After-sales data management system and management method thereof

Abstract

An after-sales data management system and a management method thereof. First, the work order text part extracted from the after-sales work order database is semantically encoded based on word granularity to obtain a sequence of work order text word granularity semantic encoding feature vectors, then, the information description of the candidate maintenance technician extracted from the after-sales maintenance technician database is semantically encoded based on word granularity to obtain a sequence of candidate maintenance technician information description word granularity semantic encoding feature vectors, then, the local matching coefficient and the global matching coefficient between the sequence of the work order text word granularity semantic encoding feature vectors and the sequence of the candidate maintenance technician information description word granularity semantic encoding feature vectors are calculated, and finally, the weighted sum of the local matching coefficient and the global matching coefficient is calculated to obtain a comprehensive matching coefficient, and based on the comprehensive matching coefficient, it is determined whether to refer to the candidate maintenance technician to handle the work order to be assigned.

Description

Technical Field

[0001] The present disclosure relates to the field of after-sales data management, and in particular to an after-sales data management system and a management method thereof. Background Art

[0002] After-sales data management refers to the management process of recording, handling and tracking questions, requests or faults raised by customers after the company sells products or provides services. The after-sales data management system is a system used to effectively manage and process after-sales work orders, customer feedback and other information, aiming to improve customer satisfaction, optimize after-sales service processes, and enhance the company's service quality and efficiency.

[0003] However, in traditional after-sales data management systems, work orders usually need to be distributed by manual operators based on experience or simple rules. This method has problems such as uneven distribution, omissions and errors, resulting in low efficiency in work order processing. In addition, the manual distribution method lacks the support of intelligent technology, cannot conduct in-depth analysis and processing of work order content, and cannot comprehensively consider customer needs and problems, resulting in failure to meet customer needs well, affecting the effectiveness and timeliness of work order processing. Moreover, with the expansion of enterprise scale and increase in business volume, the traditional manual dispatching method is difficult to cope with the processing needs of a large number of work orders, which easily leads to work order accumulation and delays.

[0004] Therefore, an optimized after-sales data management system is desired, which can realize intelligent work order dispatching and processing. Summary of the invention

[0005] The present disclosure is made in view of the above problems. One object of the present disclosure is to provide an after-sales data management system and a management method thereof.

[0006] An embodiment of the present disclosure provides an after-sales data management system, which includes:

[0007] The work order data extraction module to be assigned is used to extract the work orders to be assigned from the after-sales work order database;

[0008] A work order text part extraction module, used to extract the text part from the work order to be assigned to obtain the work order text part, wherein the work order text part includes basic customer information, customer request, problem description and contact information;

[0009] A work order text semantic understanding module, used for performing word granularity-based semantic encoding on the work order text portion to obtain a sequence of work order text word granularity semantic encoding feature vectors;

[0010] A candidate maintenance technician information extraction module is used to extract information descriptions of candidate maintenance technicians from an after-sales maintenance technician database;

[0011] A candidate maintenance technician information semantic understanding module is used to perform word granularity-based semantic coding on the candidate maintenance technician information description to obtain a sequence of word granularity semantic coding feature vectors of the candidate maintenance technician information description;

[0012] A local matching coefficient calculation module, used to calculate the local matching coefficient between the sequence of word granularity semantic coding feature vectors of the work order text and the sequence of word granularity semantic coding feature vectors of the candidate maintenance technician information description;

[0013] A global matching coefficient calculation module, used to calculate the global matching coefficient between the sequence of word-granular semantic coding feature vectors of the work order text and the sequence of word-granular semantic coding feature vectors of the candidate maintenance technician information description; and

[0014] The after-sales work order dispatching module is used to calculate the weighted sum of the local matching coefficient and the global matching coefficient to obtain a comprehensive matching coefficient, and determine whether to refer the candidate maintenance technician to handle the work order to be dispatched based on the comprehensive matching coefficient.

[0015] For example, according to the after-sales data management system of an embodiment of the present disclosure, the work order text semantic understanding module is used to:

[0016] After word segmentation processing is performed on the work order text portion, a semantic encoder including a word embedding layer is passed through to obtain a sequence of word granularity semantic encoding feature vectors of the work order text.

[0017] For example, according to the after-sales data management system of an embodiment of the present disclosure, the candidate maintenance technician information semantic understanding module is used to:

[0018] The information description of the candidate maintenance technician is processed by word segmentation and then passed through the semantic encoder of the word embedding layer to obtain a sequence of word granularity semantic encoding feature vectors of the information description of the candidate maintenance technician.

[0019] For example, according to the after-sales data management system of an embodiment of the present disclosure, the global matching coefficient calculation module includes:

[0020] A first vector cascading unit is used to cascade the sequence of the work order text word granularity semantic coding feature vectors and the sequence of the candidate maintenance technician information description word granularity semantic coding feature vectors to obtain a work order text word granularity semantic coding cascade feature vector and a candidate maintenance technician information description word granularity semantic coding cascade feature vector;

[0021] A second vector cascading unit is used to cascade the work order text word granularity semantic encoding cascade feature vector and the candidate maintenance technician information description word granularity semantic encoding cascade feature vector to obtain a work order text-candidate maintenance technician information association feature vector;

[0022] A linear transformation unit, used to use a transformation matrix and a bias vector to linearly transform the work order text-candidate maintenance technician information association feature vector to obtain a linearly transformed work order text-candidate maintenance technician information association feature vector;

[0023] An activation unit, used for activating the linear transformation work order text-candidate maintenance technician information associated feature vector using a sigmoid activation function to obtain an activation linear transformation threshold vector;

[0024] A weighting unit, configured to calculate a weighted sum of the work order text word granularity semantic encoding cascade feature vector and the candidate maintenance technician information description word granularity semantic encoding cascade feature vector using each position feature value in the activation linear transformation threshold vector as a threshold value to obtain a work order text-candidate maintenance technician information word granularity semantic matching feature vector; and

[0025] A global matching coefficient calculation unit is used to calculate the transpose of the work order text-candidate maintenance technician information word granularity semantic matching feature vector and the vector multiplication between the work order text-candidate maintenance technician information word granularity semantic matching feature vector, and then divide the obtained autocorrelation value by the square of the norm of the work order text-candidate maintenance technician information word granularity semantic matching feature vector to obtain the global matching coefficient.

[0026] For example, according to the after-sales data management system of an embodiment of the present disclosure, the linear transformation unit is used to:

[0027] After calculating the matrix multiplication of the work order text-candidate maintenance technician information associated feature vector and the transformation matrix, the obtained feature vector is added to the bias vector by position to obtain the linear transformation work order text-candidate maintenance technician information associated feature vector.

[0028] For example, according to the after-sales data management system of an embodiment of the present disclosure, the weighting unit is used to:

[0029] Calculating an exponential function of the work order text word granularity semantic encoding cascade feature vector with a natural constant as the base to obtain a work order text word granularity semantic class support feature vector;

[0030] Calculating an exponential function of the granular semantic encoding cascade feature vector of the candidate maintenance technician information description word with a natural constant as the base to obtain a granular semantic class support feature vector of the candidate maintenance technician information description word;

[0031] Calculate the product of the feature value of each position in the work order text word granularity semantic class support feature vector and the threshold value to obtain the weighted item work order text word granularity semantic class support feature vector;

[0032] Calculate the product of the feature value of each position in the candidate maintenance technician information description word granular semantic class support feature vector and one minus a threshold value to obtain a weighted candidate maintenance technician information description word granular semantic class support feature vector; and

[0033] The weighted item work order text word granularity semantic class support feature vector and the weighted item candidate maintenance technician information description word granularity semantic class support feature vector are added according to the position points to obtain the work order text-candidate maintenance technician information word granularity semantic matching feature vector.

[0034] For example, according to the after-sales data management system of an embodiment of the present disclosure, the after-sales work order dispatching module includes:

[0035] a comprehensive matching coefficient calculation unit, configured to calculate a weighted sum of the local matching coefficient and the global matching coefficient to obtain a comprehensive matching coefficient; and

[0036] The work order dispatching unit is used to determine whether to designate the candidate maintenance technician to handle the to-be-dispatched work order based on a comparison between the comprehensive matching coefficient and a predetermined threshold.

[0037] For example, according to the after-sales data management system of an embodiment of the present disclosure, in response to the comprehensive matching coefficient being greater than the predetermined threshold, it is determined to designate the candidate maintenance technician to handle the work order to be assigned.

[0038] The embodiment of the present disclosure also provides an after-sales data management method, which includes:

[0039] Extract the work orders to be dispatched from the after-sales work order database;

[0040] Extracting a text portion from the work order to be assigned to obtain a work order text portion, wherein the work order text portion includes basic customer information, customer request, problem description, and contact information;

[0041] Performing word-granularity-based semantic coding on the work order text portion to obtain a sequence of work order text word-granularity semantic coding feature vectors;

[0042] Extracting information descriptions of candidate maintenance technicians from an after-sales maintenance technician database;

[0043] Performing word-granularity-based semantic coding on the information description of the candidate maintenance technician to obtain a sequence of word-granularity semantic coding feature vectors of the information description of the candidate maintenance technician;

[0044] Calculating a local matching coefficient between a sequence of word-granular semantic encoding feature vectors of the work order text and a sequence of word-granular semantic encoding feature vectors of the candidate maintenance technician information description;

[0045] Calculating a global matching coefficient between a sequence of word-granular semantic encoding feature vectors of the work order text and a sequence of word-granular semantic encoding feature vectors of the candidate maintenance technician information description; and

[0046] A weighted sum of the local matching coefficient and the global matching coefficient is calculated to obtain a comprehensive matching coefficient, and based on the comprehensive matching coefficient, it is determined whether to designate the candidate maintenance technician to handle the work order to be assigned.

[0047] For example, according to the after-sales data management method of an embodiment of the present disclosure, the word granularity-based semantic encoding of the work order text portion is performed to obtain a sequence of work order text word granularity semantic encoding feature vectors, including:

[0048] After word segmentation processing is performed on the work order text portion, a semantic encoder including a word embedding layer is passed through to obtain a sequence of word granularity semantic encoding feature vectors of the work order text.

[0049] According to the after-sales data management system and management method of the embodiment of the present disclosure, it extracts the work order to be assigned from the after-sales work order database, and extracts the information description of the candidate maintenance technician from the after-sales maintenance technician database, and uses the data processing and text semantic understanding algorithm based on artificial intelligence and natural language processing technology to perform semantic analysis on the text part of the work order to be assigned and the information description of the candidate maintenance technician in the back end, so as to capture the text semantic features of the work order to be assigned and the information of the candidate maintenance technician, so as to match and analyze the semantic information of the two later to automatically assign the work order to the most suitable after-sales maintenance technician. In this way, the intelligent assignment and processing of work orders can be realized, so that the management of after-sales data meets the actual needs of customers, thereby improving the quality and efficiency of after-sales service. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the following briefly introduces the drawings of the embodiments of the present disclosure. Obviously, the drawings described below only relate to some embodiments of the present disclosure, and are not intended to limit the present disclosure.

[0051] Figure 1A schematic diagram of the application architecture of the after-sales data management system in an embodiment of the present disclosure is shown;

[0052] Figure 2 A schematic diagram of the structure of an after-sales data management system in an embodiment of the present disclosure is shown;

[0053] Figure 3 A schematic diagram showing the structure of a global matching coefficient calculation module of an after-sales data management system in an embodiment of the present disclosure is shown;

[0054] Figure 4 A schematic diagram of the structure of an after-sales work order dispatching module of an after-sales data management system in an embodiment of the present disclosure is shown;

[0055] Figure 5 A flow chart of the after-sales data management method in an embodiment of the present disclosure is shown;

[0056] Figure 6 An application scenario diagram of the after-sales data management system in an embodiment of the present disclosure is shown; and

[0057] Figure 7 A schematic diagram of a storage medium according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0058] The technical solutions in the embodiments of the present disclosure will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present disclosure, rather than all embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present disclosure.

[0059] The terms used in this specification are those general terms currently widely used in the art in consideration of the functions of the present disclosure, but these terms may vary according to the intention of a person of ordinary skill in the art, precedents, or new technologies in the art. In addition, specific terms may be selected by the applicant, and in this case, their detailed meanings will be described in the detailed description of the present disclosure. Therefore, the terms used in the specification should not be understood as simple names, but rather as an overall description based on the meaning of the terms and the present disclosure.

[0060] Although the present disclosure makes various references to certain modules in the system according to embodiments of the present disclosure, any number of different modules can be used and run on a user terminal and / or server. The modules are illustrative only, and different aspects of the systems and methods can use different modules.

[0061] Flowcharts are used in this disclosure to illustrate the operations performed by the system according to the embodiments of the present disclosure. It should be understood that the preceding or following operations are not necessarily performed precisely in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. At the same time, other operations may also be added to these processes, or one or more operations may be removed from these processes.

[0062] Figure 1 A schematic diagram of the application architecture of the after-sales data management system in an embodiment of the present disclosure is shown, including a server 100 and a terminal device 200.

[0063] The terminal device 200 and the server 100 can be connected via the Internet to achieve mutual communication. Optionally, the above-mentioned Internet uses standard communication technology and / or protocol. The Internet is usually the Internet, but it can also be any network, including but not limited to any combination of a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a mobile, wired or wireless network, a dedicated network or a virtual private network. In some embodiments, the data exchanged through the network is represented by technologies and / or formats including Hyper Text Markup Language (HTML), Extensible Markup Language (XML), etc. In addition, conventional encryption technologies such as Secure Socket Layer (SSL), Transport Layer Security (TLS), Virtual Private Network (VPN), Internet Protocol Security (IPsec) can also be used to encrypt all or some links. In other embodiments, customized and / or dedicated data communication technologies can also be used to replace or supplement the above-mentioned data communication technologies.

[0064] The server 100 can provide various network services for the terminal device 200, wherein the server 100 can be a single server, a server cluster consisting of several servers, or a cloud computing center. Specifically, the server 100 may include a processor 110 (Center Processing Unit, CPU), a memory 120, an input device 130, and an output device 140, etc. The input device 130 may include a keyboard, a mouse, a touch screen, etc., and the output device 140 may include a display device, such as a liquid crystal display (Liquid Crystal Display, LCD), a cathode ray tube (Cathode Ray Tube, CRT), etc.

[0065] The memory 120 may include a read-only memory (ROM) and a random access memory (RAM), and provides the processor 110 with program instructions and data stored in the memory 120. In the embodiment of the present disclosure, the memory 120 may be used to store the corresponding method of the after-sales data management system in the embodiment of the present disclosure.

[0066] The processor 110 calls the program instructions stored in the memory 120, and the processor 110 is used to execute the steps of any one of the after-sales data management methods in the embodiments of the present disclosure according to the obtained program instructions.

[0067] In addition, the application architecture diagram in the embodiment of the present disclosure is intended to more clearly illustrate the technical solution in the embodiment of the present disclosure, and does not constitute a limitation on the technical solution provided in the embodiment of the present disclosure. Of course, for other application architectures and business applications, the technical solution provided in the embodiment of the present disclosure is also applicable to similar problems.

[0068] The after-sales data management system provided according to at least one embodiment of the present disclosure is described below in a non-restrictive manner through several examples or embodiments. As described below, different features in these specific examples or embodiments can be combined with each other without conflicting with each other to obtain new examples or embodiments, and these new examples or embodiments also fall within the scope of protection of the present disclosure.

[0069] In response to the above technical problems, the technical concept of this application is to extract the work orders to be assigned from the after-sales work order database, and extract the information description of the candidate maintenance technicians from the after-sales maintenance technician database, and use the data processing and text semantic understanding algorithm based on artificial intelligence and natural language processing technology in the back end to perform semantic analysis on the text part of the work order to be assigned and the information description of the candidate maintenance technician, so as to capture the text semantic features of the work order to be assigned and the information of the candidate maintenance technician, so as to match and analyze the semantic information of the two later to automatically assign the work order to the most suitable after-sales maintenance technician. In this way, the intelligent assignment and processing of work orders can be realized, so that the management of after-sales data meets the actual needs of customers, thereby improving the quality and efficiency of after-sales service.

[0070] Figure 2 The schematic diagram of the structure of the after-sales data management system 800 in the embodiment of the present disclosure is shown. The after-sales data management system 800 includes: a work order data extraction module 810 to be assigned, which is used to extract the work order to be assigned from the after-sales work order database; a work order text part extraction module 820, which is used to extract the text part from the work order to be assigned to obtain the work order text part, wherein the work order text part includes basic customer information, customer request, problem description and contact information; a work order text semantic understanding module 830, which is used to perform word granularity-based semantic encoding on the work order text part to obtain a sequence of work order text word granularity semantic encoding feature vectors; an alternative maintenance technician information extraction module 840, which is used to extract the information description of the alternative maintenance technician from the after-sales maintenance technician database; an alternative maintenance technician information semantic understanding module 850, which is used to perform word granularity-based semantic encoding on the information description of the alternative maintenance technician to obtain a sequence of word-granular semantic coding feature vectors of the candidate maintenance technician information description words; a local matching coefficient calculation module 860, used to calculate the local matching coefficients between the sequence of word-granular semantic coding feature vectors of the work order text and the sequence of word-granular semantic coding feature vectors of the candidate maintenance technician information description words; a global matching coefficient calculation module 870, used to calculate the global matching coefficients between the sequence of word-granular semantic coding feature vectors of the work order text and the sequence of word-granular semantic coding feature vectors of the candidate maintenance technician information description words; and an after-sales work order dispatching module 880, used to calculate the weighted sum of the local matching coefficient and the global matching coefficient to obtain a comprehensive matching coefficient, and determine whether to refer to the candidate maintenance technician to handle the work order to be dispatched based on the comprehensive matching coefficient.

[0071] Specifically, in the technical solution of the present application, first, the work order to be assigned is extracted from the after-sales work order database. It should be understood that the work order to be assigned usually contains a large amount of text semantic information, including basic customer information, request content, problem description and contact information, etc. Therefore, it is necessary to extract the text part from the work order to be assigned to obtain the work order text part, wherein the work order text part includes basic customer information, customer request, problem description and contact information. By extracting these key information from the work order, the system can accurately grasp the customer's needs and problems, and provide basic data for subsequent processing and work order assignment.

[0072] Next, considering that the work order text contains a large amount of semantic information about the customer's basic information, needs, and problems, these semantic information plays an important role in the subsequent work order dispatch. Therefore, in order to better understand the content of the work order text, in the technical solution of the present application, the work order text is further processed by word segmentation and then passed through a semantic encoder containing a word embedding layer to obtain a sequence of word granularity semantic encoding feature vectors of the work order text. The work order text can be segmented according to the unit of words through word segmentation, which helps to accurately capture the meaning and context of each word in the work order, and at the same time extract the semantic contextual association relationship between each word in the work order, so as to facilitate a more comprehensive and detailed understanding of the semantics of the work order type, urgency, and customer needs, and provide data basis for the subsequent work order dispatch.

[0073] Correspondingly, the work order text semantic understanding module 830 is used to: perform word segmentation processing on the work order text part and then pass it through a semantic encoder including a word embedding layer to obtain a sequence of word granularity semantic encoding feature vectors of the work order text.

[0074] Then, since there are many maintenance technicians in the after-sales maintenance technician database, the skills and characteristics of these maintenance technicians are different. In order to capture the semantic feature information of the capabilities, experience and professional fields of these maintenance technicians, so as to provide a more effective basis for the work order assignment task, in the technical solution of this application, it is necessary to extract the information description of the candidate maintenance technician from the after-sales maintenance technician database. Then, the information description of the candidate maintenance technician is processed by word segmentation and then passed through the semantic encoder of the word embedding layer to obtain a sequence of word granularity semantic encoding feature vectors of the information description of the candidate maintenance technician. It should be understood that word segmentation and word embedding technology help to accurately capture the meaning and context of each word in the information description of the candidate maintenance technician, and further context semantic encoding of the word embedding information of each word can be performed, so that the word embedding semantic context information and association of each word in the information description of the candidate maintenance technician can be extracted, which helps the system to better understand the capabilities, experience and professional fields of the technicians, so as to intelligently match and assign the technicians who are most suitable for handling the work order for the work order, and improve service efficiency and quality.

[0075] Accordingly, the candidate maintenance technician information semantic understanding module 850 is used to: perform word segmentation processing on the candidate maintenance technician information description and then pass it through the semantic encoder of the word embedding layer to obtain a sequence of word granularity semantic encoding feature vectors of the candidate maintenance technician information description.

[0076] Further, considering that the word-granularity contextual semantic features in the work order text part and the information description of the candidate maintenance technician may have a mutual local semantic association relationship, such as the word-granularity semantics about customer problems and needs in the work order text part and the word-granularity semantics about the skills and experience of the candidate maintenance technician in the information description of the candidate maintenance technician. Therefore, in order to be able to more finely analyze and compare the local semantics in the work order text part and the information description of the candidate maintenance technician, so as to determine which technicians are most suitable for handling a specific work order, in the technical solution of the present application, the local matching coefficient between the sequence of the word-granularity semantic encoding feature vectors of the work order text and the sequence of the word-granularity semantic encoding feature vectors of the information description of the candidate maintenance technician is calculated. By calculating the local matching coefficient, it can help evaluate the local semantic similarity between the work order text and the information description of the candidate maintenance technician, that is, the degree of matching between them at the level of specific words or phrases. This helps to more accurately match the association between the work order requirements and the capabilities of the technicians, thereby improving the accuracy and efficiency of the technician allocation.

[0077] The local matching coefficient calculation module 860 is used to calculate the local matching coefficient between the sequence of the word granularity semantic encoding feature vectors of the work order text and the sequence of the word granularity semantic encoding feature vectors of the candidate maintenance technician information description using the following local matching formula; wherein the local matching formula is:

[0078]

[0079]

[0080] Among them, v' i is the i-th work order text word granularity semantic encoding feature vector in the sequence of the work order text word granularity semantic encoding feature vectors, u' j is the jth candidate maintenance technician information description word granular semantic coding feature vector in the sequence of candidate maintenance technician information description word granular semantic coding feature vectors, ||·|| represents the Frobenius norm of the vector, S ij represents the eigenvalues ​​of each position in the semantic cosine similarity matrix of work order-candidate maintenance technicians, [S ij ] + is max(S ij ,0), k is the number of feature vectors in the sequence of the word-granular semantic encoding feature vectors of the work order text, is the eigenvalue of each position in the work order-candidate maintenance technician semantic attention matrix, n is the number of feature vectors in the sequence of word-granular semantic encoding feature vectors of the candidate maintenance technician information description, p is the number of vectors in the sequence of word-granular semantic encoding feature vectors of the work order text, is the eigenvalue of each position in the semantic attention matrix of candidate maintenance technicians and work orders, α ij is the attention weight value of the word-granular semantic encoding feature containing the information description word-granular semantic encoding feature of the candidate maintenance technician corresponding to the i-th work order text word-granular semantic encoding feature vector, represents the granular semantic encoding feature vector of the information description word of the candidate maintenance technician in the sequence of the granular semantic encoding feature vector of the information description word of the candidate maintenance technician, λ is the inverse temperature coefficient of the Softmax function, β ij represents the attention weight value of the word-granular semantic coding feature of the work order text corresponding to the word-granular semantic coding feature vector of the j-th candidate maintenance technician information description, represents the jth work order text word granularity semantic encoding feature vector in the work order text word granularity semantic encoding feature vector sequence, represents the eigenvalues ​​of each position in the semantic cosine similarity matrix of the enhanced work order and candidate maintenance technicians, represents the eigenvalues ​​of each position in the semantic cosine similarity matrix of enhanced candidate maintenance technicians and enhanced work orders, μ1 is a hyperparameter, S t (I, T) represents the local matching coefficient, and exp(·) represents an exponential function with the natural constant e as the base.

[0081] Furthermore, after considering the local semantic similarity and matching degree at the specific word or phrase level in the work order text part and the information description of the candidate maintenance technician, in order to screen out maintenance technicians who are more matched with specific work order categories and customer needs, it is necessary to further evaluate the overall similarity or matching degree between the entire work order text and the information description of the candidate technician. Based on this, in the technical solution of the present application, the global matching coefficient between the sequence of the word granularity semantic encoding feature vectors of the work order text and the sequence of the word granularity semantic encoding feature vectors of the alternative maintenance technician information description is further calculated. By calculating the global matching coefficient, it is possible to comprehensively consider the matching of all word or phrase levels between the work order text and the technician information description, that is, the global matching coefficient can provide an overall similarity measure between the work order text and the candidate technician information description, helping the system to comprehensively evaluate the matching degree between the work order requirements and the technician's capabilities. This helps to improve the accuracy and efficiency of technician allocation, thereby improving service quality and customer satisfaction.

[0082] Correspondingly, the global matching coefficient calculation module 870 includes: a first vector cascade unit 871, which is used to cascade the sequence of the work order text word granularity semantic coding feature vectors and the sequence of the candidate maintenance technician information description word granularity semantic coding feature vectors to obtain the work order text word granularity semantic coding cascade feature vector and the candidate maintenance technician information description word granularity semantic coding cascade feature vector; a second vector cascade unit 872, which is used to cascade the work order text word granularity semantic coding cascade feature vector and the candidate maintenance technician information description word granularity semantic coding cascade feature vector to obtain the work order text-alternative maintenance technician information associated feature vector; a linear transformation unit 873, which is used to use the transformation matrix and the bias vector to linearly transform the work order text-alternative maintenance technician information associated feature vector to obtain a linearly transformed work order text-alternative maintenance technician information associated feature vector; an activation unit 874, which is used to use The sigmoid activation function activates the linear transformation work order text-alternative maintenance technician information associated feature vector to obtain an activated linear transformation threshold vector; a weighting unit 875 is used to calculate the weighted sum of the work order text word granularity semantic encoding cascade feature vector and the alternative maintenance technician information description word granularity semantic encoding cascade feature vector using each position feature value in the activated linear transformation threshold vector as a threshold value to obtain a work order text-alternative maintenance technician information word granularity semantic matching feature vector; and a global matching coefficient calculation unit 876 is used to calculate the transpose of the work order text-alternative maintenance technician information word granularity semantic matching feature vector and the vector multiplication between the work order text-alternative maintenance technician information word granularity semantic matching feature vector, and then divide the obtained autocorrelation value by the square of the norm of the work order text-alternative maintenance technician information word granularity semantic matching feature vector to obtain the global matching coefficient.

[0083] Among them, the linear transformation unit 873 is used to: calculate the matrix multiplication of the work order text-candidate maintenance technician information associated feature vector and the transformation matrix, and then add the obtained feature vector to the bias vector by position to obtain the linear transformation work order text-candidate maintenance technician information associated feature vector.

[0084] Among them, the weighting unit 875 is used to: calculate the exponential function of the work order text word granularity semantic encoding cascade feature vector with a natural constant as the base to obtain the work order text word granularity semantic class support feature vector; calculate the exponential function of the candidate maintenance technician information description word granularity semantic encoding cascade feature vector with a natural constant as the base to obtain the candidate maintenance technician information description word granularity semantic class support feature vector; calculate the product of the feature value of each position in the work order text word granularity semantic class support feature vector and the threshold value to obtain the weighted item work order text word granularity semantic class support feature vector; calculate the product of the feature value of each position in the candidate maintenance technician information description word granularity semantic class support feature vector and one minus the threshold value to obtain the weighted item candidate maintenance technician information description word granularity semantic class support feature vector; and, add the weighted item work order text word granularity semantic class support feature vector and the weighted item candidate maintenance technician information description word granularity semantic class support feature vector according to the position point to obtain the work order text-alternative maintenance technician information word granularity semantic matching feature vector.

[0085] In a specific example, the global matching coefficient calculation module 870 is used to calculate the global matching coefficient between the sequence of the word granularity semantic encoding feature vectors of the work order text and the sequence of the word granularity semantic encoding feature vectors of the candidate maintenance technician information description using the following global matching formula; wherein the global matching formula is:

[0086]

[0087] Among them, v' p is the pth work order text word granularity semantic encoding feature vector in the sequence of the work order text word granularity semantic encoding feature vectors, u' n is the jth work order text word granularity semantic encoding feature vector in the sequence of the work order text word granularity semantic encoding feature vectors, n is the number of feature vectors in the sequence of the candidate maintenance technician information description word granularity semantic encoding feature vectors, p is the number of vectors in the sequence of the work order text word granularity semantic encoding feature vectors, is the concatenated feature vector of the word-granular semantic encoding of the work order text, Represents the cascade feature vector of word-granular semantic encoding of the work order text, v g is the semantic fusion feature vector, W g is the transformation matrix, b is the bias vector, sigmoid(·) represents the activation function, concat(·,·) represents the cascade operation, T is the activation linear transformation threshold vector, t is the threshold value of each position in the activation linear transformation threshold vector, t∈[0,1], ||·|| represents the one-norm of the vector, S g is the global matching coefficient.

[0088] It should be understood that the local matching coefficient and the global matching coefficient respectively represent the local semantic matching degree and the global semantic matching degree in the work order text part and the information description of the candidate maintenance technician. In order to utilize these two different scales of semantic matching information to more accurately perform work order assignment, in the technical solution of the present application, the weighted sum of the local matching coefficient and the global matching coefficient is further calculated to obtain a comprehensive matching coefficient. By weighting and combining the local matching coefficient and the global matching coefficient, the local matching and overall matching between the work order text and the technician's information description can be comprehensively considered to obtain a more comprehensive and accurate matching metric. This can help the system better weigh the importance of local matching and global matching, thereby improving the matching quality and accuracy, providing decision makers with more comprehensive information, and helping them make decisions on whether to assign candidate technicians to handle work orders.

[0089] Next, based on the comparison between the comprehensive matching coefficient and the predetermined threshold, it is determined whether to assign the candidate maintenance technician to handle the work order to be assigned. That is to say, based on the comparison between the comprehensive matching coefficient and the predetermined threshold, it can help determine whether to assign the candidate maintenance technician to handle the work order to be assigned. In particular, in response to the comprehensive matching coefficient being greater than the predetermined threshold, it is determined to assign the candidate maintenance technician to handle the work order to be assigned. In this way, the work order can be automatically assigned to the most suitable after-sales maintenance technician, thereby realizing the intelligent assignment and processing of the work order, so that the management of after-sales data meets the actual needs of customers, thereby improving the quality and efficiency of after-sales service.

[0090] Accordingly, the after-sales work order dispatching module 880 includes: a comprehensive matching coefficient calculation unit 881, which is used to calculate the weighted sum of the local matching coefficient and the global matching coefficient to obtain a comprehensive matching coefficient; and a work order dispatching unit 882, which is used to determine whether to assign the candidate maintenance technician to handle the work order to be assigned based on the comparison between the comprehensive matching coefficient and a predetermined threshold. Accordingly, in response to the comprehensive matching coefficient being greater than the predetermined threshold, it is determined to assign the candidate maintenance technician to handle the work order to be assigned.

[0091] Further, in a preferred example, calculating the global matching coefficient between the sequence of the work order text word granularity semantic encoding feature vectors and the sequence of the candidate maintenance technician information description word granularity semantic encoding feature vectors includes the following steps: cascading the sequence of the work order text word granularity semantic encoding feature vectors and the sequence of the candidate maintenance technician information description word granularity semantic encoding feature vectors to obtain the work order text word granularity semantic encoding cascade feature vectors and the candidate maintenance technician information description word granularity semantic encoding cascade feature vectors respectively; and calculating the cosine similarity between the work order text word granularity semantic encoding cascade feature vectors and the candidate maintenance technician information description word granularity semantic encoding cascade feature vectors as the global matching coefficient.

[0092] Further, in this preferred example, the cosine similarity between the work order text word granularity semantic encoding cascade feature vector and the candidate maintenance technician information description word granularity semantic encoding cascade feature vector is calculated as the global matching coefficient, including: cascading the work order text word granularity semantic encoding cascade feature vector and the candidate maintenance technician information description word granularity semantic encoding cascade feature vector to obtain a work order text-candidate maintenance technician semantic cascade representation feature vector; calculating the sum of the work order text-candidate maintenance technician semantic cascade representation feature vector and the square root of the length of the work order text-candidate maintenance technician semantic cascade representation feature vector and the reciprocal of the square root of the bi-norm of the work order text-candidate maintenance technician semantic cascade representation feature vector to obtain a work order text-candidate maintenance technician semantic cascade offset feature vector; calculating the exponential function of the work order text-candidate maintenance technician semantic cascade offset feature vector with a natural constant as the base to obtain the work order text-candidate maintenance technician semantic cascade representation feature vector. Cascade offset prediction feature vector; calculating the dot product of the work order text-candidate maintenance technician semantic cascade representation feature vector and the norm and weight hyperparameter of the work order text-candidate maintenance technician semantic cascade representation feature vector to obtain the work order text-candidate maintenance technician semantic cascade constraint feature vector; calculating the dot sum of the work order text-candidate maintenance technician semantic cascade offset prediction feature vector and the work order text-candidate maintenance technician semantic cascade constraint feature vector to obtain the optimized work order text-candidate maintenance technician semantic cascade representation feature vector; splitting the optimized work order text-candidate maintenance technician semantic cascade representation feature vector into an optimized work order text word granularity semantic encoding cascade feature vector and an optimized candidate maintenance technician information description word granularity semantic encoding cascade feature vector; calculating the cosine similarity between the optimized work order text word granularity semantic encoding cascade feature vector and the optimized candidate maintenance technician information description word granularity semantic encoding cascade feature vector as the global matching coefficient.

[0093] Specifically, due to the content differences between the work order text and the candidate maintenance technician information in the text source domain, the word-granularity semantic encoding cascade feature representation of the work order text and the word-granularity semantic cascade feature representation of the candidate maintenance technician information description have a more significant feature distribution misalignment in the high-dimensional feature space, resulting in poor computational regression constraints of the cosine similarity between the two in the cosine distance space. Therefore, in the above preferred example, the work order text word granularity semantic encoding cascade feature vector and the candidate maintenance technician information description word granularity semantic encoding cascade feature vector are first feature aggregated in a cascade manner, and then the structured norm representation of the work order text-candidate maintenance technician semantic cascade representation feature vector is used as the local canonical coordinates for each eigenvalue of the work order text-candidate maintenance technician semantic cascade representation feature vector to determine the class rotation offset of the vector overall distribution of the work order text-candidate maintenance technician semantic cascade representation feature vector relative to the eigenvalue for the class offset prediction direction of each eigenvalue of the work order text-candidate maintenance technician semantic cascade representation feature vector as the center, and the bounding box of the vector eigenvalue distribution of the work order text-candidate maintenance technician semantic cascade representation feature vector is used to constrain the eigenvalue. Through the above method, the computational regression constraint of the cosine similarity of the work order text word granularity semantic encoding cascade feature vector and the candidate maintenance technician information description word granularity semantic encoding cascade feature vector in the cosine distance space is improved, thereby improving the calculation accuracy of the global matching coefficient. In this way, work orders can be more accurately assigned to the most appropriate after-sales maintenance technicians, thereby realizing intelligent dispatch and processing of work orders.

[0094] Based on the above embodiments, see Figure 5 FIG. 1 is a flowchart of an after-sales data management method in an embodiment of the present disclosure. For example, the after-sales data management method may be executed by a server, which may be Figure 1 The server 100 shown in FIG. Figure 5As shown, according to the after-sales data management method of the embodiment of the present disclosure, the method includes the following steps: S510, extracting the to-be-assigned work order from the after-sales work order database; S520, extracting the text part from the to-be-assigned work order to obtain the work order text part, wherein the work order text part includes basic customer information, customer request, problem description and contact information; S530, performing word granularity-based semantic encoding on the work order text part to obtain a sequence of word granularity semantic encoding feature vectors of the work order text; S540, extracting the information description of the candidate maintenance technician from the after-sales maintenance technician database; S550, performing word granularity-based semantic encoding on the information description of the candidate maintenance technician to obtain the candidate maintenance technician. A sequence of word-granular semantic coding feature vectors for technician information description; S560, calculating the local matching coefficients between the sequence of word-granular semantic coding feature vectors for the work order text and the sequence of word-granular semantic coding feature vectors for the candidate maintenance technician information description; S570, calculating the global matching coefficients between the sequence of word-granular semantic coding feature vectors for the work order text and the sequence of word-granular semantic coding feature vectors for the candidate maintenance technician information description; and, S580, calculating the weighted sum of the local matching coefficients and the global matching coefficients to obtain a comprehensive matching coefficient, and determining whether to refer to the candidate maintenance technician to handle the work order to be assigned based on the comprehensive matching coefficient.

[0095] In the above-mentioned after-sales data management method, the work order text part is semantically encoded based on word granularity to obtain a sequence of work order text word granularity semantic encoding feature vectors, including: after word segmentation processing of the work order text part, passing it through a semantic encoder including a word embedding layer to obtain a sequence of work order text word granularity semantic encoding feature vectors.

[0096] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned after-sales data management method have been referred to above. Figures 2 to 4 It has been described in detail in the description of the after-sales data management system 800, and therefore, its repeated description will be omitted.

[0097] Figure 6 FIG. 1 is an application scenario diagram of the after-sales data management system according to an embodiment of the present disclosure. Figure 6 As shown, in this application scenario, first, the work orders to be assigned are extracted from the after-sales work order database (for example, Figure 6 D1 shown in FIG. 1 ), and extracting information descriptions of candidate maintenance technicians from an after-sales maintenance technician database (e.g., Figure 6 Then, the information description of the work order to be assigned and the candidate maintenance technician is input into a server (for example, Figure 6In S) shown in , the server is able to use the after-sales data management algorithm to process the information description of the work order to be dispatched and the alternative maintenance technician to determine whether to refer the alternative maintenance technician to handle the work order to be dispatched.

[0098] Based on the above embodiments, another exemplary embodiment of an electronic device is also provided in the embodiments of the present disclosure. In some possible implementations, the electronic device in the embodiments of the present disclosure may include a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor may implement the steps of the after-sales data management method in the above embodiments when executing the program.

[0099] For example, the electronic device is disclosed in the present invention Figure 1 Taking the server 100 in the example as an example, the processor in the electronic device is the processor 110 in the server 100, and the memory in the electronic device is the memory 120 in the server 100.

[0100] An embodiment of the present disclosure also provides a computer-readable storage medium. Figure 7 FIG. 1 is a schematic diagram of a computer-readable storage medium 1000 according to an embodiment of the present disclosure. Figure 7 As shown, the computer-readable storage medium 1000 stores computer-executable instructions 1001. When the computer-executable instructions 1001 are executed by the processor, the after-sales data management method according to the embodiment of the present disclosure described with reference to the above figures can be executed. The computer-readable storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. The volatile memory may include, for example, a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may include, for example, a read-only memory (ROM), a hard disk, a flash memory, etc.

[0101] The embodiment of the present disclosure also provides a computer program product or a computer program, which includes a computer instruction, and the computer instruction is stored in a computer-readable storage medium. The processor of the computer device reads the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device executes the after-sales data management method according to the embodiment of the present disclosure.

[0102] Those skilled in the art will appreciate that the contents disclosed in this disclosure may be subject to various modifications and improvements. For example, the various devices or components described above may be implemented by hardware, or by software, firmware, or a combination of some or all of the three.

[0103] In addition, although the present disclosure makes various references to certain units in the system according to embodiments of the present disclosure, any number of different units may be used and run on the client and / or server. The units are illustrative only, and different aspects of the systems and methods may use different units.

[0104] Those skilled in the art will appreciate that all or part of the steps in the above method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk. Optionally, all or part of the steps in the above embodiment can also be implemented using one or more integrated circuits. Accordingly, each module / unit in the above embodiment can be implemented in the form of hardware or in the form of a software functional module. The present disclosure is not limited to any particular form of combination of hardware and software.

[0105] The above is an explanation of the present disclosure and should not be considered as a limitation thereof. Although several exemplary embodiments of the present disclosure have been described, it will be readily understood by those skilled in the art that many modifications may be made to the exemplary embodiments without departing from the novel teachings and advantages of the present disclosure. Therefore, all such modifications are intended to be included within the scope of the present disclosure as defined in the claims. It should be understood that the above is an explanation of the present disclosure and should not be considered as being limited to the specific embodiments disclosed, and modifications to the disclosed embodiments and other embodiments are intended to be included within the scope of the appended claims. The present disclosure is defined by the claims and their equivalents.

Claims

1. An after-sales data management system, characterized in that: include: The work order data extraction module to be assigned is used to extract the work orders to be assigned from the after-sales work order database; A work order text part extraction module, used to extract the text part from the work order to be assigned to obtain the work order text part, wherein the work order text part includes basic customer information, customer request, problem description and contact information; A work order text semantic understanding module, used for performing word granularity-based semantic encoding on the work order text portion to obtain a sequence of work order text word granularity semantic encoding feature vectors; A candidate maintenance technician information extraction module is used to extract information descriptions of candidate maintenance technicians from an after-sales maintenance technician database; A candidate maintenance technician information semantic understanding module is used to perform word granularity-based semantic coding on the candidate maintenance technician information description to obtain a sequence of word granularity semantic coding feature vectors of the candidate maintenance technician information description; A local matching coefficient calculation module, used to calculate the local matching coefficient between the sequence of word granularity semantic coding feature vectors of the work order text and the sequence of word granularity semantic coding feature vectors of the candidate maintenance technician information description; A global matching coefficient calculation module, used to calculate the global matching coefficient between the sequence of word-granular semantic coding feature vectors of the work order text and the sequence of word-granular semantic coding feature vectors of the candidate maintenance technician information description; and an after-sales work order dispatching module, configured to calculate a weighted sum of the local matching coefficient and the global matching coefficient to obtain a comprehensive matching coefficient, and determine whether to refer the candidate maintenance technician to handle the work order to be dispatched based on the comprehensive matching coefficient; Wherein, the global matching coefficient calculation module includes: A first vector cascading unit is used to cascade the sequence of the work order text word granularity semantic coding feature vectors and the sequence of the candidate maintenance technician information description word granularity semantic coding feature vectors to obtain a work order text word granularity semantic coding cascade feature vector and a candidate maintenance technician information description word granularity semantic coding cascade feature vector; A second vector cascading unit is used to cascade the work order text word granularity semantic encoding cascade feature vector and the candidate maintenance technician information description word granularity semantic encoding cascade feature vector to obtain a work order text-candidate maintenance technician information association feature vector; A linear transformation unit, used to use a transformation matrix and a bias vector to linearly transform the work order text-candidate maintenance technician information association feature vector to obtain a linearly transformed work order text-candidate maintenance technician information association feature vector; Activate the unit for use The activation function activates the linear transformation work order text-candidate maintenance technician information associated feature vector to obtain an activation linear transformation threshold vector; A weighting unit, configured to calculate a weighted sum of the work order text word granularity semantic encoding cascade feature vector and the candidate maintenance technician information description word granularity semantic encoding cascade feature vector using each position feature value in the activation linear transformation threshold vector as a threshold value to obtain a work order text-candidate maintenance technician information word granularity semantic matching feature vector; and A global matching coefficient calculation unit is used to calculate the transpose of the work order text-candidate maintenance technician information word granularity semantic matching feature vector and the vector multiplication between the work order text-candidate maintenance technician information word granularity semantic matching feature vector, and then divide the obtained autocorrelation value by the square of the norm of the work order text-candidate maintenance technician information word granularity semantic matching feature vector to obtain the global matching coefficient.

2. The after-sales data management system according to claim 1, characterized in that: The work order text semantic understanding module is used to: After word segmentation processing is performed on the work order text portion, a semantic encoder including a word embedding layer is passed through to obtain a sequence of word granularity semantic encoding feature vectors of the work order text.

3. The after-sales data management system according to claim 2, characterized in that: The candidate maintenance technician information semantic understanding module is used to: The information description of the candidate maintenance technician is processed by word segmentation and then passed through the semantic encoder of the word embedding layer to obtain a sequence of word granularity semantic encoding feature vectors of the information description of the candidate maintenance technician.

4. The after-sales data management system according to claim 3, characterized in that: The linear transformation unit is used for: After calculating the matrix multiplication of the work order text-candidate maintenance technician information associated feature vector and the transformation matrix, the obtained feature vector is added to the bias vector by position to obtain the linear transformation work order text-candidate maintenance technician information associated feature vector.

5. The after-sales data management system according to claim 4, characterized in that: The weighting unit is used for: Calculating an exponential function of the work order text word granularity semantic encoding cascade feature vector with a natural constant as the base to obtain a work order text word granularity semantic class support feature vector; Calculating an exponential function of the granular semantic encoding cascade feature vector of the candidate maintenance technician information description word with a natural constant as the base to obtain a granular semantic class support feature vector of the candidate maintenance technician information description word; Calculate the product of the feature value of each position in the work order text word granularity semantic class support feature vector and the threshold value to obtain the weighted item work order text word granularity semantic class support feature vector; Calculate the product of the feature value of each position in the candidate maintenance technician information description word granular semantic class support feature vector and one minus a threshold value to obtain a weighted candidate maintenance technician information description word granular semantic class support feature vector; and The weighted item work order text word granularity semantic class support feature vector and the weighted item candidate maintenance technician information description word granularity semantic class support feature vector are added according to the position points to obtain the work order text-candidate maintenance technician information word granularity semantic matching feature vector.

6. The after-sales data management system according to claim 5, characterized in that: The after-sales work order dispatching module includes: a comprehensive matching coefficient calculation unit, configured to calculate a weighted sum of the local matching coefficient and the global matching coefficient to obtain a comprehensive matching coefficient; and The work order dispatching unit is used to determine whether to designate the candidate maintenance technician to handle the to-be-dispatched work order based on a comparison between the comprehensive matching coefficient and a predetermined threshold.

7. The after-sales data management system according to claim 6, characterized in that: In response to the comprehensive matching coefficient being greater than the predetermined threshold, it is determined to assign the candidate maintenance technician to handle the work order to be assigned.

8. An after-sales data management method, applied to the after-sales data management system according to any one of claims 1 to 7, characterized in that: include: Extract the work orders to be dispatched from the after-sales work order database; Extracting a text portion from the work order to be assigned to obtain a work order text portion, wherein the work order text portion includes basic customer information, customer request, problem description, and contact information; Performing word-granularity-based semantic coding on the work order text portion to obtain a sequence of work order text word-granularity semantic coding feature vectors; Extracting information descriptions of candidate maintenance technicians from an after-sales maintenance technician database; Performing word-granularity-based semantic coding on the information description of the candidate maintenance technician to obtain a sequence of word-granularity semantic coding feature vectors of the information description of the candidate maintenance technician; Calculating a local matching coefficient between a sequence of word-granular semantic encoding feature vectors of the work order text and a sequence of word-granular semantic encoding feature vectors of the candidate maintenance technician information description; Calculating a global matching coefficient between a sequence of word-granular semantic encoding feature vectors of the work order text and a sequence of word-granular semantic encoding feature vectors of the candidate maintenance technician information description; and A weighted sum of the local matching coefficient and the global matching coefficient is calculated to obtain a comprehensive matching coefficient, and based on the comprehensive matching coefficient, it is determined whether to designate the candidate maintenance technician to handle the work order to be assigned.

9. The after-sales data management method according to claim 8, characterized in that: The work order text portion is semantically encoded based on word granularity to obtain a sequence of work order text word granularity semantic encoding feature vectors, including: After word segmentation processing is performed on the work order text portion, a semantic encoder including a word embedding layer is passed through to obtain a sequence of word granularity semantic encoding feature vectors of the work order text.

Images