can be established between two nodes. Then,

[0086]

[0087] in, represents the eigenvalue in the product feature vector, represents the eigenvalue in the answer eigenvector, Represents the similarity between two eigenvalues, which is used to determine whether a graph structure edge relationship can be established between nodes. Specifically:

[0088] Setting the similarity threshold , if the comparison result between the set similarity threshold and the calculated similarity satisfies the formula , indicating that the similarity between the two feature values ​​does not meet the feature matching requirements, and the relationship between the two nodes cannot be established. The relationship between the edges of the graph structure;

[0089] If the comparison result between the set similarity threshold and the calculated similarity satisfies the formula , indicating that the similarity between the two feature values ​​meets the feature matching requirements, and the relationship between the two nodes can establish a graph structure edge, and establish The relationship between the edges of the graph structure.

[0090] It should be noted that for two nodes that can establish a graph structure edge relationship, the number of graph structure edges established between the product feature value and the answer feature value is used to achieve matching between the answer and the product, specifically:

[0091] For any eigenvalue in the answer eigenvector , for the edge of the graph structure constructed between it and the product feature value, if the formula is satisfied When , it means that the eigenvalue in the current answer eigenvector can establish more than two graph structure edge relationships with the eigenvalue in the product vector, which means that one eigenvalue in the current answer vector can match multiple eigenvalues ​​in the product vector, and the current answer data matches the product;

[0092] If the formula cannot be satisfied When , it means that the eigenvalue in the current answer eigenvector can establish no more than two variable relationships of the graph structure with the eigenvalue in the product vector. This means that one eigenvalue in the current answer vector can only match one eigenvalue in the product vector at most, and the current answer data does not match the product.

[0093] Furthermore, the feature matching between the question and the answer is achieved based on the matching result between the question data and the product, and the specific implementation is as follows:

[0094] For the feature values ​​corresponding to the answer data that can match the product , select the question feature value corresponding to the current answer feature , calculate the similarity between the current problem feature value and the product feature value , if the calculated similarity satisfies the formula , it means that the current question features match the product, and the current question features match the answer features;

[0095] If the calculated similarity satisfies the formula , it means that the current question characteristics do not match the product, the current question does not match the product, it is beyond the answer range of the customer service robot, and the current question characteristics do not match the answer characteristics;

[0096] For the feature values ​​corresponding to the answer data that cannot be matched with the product , select the question feature value corresponding to the current answer feature , calculate the similarity between the current problem feature value and the product feature value , if the calculated similarity satisfies the formula , it means that the current question features match the product, but the answer features do not match the product, and the current question features do not match the answer features;

[0097] If the calculated similarity satisfies the formula , it means that the current question features do not match the product, and the answer features do not match the product either. The current question data and answer data are beyond the answer range of the customer service robot.

[0098] S3: Based on the answer data matching results, intelligent customer service guidance is realized.

[0099] Specifically, the intelligent customer service guidance is achieved by using an intelligent guidance mechanism to achieve intelligent matching between question data and the best answer based on the matching result of the answer data.

[0100] Furthermore, the intelligent guidance mechanism is to intelligently score the answer data corresponding to each question feature by introducing an answer scoring mechanism, and to achieve intelligent matching between the question data and the best answer based on the intelligent scoring result, which is specifically implemented as follows:

[0101] For question data that can match answer features , extract the answer features corresponding to the question data at this time , intelligently score the current answer features, then we have,

[0102] The intelligent scoring of answer features is based on the combined evaluation results of user evaluation survey results and interference factors, specifically:

[0103] After each answer given by the customer service robot, collect the user's satisfaction score for the current answer At the same time, the time difference between the current question data and the corresponding answer data is collected , calculate the intelligent score of the current answer feature, then we have,

[0104]

[0105] in, Indicates the similarity between the current problem feature value and the product feature value, Indicates the similarity between the current answer features and the product features, Indicates the similarity between the current answer feature and the question feature, Indicates the user's satisfaction rating for the current answer. Indicates the time difference between the current question data and the corresponding answer data, which is the response time of the current answer feature to the question feature. , They represent weight coefficients respectively, which are set by the implementers according to the actual application scenarios. Indicates Intelligent scoring of answer features is used to achieve intelligent matching between question data and answer features, specifically:

[0106] For the intelligent scoring of the answer features corresponding to the calculated question data, in order to improve the accuracy of intelligent matching, the same question is asked multiple times, and the intelligent score of each answer is calculated, and the calculated intelligent score is constructed as a data set. ,in, Indicates The first intelligent scoring of answer features, Indicates The second intelligent scoring of answer features, Indicates The answer feature Smart scores and select the one with the highest smart score The corresponding answer is taken as the best answer to the current question;

[0107] For question data that cannot be matched with answer features , extract the answer features corresponding to the question data at this time , if the question data at this time matches the product, then the current question data is intelligently matched with the answer data in the historical database, and the intelligent score corresponding to the answer feature in the historical database is calculated, and the answer feature with the highest score is selected as the best answer to the current question data;

[0108] If the question data at this time does not match the product, it means that the current question is beyond the answer range of the customer service robot. At this time, the current customer’s historical question data is collected, and based on the matching degree between the historical question data and the product, the product with the highest matching degree is automatically recommended to achieve the guidance of intelligent customer service.

[0109] Example 2

[0110] A second embodiment of the present invention provides an intelligent customer service system based on a customer service robot, including a data collection and processing module, a question-answer intelligent matching module, and an intelligent customer service guidance module;

[0111] Specifically, the data collection and processing module is used to collect the question text data input by the customer and the answer text data of the customer service robot, and construct the collected data into a corresponding data set; the question-answer intelligent matching module compares and verifies the question features with the answer features according to the features in the data set, and realizes intelligent matching of questions and answers according to the comparison and verification results; the intelligent customer service guidance module realizes intelligent matching between question data and the best answer according to the matching results of questions and answers and through an intelligent guidance mechanism.

[0112] Furthermore, the data collection and processing module performs denoising, word segmentation and formatting on the collected data, removes invalid characters or redundant information, and supports the collection of multimodal data such as text, voice, and pictures, providing compatibility for complex scenarios. At the same time, it records the anomalies in the collection process (such as collection failure or matching timeout) in real time to facilitate subsequent optimization; the data collection and processing module provides a reliable basis for feature comparison and modeling of the question answer intelligent matching module by providing a structured, high-quality data set, and ensures data consistency in the matching process through timestamp verification and dynamic collection mechanisms;

[0113] The question-answer intelligent matching module is to construct question features and answer features, extract product features from product data, respectively calculate the correlation between answer features and product features, the correlation between question features and answer features, and the correlation between question features and product features, and build a graph structure model. At the same time, based on the correlation between answer features and product features, the matching relationship between answers and products is judged. Finally, intelligent matching between questions and answers is achieved based on the correlation between question features and answer features and the correlation between question features and product features. The question-answer intelligent matching module achieves efficient matching between questions and answers through feature modeling, similarity measurement and graph structure modeling, and at the same time combines a dynamic adjustment mechanism to adapt to diverse question requirements in complex scenarios.

[0114] The intelligent customer service guidance module introduces an answer scoring mechanism to intelligently score the answer data corresponding to each question feature, and realizes intelligent matching between the question data and the best answer based on the intelligent scoring result. It also performs intelligent scoring on the answer data corresponding to each question feature, and realizes intelligent matching between the question data and the best answer based on the intelligent scoring result. Different guidance plans are formulated based on the matching relationship between different questions and answers to achieve matching between questions and the best answers. By introducing an intelligent scoring mechanism, user satisfaction, response time, and feature similarity between questions and answers are comprehensively considered to generate intelligent scoring results. Through multiple scoring and optimal selection strategies, accurate matching between questions and the best answers is guaranteed, which not only improves the credibility of the matching results, but also dynamically adjusts the recommendation results based on real-time feedback from users, and provides historical data association recommendations for unmatched questions, thereby realizing intelligent and diversified guidance methods.

[0115] Furthermore, if the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0116] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0117] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0118] Additionally, in order to provide a concise description of exemplary embodiments, all features of an actual embodiment may not be described (ie, those features that are not relevant to the best mode presently contemplated for carrying out the invention or those features that are not relevant to implementing the invention).

[0119] It will be appreciated that in the development of any actual implementation, as in any engineering or design project, numerous implementation-specific decisions may be made. Such a development effort may be complex and time-consuming, but will be a routine task of design, fabrication, and production for those of ordinary skill having the benefit of this disclosure without undue experimentation.

[0120] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An intelligent customer service guidance method based on a customer service robot, characterized in that: The following steps are included: Collect customer service robot question data and answer data, and process the collected data, specifically: Collect the text data of questions entered by customers and the text data of answers output by the customer service robot, construct the text data into corresponding data sets, and use data processing algorithms to process the constructed data sets; The data processing algorithm used to process the constructed data set is specifically as follows: The timestamp T(Q i ) and the timestamp T(A) of the answer data i ), the time matching is completed through the difference of timestamps, specifically: ΔT=T(Q i )-T(A i ) Among them, T(Q i ) represents the timestamp of the i-th question data, T(A i ) represents the timestamp of the i-th answer data, ΔT represents the time difference between the question data and the corresponding answer data, which is used to match the collection time, then, Set the time difference threshold ΔT′, and complete the acquisition time matching based on the comparison result between the set threshold and the calculated time difference, specifically: If the comparison result between the two satisfies the formula ΔT≤ΔT′, it means that the collected answer data matches the question data; If the comparison result between the two satisfies the formula ΔT>ΔT′, it means that the collected answer data does not match the question data, and the answer data corresponding to the question data is collected again, and the time difference calculation is performed on the re-collected answer data until the collection time matching formula is satisfied; Based on the processed data, feature matching between questions and answers is achieved, specifically: Construct answer feature vectors and question feature vectors respectively, and construct product feature vectors based on product data. Use the feature values ​​in the product feature vectors and the feature values ​​in the answer feature vectors as nodes, and use the associations between nodes as edges to construct a graph structure model. Use the similarity between the feature values ​​to determine whether the edges of the graph structure model can be established between the nodes. Use the number of graph structure edges established between the product feature values ​​and the answer feature values ​​to achieve matching between the answer and the product. At the same time, use the matching results between the question features and the product to achieve feature matching between the question and the answer. And, based on the answer data matching results, intelligent customer service guidance is achieved, specifically: An answer scoring mechanism is introduced to intelligently score the answer data corresponding to each question feature, and intelligent matching between question data and the best answer is achieved based on the intelligent scoring results. At the same time, different intelligent customer service guidance plans are formulated for different answer data matching results to achieve intelligent matching between question data and the best answer.

2. The intelligent customer service guidance method based on a customer service robot according to claim 1, characterized in that: The graph structure model is specifically constructed as follows: The eigenvalues ​​p in the product feature vector are j And the eigenvalue a in the answer eigenvector ik As nodes, the correlation between nodes is used as edges to construct a graph structure model. According to the constructed graph structure model, the correlation between products and answers is judged, then, According to the similarity between the eigenvalues, we can determine whether a graph structure edge relationship can be established between two nodes. Then, Among them, p j represents the eigenvalue in the product feature vector, a ik represents the eigenvalue in the answer eigenvector, S(p j ,a ik ) represents the similarity between two eigenvalues, which is used to determine whether a graph structure edge relationship can be established between nodes. Specifically: Set the similarity threshold S T , if the comparison result between the set similarity threshold and the calculated similarity satisfies the formula S(p j ,a ik )<S T , indicating that the similarity between the two feature values ​​does not meet the feature matching requirements, and the relationship between the two nodes cannot be established. j ,a ik )=0; If the comparison result between the set similarity threshold and the calculated similarity satisfies the formula S(p j ,a ik )≥S T , indicating that the similarity between the two feature values ​​meets the feature matching requirements, the relationship between the two nodes can establish a graph structure edge, and establish E(p j ,a ik )=1.

3. The intelligent customer service guidance method based on a customer service robot as claimed in claim 2, characterized in that: The matching between the implementation answer and the product is as follows: For any eigenvalue a in the answer eigenvector ik , for the edge of the graph structure constructed between it and the product feature value, if the formula (E(p j ,a ik )=1)∧(E(p f≠j ,a ik )=1), it means that the eigenvalue in the current answer eigenvector can establish more than two graph structure edge relationships with the eigenvalue in the product vector, which means that one eigenvalue in the current answer vector can match multiple eigenvalues ​​in the product vector, and the current answer data matches the product; If the formula (E(p j ,a ik )=1)∧(E(p f≠j ,a ik )=1), it means that the eigenvalue in the current answer feature vector can establish no more than two edge relationships with the eigenvalue in the product vector, which means that one eigenvalue in the current answer vector can only match one eigenvalue in the product vector at most, and the current answer data does not match the product.

4. The intelligent customer service guidance method based on a customer service robot as claimed in claim 3, characterized in that: The feature matching between the question and the answer is specifically implemented as follows: For the feature value a corresponding to the answer data that can match the product ik , select the question feature value q corresponding to the current answer feature ik , calculate the similarity S(p j ,q ik ), if the calculated similarity satisfies the formula S(p j ,q ik )≥S T , it means that the current question features match the product, and the current question features match the answer features; If the calculated similarity satisfies the formula S(p j ,q ik )<S T , it means that the current question characteristics do not match the product, the current question does not match the product, it is beyond the answer range of the customer service robot, and the current question characteristics do not match the answer characteristics; For the answer data that cannot be matched with the product, the corresponding feature value a ik , select the question feature value q corresponding to the current answer feature ik , calculate the similarity S(p j ,q ik ), if the calculated similarity satisfies the formula S(p j ,q ik )≥S T , it means that the current question features match the product, but the answer features do not match the product, and the current question features do not match the answer features; If the calculated similarity satisfies the formula S(p j ,q ik )<S T , it means that the current question features do not match the product, and the answer features do not match the product either. The current question data and answer data are beyond the answer range of the customer service robot.

5. The intelligent customer service guidance method based on a customer service robot according to claim 4, characterized in that: The answer scoring mechanism introduced to intelligently score the answer data corresponding to each question feature is as follows: After each answer given by the customer service robot, collect the user's satisfaction score S for the current answer a At the same time, the time difference ΔT between the current question data and the corresponding answer data is collected, and the intelligent score of the current answer feature is calculated, then, S i =S(p j ,q ik )+S(p j ,a ik )+S(q ik ,a ik )+w1·S a +w2·ΔT Among them, S(p j ,q ik ) represents the similarity between the current problem feature value and the product feature value, S(p j ,a ik ) represents the similarity between the current answer feature and the product feature, S(q ik ,a ik ) represents the similarity between the current answer feature and the question feature, S a represents the user's satisfaction score for the current answer, ΔT represents the time difference between the current question data and the corresponding answer data, and is the response time of the current answer feature to the question feature. w1 and w2 represent weight coefficients respectively. S i Represents the intelligent score of the i-th answer feature, which is used to achieve intelligent matching between question data and answer features.

6. The intelligent customer service guidance method based on a customer service robot according to claim 5, characterized in that: The intelligent matching between the question data and the best answer is specifically implemented as follows: For the intelligent scoring of the answer features corresponding to the calculated question data, the same question is asked multiple times, and the intelligent score of each answer is calculated. The calculated intelligent score is constructed as a data set. [S i1 ,S i2 ,...,S in ], where S i1 represents the first intelligent scoring of the i-th answer feature, S i2 represents the second intelligent scoring of the i-th answer feature, S in represents the nth intelligent score of the feature of the ith answer, and selects the max([S i1 ,S i2 ,...,S in ]) The answer corresponding to the question is the best answer to the current question; For question data Q that cannot be matched with answer features i , extract the answer feature A corresponding to the question data at this time i , if the question data at this time matches the product, then the current question data is intelligently matched with the answer data in the historical database, and the intelligent score corresponding to the answer feature in the historical database is calculated, and the answer feature with the highest score is selected as the best answer to the current question data; If the question data at this time does not match the product, it means that the current question is beyond the answer range of the customer service robot. At this time, the current customer’s historical question data is collected, and based on the matching degree between the historical question data and the product, the product with the highest matching degree is automatically recommended to achieve the guidance of intelligent customer service.

7. An intelligent customer service guidance system based on a customer service robot according to any one of claims 1 to 6, characterized in that: Including data collection and processing module, question and answer intelligent matching module and intelligent customer service guidance module; The data collection and processing module is used to collect the question text data input by the customer and the answer text data of the customer service robot, and construct the collected data into a corresponding data set; The question-answer intelligent matching module compares and verifies the question features with the answer features according to the features in the data set, and realizes intelligent matching of the question and the answer according to the comparison and verification results; The intelligent customer service guidance module realizes intelligent matching between question data and the best answer based on the matching results between questions and answers and through an intelligent guidance mechanism.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images