Service work order classification method, electronic equipment, storage medium and program product

Abstract

The invention provides a service work order classification method, electronic equipment, a storage medium and a program product, and belongs to the field of artificial intelligence. The method comprises the steps of performing clustering operation on at least one sample service work order to obtain at least one service work order set; based on the first large language model and the first cue word information, performing scene recognition processing on sample service work orders included in the service work order set to obtain at least one work order problem scene and a sample service work order corresponding to each work order problem scene in the service work order set, each work order problem scene comprises a problem phenomenon field, a key error reporting field, a troubleshooting scheme field, a processing scheme field and field values of the fields. According to the method and the device, the work order problem scene to which the service work order belongs can be automatically determined, and the work order problem scene comprises a plurality of fields and field values, so that the classification result is more accurate, and the usability is higher.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a service work order classification method, electronic equipment, storage medium and program product. Background Art

[0002] A service ticket is a record created by an enterprise through a customer service management system or other similar tools to track and manage customer issues or needs. As a service request management tool, service tickets play an important role in the operation of an enterprise. They can not only effectively manage customer issues and needs and improve user satisfaction, but also improve user work efficiency and help enterprises improve products and services. In order to better handle service tickets, service tickets need to be classified.

[0003] Related technologies use manual methods to classify service tickets. The classification results are heavily dependent on user experience and contain less information, resulting in inaccurate classification results and lack of usability. Summary of the invention

[0004] The embodiment of the present application provides a service ticket classification method, electronic device, storage medium and program product, which can automatically determine the service ticket problem scenario to which the service ticket belongs, and the service ticket problem scenario includes multiple fields and field values, and the classification result is more accurate and has higher usability. The technical solution is as follows: In a first aspect, a service ticket classification method is provided, the method comprising: Performing a clustering operation on at least one sample service work order to obtain at least one service work order set, wherein the sample service work orders included in any service work order set in the at least one service work order set have the same work order category; For any service ticket set, based on the first language model and the first prompt word information, scenario recognition processing is performed on the sample service tickets included in the any service ticket set to obtain at least one ticket problem scenario and the sample service tickets corresponding to each ticket problem scenario in the any service ticket set. Each ticket problem scenario includes a problem scenario description field and its field value. The problem scenario description field includes a problem phenomenon field, a key error field, a troubleshooting solution field and a processing solution field.

[0005] In a second aspect, a service work order classification device is provided, the device comprising: A clustering module, configured to perform a clustering operation on at least one sample service work order to obtain at least one service work order set, wherein the sample service work orders included in each service work order set in the at least one service work order set have the same work order category; An identification module is used to perform scenario recognition processing on the sample service work orders included in any service work order set based on the first language model and the first prompt word information, to obtain at least one work order problem scenario and the sample service work orders corresponding to each work order problem scenario in any service work order set, each work order problem scenario includes a problem scenario description field and its field value, and the problem scenario description field includes a problem phenomenon field, a key error field, a troubleshooting solution field and a processing solution field.

[0006] In a third aspect, an electronic device is provided, comprising a processor and a memory; the memory stores at least one program code; the at least one program code is used to be called and executed by the processor to implement the service ticket classification method described in the first aspect.

[0007] In a fourth aspect, a computer-readable storage medium is provided, wherein at least one computer program is stored in the computer-readable storage medium, and when the at least one computer program is executed by a processor, the service ticket classification method described in the first aspect can be implemented.

[0008] In a fifth aspect, a computer program product is provided, the computer program product comprising a computer program, and when the computer program is executed by a processor, the service ticket classification method described in the first aspect can be implemented.

[0009] The beneficial effects of the technical solution provided by the embodiment of the present application are: The embodiment of the present application can obtain at least one service work order set by clustering at least one sample service work order, and the sample service work orders included in each service work order set in the at least one service work order set have the same work order category. Then, the first large language model plus the first prompt information are used to perform scene recognition on the sample service work orders included in each service work order set, and at least one work order problem scene and the sample service work order corresponding to each work order problem scene in any service work order set are obtained. The work order problem scene includes a problem scene description field and its field value, and the problem scene description field includes a problem phenomenon field, a key error field, a troubleshooting solution field, and a processing solution field, etc., which can fully describe the situation of the service work order. Since there is no need to use manual classification, with the help of a large language model, the work order problem scene to which the sample service work order belongs can be automatically determined, and the determined work order problem scene is summarized by the large language model from the sample service work order of the actual scene, which can accurately reflect the category of the sample service work order, so the classification result is more accurate, and the classification is more universal and more usable. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0011] Figure 1 This is a flow chart of a service work order classification method provided in an embodiment of the present application; Figure 2 is a flow chart of a training method for a vector conversion model provided in an embodiment of the present application; Figure 3 This is an overall flow chart of service work order classification provided by an embodiment of the present application; Figure 4 Schematic diagram of the index structure of the work order knowledge base constructed in the embodiment of the present application; Figure 5 is a flowchart of another service work order classification method provided in an embodiment of the present application; Figure 6 It is a structural diagram of a service work order classification device provided in an embodiment of the present application; Figure 7 A structural block diagram of an electronic device provided by an exemplary embodiment of the present application is shown. DETAILED DESCRIPTION

[0012] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0013] It can be understood that the terms "each", "multiple", and "any" used in the embodiments of the present application include two or more, each refers to each of the corresponding multiple, and any refers to any one of the corresponding multiple. For example, the multiple words include 10 words, and each word refers to each of the 10 words, and any word refers to any one of the 10 words.

[0014] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0015] Artificial intelligence is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain better results. In other words, artificial intelligence is a comprehensive technology in computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines, so that machines have the functions of perception, reasoning and decision-making.

[0016] Artificial intelligence technology is a comprehensive discipline that covers a wide range of fields, including both hardware-level and software-level technologies. The basic technologies of artificial intelligence generally include sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics and other technologies. Artificial intelligence software technology mainly includes computer vision technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0017] With the research and progress of artificial intelligence technology, artificial intelligence technology has been studied and applied in many fields, such as common smart homes, smart wearable devices, virtual assistants, smart speakers, smart marketing, unmanned driving, autonomous driving, robots, smart medical care, smart customer service, etc. It is believed that with the development of technology, artificial intelligence technology will be applied in more fields and play an increasingly important role. The solution provided in the embodiment of this application involves natural language processing, machine learning / deep learning and other technologies of artificial intelligence.

[0018] As mentioned above, service tickets play an important role in the operation of an enterprise. Therefore, how to handle service tickets is crucial. Different service tickets have different problems and different handling methods. In order to better handle service tickets, service tickets need to be classified before they are processed. Based on the classification results, problems in service tickets can be handled in a timely manner to improve user satisfaction and enhance product service experience.

[0019] In order to improve the processing efficiency of service tickets, currently, relevant technologies will classify historical service tickets based on user experience or traditional categories, and then build a ticket knowledge base based on the classification results of historical service tickets. User experience depends on the user's understanding of product problems, and traditional categories depend on the understanding of previous products. Whether it is a classification method based on user experience or a classification method based on traditional categories, the determined classification results may be out of touch with the actual application scenario of the product, resulting in inaccurate classification results and lack of universality and availability. Furthermore, since the ticket knowledge base is constructed based on historical service tickets and their classification results, if the classification results are inaccurate, the constructed ticket knowledge base will also not be accurate. And because the classification of the ticket knowledge base is not universal and available, a large number of processing solutions in the ticket knowledge base cannot be applied and become invalid solutions.

[0020] In addition, when expanding the capacity of the work order knowledge base, the relevant technology adopts a stacking method. The so-called stacking method means that when optimizing and merging similar solutions, the same problem may have a large number of different solutions. Users will write a question when they find one, and the number of documents in this category will be quickly increased by exhaustive method. However, when searching using the work order knowledge base, one problem often hits multiple solutions. Users cannot tell which solution is truly effective, and may piece together the hit solutions to get the final solution. The pieced-together final solution may not solve the problem. Faced with multiple solutions to the same problem, users may also nest multiple solutions. The so-called nesting means that when there are multiple solutions to the same problem, and there are differences between the multiple solutions, when adding a new solution, the user will reference the link of the same solution, and then add new steps, resulting in multiple nesting of some documents. In actual problem handling, the problem is often unable to be solved, and a large number of out-of-solution operations are required for adjustment, which increases the probability of fault triggering and prolongs the problem handling time.

[0021] At present, in the service ticket processing scenario, users prefer to understand whether the processing solution is suitable for the current problem, and then perform adaptation verification, which makes it more important to determine the problem of the service ticket than to simply obtain the processing solution. However, the classification results determined by the relevant technical methods are inaccurate, and the classification results are usually given based on the problem phenomenon, so the determined problem phenomenon is inaccurate. Moreover, the current service ticket processing methods all give processing solutions on the premise of clarifying the problem phenomenon. When the determined problem phenomenon is inaccurate, the processing solution based on the problem phenomenon is also inaccurate, thereby increasing the risk of secondary failures.

[0022] In addition, in the real service ticket processing scenario, determining the root cause of the service ticket problem helps to associate the problem phenomenon, troubleshooting solution and treatment solution according to the anchor point, and then based on the association result, the problems existing in the service ticket are regularly checked and solved to eliminate manual judgment factors and improve the processing efficiency of the service ticket. However, the classification results determined by the relevant technology only give the problem phenomenon and treatment solution, and do not give information on other dimensions, and do not associate the information on these dimensions, resulting in a small amount of information in the determined classification results, and the processing efficiency of the service ticket based on the ticket knowledge base is low and the accuracy is not high.

[0023] In order to solve the problems of low accuracy and lack of availability when classifying service tickets in related technologies, and to meet the needs of users in service ticket processing scenarios, the embodiment of this application proposes the concept of 4K, which specifically includes: K1: problem phenomenon (description of on-site problem manifestation and root cause), K2: key error (referring to the core log, error message or API call failure generated by the root cause of the problem, etc.), K3: troubleshooting plan (a set of methods for locating the root cause of the problem), K4: processing plan (targeted implementation method for solving the above problems). Then, at the 4K level, with the help of the powerful processing power of the large language model, scenario-based 4K tags are automatically generated for sample service tickets in real scenarios, and the scenario-based 4K tags are associated to avoid interference from human factors, improve the accuracy and availability of the classification results, and increase the probability of the processing plan hitting the problem.

[0024] Based on the proposed 4K concept and with the help of a large language model, the embodiment of the present application identifies work order problem scenarios (scenario-based 4K labels) for sample service work orders by adopting multiple operation steps, including text preprocessing (including data cleaning, formatting, etc.), vectorization training (including training vector conversion model), refined classification (including clustering to obtain a service work order set and generating work order problem labels for the service work order set, etc.), and intelligent scene generation.

[0025] The embodiment of the present application provides a service ticket classification method. Taking the electronic device executing the embodiment of the present application as an example, the electronic device can be a terminal with strong computing power, such as a laptop computer, a desktop computer, etc., or a server. The server can be a separate physical server, or a cluster or distributed system composed of multiple physical servers. Figure 1 , the method flow provided in the embodiment of the present application includes: 101. Determine a target feature vector corresponding to at least one sample service work order.

[0026] Among them, the sample service work order includes historical service work orders and newly added service work orders, etc. The historical service work order refers to a service work order generated before a certain moment and marked with a work order category label. This moment can be the moment of review after the processing of the product-related service work order is completed. The work order category label marked by the historical service work order can be marked manually, and the marked work order category label is an N-level label, where N is greater than or equal to 1. Labels of different levels can reflect the degree of abstraction of the cause of the problem and the handling method of the service work order, which is obtained by gradually refining the cause of the problem and the handling method of the service work order from the product level. In the embodiment of the present application, the N-level labels marked by the historical service work order may include a first-level work order category label and a second-level work order category label, etc. A newly added service work order in the embodiment of the present application refers to a service work order generated before a certain moment but not marked with a work order category label.

[0027] Optionally, some sensitive data and some useless data may exist in the sample service work orders that are usually obtained. In order to avoid the leakage of sensitive information and reduce the resources consumed by the sample service work order processing process, at least one sample service work order may be cleaned. When performing data cleaning on at least one sample service work order, the cleaning content may be determined first, and then the cleaning order may be determined for the determined cleaning content, and then the regular expression written by the user based on the determined cleaning content and cleaning order may be obtained. By executing the regular expression, each sample service work order may be cleaned to obtain at least one cleaned sample service work order. Among them, the cleaning content includes sensitive fields, Htttp links, solution codes, etc. The cleaning order may be determined by the user according to actual cleaning needs, such as cleaning the Htttp links first, then cleaning the solution codes, and finally determining the sensitive fields.

[0028] Furthermore, after at least one cleaned sample service work order is obtained, a preset format may be used to format the at least one cleaned sample service work order to unify the data format in the at least one cleaned sample service work order, thereby obtaining at least one structured sample service work order. The preset format may be a structured data format including specified fields and field values. For example, when the specified fields include fields such as work order title, troubleshooting results, and troubleshooting process, the preset format may be "work order title: troubleshooting results: troubleshooting process:".

[0029] The embodiment of the present application can avoid sensitive data leakage, reduce the computing resources consumed in the service work order processing process, and improve the processing efficiency of the service work order by performing data cleaning, formatting and other processing on at least one sample service work order.

[0030] Among them, the target feature vector can update the distance in the vector space of sample service work orders belonging to the same work order category. Specifically, the target feature vectors corresponding to the sample service work orders of the same category have similarities in the high-dimensional vector space, and the target feature vectors corresponding to the sample service work orders of different categories have differences in the high-dimensional vector space. When determining the target feature vector corresponding to at least one sample service work order, the embodiments of the present application may adopt some algorithms that can determine the target feature vector, and may also use the trained model. Taking the use of the trained model to determine the target feature vector corresponding to at least one sample service work order as an example, the determination process may include the following steps: 1011. Call a vector initialization model to perform feature extraction on at least one sample service work order to obtain an initial feature vector corresponding to the at least one sample service work order.

[0031] The vector initialization model is used to convert each sample service ticket into a high-dimensional vector form. The vector initialization model is a model that can convert text into computer-recognizable characters, such as the BGE model. When each sample service ticket is input into the vector initialization model, it is processed by the vector initialization model to obtain the initial feature vector corresponding to each sample service ticket.

[0032] 1012. Call a vector conversion model to convert an initial feature vector of at least one sample service work order to obtain a target feature vector corresponding to the at least one sample service work order.

[0033] Among them, the vector conversion model is used to convert the initial feature vector of the sample service work order, change the position of the initial feature vector in the high-dimensional space, and obtain the target feature vector corresponding to each sample service work order. The target feature vectors corresponding to the sample service work orders of the same category have similarities in the high-dimensional vector space, and the target feature vectors corresponding to the sample service work orders of different categories have differences in the high-dimensional vector space. Based on the geometric properties of the target feature vector in the high-dimensional vector space, when clustering the sample service work orders, the sample service work orders with the same semantic information can be clustered into one category. The vector conversion model has a neural network structure, which can be a ResNet model, etc. The vector conversion model can be trained using historical service work orders, and the historical service work orders are annotated with N-level work order category labels, such as first-level work order category labels and second-level work order category labels. Figure 2 The training process of the vector conversion model is shown, and the training process includes the following steps: 201. Call the vector initialization model to determine initial feature vectors corresponding to multiple historical service work orders.

[0034] 202. Call the vector conversion model to be trained to convert the initial feature vectors corresponding to the multiple historical service work orders to obtain the target feature vectors corresponding to the multiple historical service work orders.

[0035] 203. According to the N-level work order category labels and target feature vectors corresponding to each of the multiple historical service work orders, determine N loss function values ​​corresponding to the N-level work order category labels.

[0036] For the N-level work order category labels marked by the historical service work orders, N classification tasks can be set, each classification task corresponds to a first-level classification label, and each classification task corresponds to a loss function, and each loss function can be a cross entropy function. The N loss function values ​​respectively reflect that the target feature vectors corresponding to the historical service work orders with the same level of work order category labels have a closer distance. Taking the N-level work order category labels as the first-level work order category labels and the second-level work order category labels as an example, the constructed loss function includes the first loss function and the second loss function. The first loss function corresponds to the first-level work order category label, and the first loss function value reflects that the target feature vectors corresponding to the historical service work orders with the same first-level work order category label have a closer distance; the second loss function corresponds to the second-level work order category label, and the second loss function value reflects that the target feature vectors corresponding to the historical service work orders with the same second-level work order category label have a closer distance.

[0037] 204. Based on the N loss function values, determine a target loss function value to train the vector conversion model to be trained according to the target loss function value.

[0038] By using the weight values ​​corresponding to the N loss function values ​​respectively and performing weighted addition on the N loss function values, the target loss function value can be obtained. If the target loss function value is greater than the preset threshold, the model parameters of the vector conversion model to be trained are adjusted, and the above steps are performed based on the vector conversion model after the parameter adjustment until the training cutoff condition is met, and the training cutoff condition includes that the number of adjustments reaches the set number of times, or the target loss function value is less than the preset threshold, etc. The vector conversion model corresponding to the training cutoff condition is obtained, and the vector conversion model is used as the trained vector conversion model. The historical service work order of the embodiment of the present application has multi-level labels, and the vector conversion model trained for the classification task set for the multi-level labels can make the target feature vector have good classification results on different level labels.

[0039] Based on the trained vector conversion model, for a historical service work order in at least one sample service work order, the initial feature vector corresponding to the historical service work order can be input into the trained vector conversion model, and the work order target feature vector corresponding to the historical service work order can be output; for a newly added service work order in at least one sample service work order, the initial feature vector corresponding to the newly added service work order can be input into the trained vector conversion model, and the work order target feature vector corresponding to the newly added service work order can be output. The embodiment of the present application improves the accuracy of the determined target feature vector by using the trained feature conversion model to process the historical service work orders and the newly added service work orders.

[0040] 102. Perform a clustering operation on at least one sample service work order based on a target feature vector corresponding to each of the at least one sample service work orders to obtain at least one service work order set.

[0041] In order to be able to mine the common attributes of sample service orders of the same category, after obtaining the target feature vector corresponding to at least one sample service order, at least one sample service order will be clustered to obtain at least one service order set, each service order set includes at least one sample service order, and the sample service orders included in the same service order set have the same order category. When clustering at least one sample service order based on the target feature vector corresponding to at least one sample service order, a preset clustering algorithm (such as K-means clustering algorithm, etc.) can be used for clustering, and clustering can also be performed based on a trained clustering model. Of course, other methods can also be used for clustering, which will not be explained here one by one. Taking clustering using a preset clustering algorithm as an example, the clustering process includes the following steps: 1021. Use a preset clustering algorithm to perform a clustering operation on at least one target feature vector corresponding to at least one sample service work order to obtain at least one cluster.

[0042] When clustering at least one sample service ticket using a preset clustering algorithm, a preset number of categories may be pre-set, and then the distance between at least one ticket target feature vector (such as Euclidean distance, cosine similarity, etc.) is calculated to form a cluster of ticket target feature vectors whose distance is less than a distance threshold. After clustering is completed, at least one cluster may be obtained, and each cluster includes at least one target feature vector.

[0043] Optionally, the multiple historical service work orders included in at least one sample service work order are marked with N-level work order category labels. To increase the speed of clustering operations, the historical service work orders with the same N-level work order category labels can be divided into an initial cluster according to the N-level work order category labels marked by the multiple historical service work orders, and at least one initial cluster is obtained. Then, the distance between the target feature vector of the newly added service work order in at least one sample service work order and the target feature vector of each historical service work order in each initial cluster is calculated. If the distance between the target feature vector of any newly added service work order and the target feature vector of each historical service work order in any initial cluster is less than a distance threshold, the newly added service work order is added to the initial cluster. This method is used to process other newly added service work orders in at least one sample service work order, and finally at least one cluster can be obtained.

[0044] 1022. Group the sample service work orders corresponding to the target feature vector in each cluster into an initial service work order set to obtain at least one initial service work order set.

[0045] 1023. In response to an adjustment operation of the user, adjust at least one initial service work order set to obtain at least one service work order set.

[0046] The adjustment process of at least one initial service work order combination includes: determining whether a preset number of categories need to be adjusted; if no adjustment is required, manually checking whether at least one initial service work order set needs to be adjusted; if no manual adjustment is required, treating at least one initial service work order set as at least one service work order; if at least one initial service work order set needs to be manually adjusted, determining whether any initial service work order set in at least one initial service work order set needs to be split; if any service work order set in at least one initial service work order set needs to be split, re-using a preset clustering algorithm to split the service work order set; if any service work order set in at least one initial service work order set does not need to be split, merging the same or similar initial service work order sets in at least one initial service work order set to ultimately obtain at least one service work order set.

[0047] The target feature vector in the embodiment of the present application can make the target feature vectors corresponding to sample service work orders of the same category have similarity in the high-dimensional vector space, and the target feature vectors corresponding to sample service work orders of different categories have differences in the high-dimensional vector space. By performing a clustering operation on at least one sample service work order based on the target feature vector corresponding to each of the at least one sample service work orders, the sample service work orders with the same work order category can be clustered into a service work order set, so that in subsequent operations, scene recognition processing is performed on the service work order set with the same work order category, thereby ensuring the accuracy of the recognition result.

[0048] 103. For any service ticket set, based on the first language model and the first prompt word information, scene recognition processing is performed on the sample service tickets included in any service ticket set to obtain at least one ticket problem scene and the sample service tickets corresponding to each ticket problem scene in any service ticket set.

[0049] Among them, at least one sample service ticket belonging to the service ticket set is mounted under each ticket problem scenario, and each ticket problem scenario includes a problem scenario description field and its field value, and the problem scenario description field includes a problem phenomenon field, a key error field, a troubleshooting solution field, and a processing solution field, etc. For any service ticket set, based on the first language model and the first prompt word information, the sample service tickets included in any service ticket set are subjected to scenario recognition processing, and at least one ticket problem scenario and the sample service ticket corresponding to each ticket problem scenario in any service ticket set can be obtained. During specific processing, a first prompt word information can be generated based on each sample service work order included in any service work order set, and the first prompt word information can be input into the first large language model so that the first large language model generates at least one work order problem scenario corresponding to any service work order set; the sample service work orders included in any service work order set can also be processed in batches to obtain at least one batch of sample service work orders, and for each batch of sample service work orders, the first prompt word information corresponding to at least one batch of sample service work orders is generated, and then the first prompt word information corresponding to at least one batch of sample service work orders is input into the first large language model so that the first large language model generates at least one work order problem scenario corresponding to at least one batch of sample service work orders. Of course, other processing methods can also be used, which will not be described one by one in the embodiments of the present application. Taking the processing of sample service work orders included in any service work order set in batches as an example, the following steps can be included: 1031. Process the sample service work orders included in any service work order set in batches to obtain at least one batch of sample service work orders.

[0050] Due to the limitation of the number of input tokens of the large language model, when the service ticket set includes a large number of sample service tickets, it is necessary to process the sample service tickets included in the service ticket set in batches to obtain at least one batch of sample service tickets. The number of sample service tickets in each batch can be 5, 10, etc., and the embodiments of the present application do not make any specific limitations on this.

[0051] 1032. Generate first prompt word information corresponding to at least one batch of sample service work orders in sequence.

[0052] Fill at least one batch of sample service work orders into the corresponding position of the first prompt template to obtain the first prompt word information corresponding to at least one batch of sample service work orders. Among them, the first prompt template is used to instruct the first language model to extract the field value of the problem phenomenon field and the field value of the key error field from the filled sample service work order. The content of the first prompt template can be: You are a work order processing expert and are very good at summarizing work order scenarios. I will give you the specific information of several work orders, including problem descriptions and solutions, and there may be key errors. Please summarize several combinations of problem phenomena + key error information, and hang the work order id that meets this combination under this combination. Try to hang multiple work orders under the same combination. If there is no key error, please summarize. The answer format is returned in json, for example: {{Problem phenomenon: ***, key error information: ***, work order id list: {x, x,...}, {...}}, note that only json is given, no other information.}.

[0053] 1033. Input first prompt word information corresponding to at least one batch of sample service work orders into the first large language model respectively, so that the first large language model generates at least one work order problem scenario corresponding to at least one batch of sample service work orders.

[0054] The first prompt word information corresponding to the obtained at least one batch of sample service work orders is input into the first large language model, and the first large language model recognizes the input batch of sample service work orders, thereby generating at least one work order problem scenario corresponding to the at least one batch of sample service work orders. Each work order problem scenario includes the field value of the problem phenomenon field and the field value of the key error field corresponding to the corresponding batch of sample service work orders.

[0055] 1034. Based on a sample service ticket belonging to a target ticket problem scenario in any service ticket set, determine a field value of a troubleshooting solution field and a field value of a processing solution field corresponding to the target ticket problem scenario.

[0056] Among them, the target work order problem scenario is any one of at least one work order problem scenario. The embodiment of the present application processes each batch of sample service work orders with the help of the first large language model and the first prompt word information, which not only improves the processing speed of the model, but also generates a corresponding target work order problem scenario for each batch of sample service work orders with the help of the powerful recognition ability of the large language model. The target work order problem scenario includes the field value of the troubleshooting solution field and the field value of the processing solution field, which contains more information than the related technology and has a more accurate classification result.

[0057] Specifically, based on a sample service ticket belonging to a target ticket problem scenario in any service ticket set, the field value of the troubleshooting solution field and the field value of the processing solution field corresponding to the target ticket problem scenario are determined, including but not limited to the following two cases: In one possible implementation, if the sample service work order belonging to the target work order problem scenario contains corpus related to the troubleshooting solution and corpus related to the processing solution, the field value of the troubleshooting solution field and the field value of the processing solution field corresponding to the target work order problem scenario are extracted from the corpus related to the troubleshooting solution and the corpus related to the processing solution.

[0058] In this case, since the sample service ticket under the target ticket problem scenario contains corpus related to the troubleshooting solution, the field value of the troubleshooting solution field can be extracted from the corpus related to the troubleshooting solution. Similarly, since the sample service ticket under the target ticket problem scenario contains corpus related to the processing solution, the field value of the processing solution field can be extracted from the corpus related to the processing solution. If there is only one sample service ticket containing the expectation related to the troubleshooting solution in the target ticket problem scenario, the field value of the troubleshooting solution field is directly extracted from the expectation related to the troubleshooting solution contained in the sample service ticket. If there is only one sample service ticket containing the corpus related to the processing solution in the target ticket problem scenario, the field value of the processing solution field is directly extracted from the corpus related to the processing solution. If multiple sample service tickets contain the corpus related to the troubleshooting solution in the target ticket problem scenario, and the multiple sample service tickets contain the same or similar expectations related to the troubleshooting solution, the field value of the troubleshooting solution field is directly extracted from the expectations related to the troubleshooting solution contained in the multiple sample service tickets. If multiple sample service tickets contain the corpus related to the processing solution in the target ticket problem scenario, and the multiple sample service tickets contain the same or similar expectations related to the troubleshooting solution, the field value of the troubleshooting solution field is directly extracted from the expectations related to the troubleshooting solution contained in the multiple sample service tickets. If the corpus is the same or similar, the field value of the solution field is directly extracted from the solution-related corpus contained in the multiple sample service tickets; if multiple sample service tickets in the target ticket problem scenario contain corpus related to the troubleshooting solution, but the sample service tickets contain different corpus related to the troubleshooting solution, then multiple different corpora can be voted, and according to the voting results, the field value of the troubleshooting solution field is extracted from the expectation related to the troubleshooting solution contained in the sample service ticket with the highest number of votes. If multiple sample service tickets in the target ticket problem scenario contain corpus related to the solution, but the sample service tickets contain different corpus related to the solution, then multiple different expectations can be voted, and according to the voting results, the field value of the solution field is extracted from the solution-related corpus contained in the sample service ticket with the highest number of votes. When extracting the field value of the solution field from the solution-related corpus, a large language model can be called or natural language processing technology can be used to perform semantic analysis on the solution-related corpus, and the field value of the solution field is extracted from it through semantic analysis. Similarly, when extracting the field value of the troubleshooting plan field from the corpus related to the troubleshooting plan, a large language model can be called or natural language processing technology can be used to perform semantic analysis on the corpus related to the troubleshooting plan, and through semantic analysis, the field value of the troubleshooting plan field can be extracted.

[0059] In another possible implementation, if the sample service work order belonging to the target work order problem scenario does not contain corpus related to the troubleshooting solution and corpus related to the processing solution, the second largest language model can be called to generate the field value of the troubleshooting solution field and the field value of the processing solution field corresponding to the target work order problem scenario. For example, the sample service work order belonging to the target work order problem scenario can be filled into the specified position of the second prompt template to obtain the prompt word information, and then the prompt word information is input into the second largest language model to instruct the second largest language model to analyze and summarize the field value of the troubleshooting solution field and the field value of the processing solution field from the information filled in the specified position of the third prompt template.

[0060] Furthermore, with the use of the work order knowledge base, when the sample service work order belonging to the target work order problem scenario contains corpus related to the troubleshooting solution and corpus related to the processing solution, the field value of the troubleshooting solution field can be extracted from the prediction related to the troubleshooting solution contained in the sample service work order, and the field value of the processing solution field can be extracted from the corpus related to the processing solution contained in the sample service work order. Then, the extracted field value of the troubleshooting solution field and the field value of the processing solution field can be replaced with the field value of the troubleshooting solution field and the field value of the processing solution field generated by the target work order problem scenario to improve the accuracy of the 4K scene label.

[0061] In another embodiment of the present application, since the first large language model performs scenario recognition on the sample service work orders included in the service work order set in batches, and the sample service work orders included in the same service work order set are of the same category, this makes it possible that the same or similar scenarios exist in the work order problem scenarios generated by the first large language model for at least one batch of sample service work orders included in the service work order set. To this end, based on the sample service work orders in any of the service work order sets that belong to the target work order problem scenario, before determining the field value of the troubleshooting plan field and the field value of the processing plan field corresponding to the target work order problem scenario, the work order problem scenarios corresponding to at least one batch of sample service work orders included in the same service work order set may also be merged.

[0062] For any service work order set, when merging the work order problem scenarios corresponding to at least one batch of sample service work orders included in the service work order set, the third largest language model can be called to merge at least one work order problem scenario corresponding to at least one batch of sample service work orders to obtain the merged work order problem scenario corresponding to the service work order set. When calling the fourth largest language model for merging, the work order problem scenarios corresponding to at least one batch of sample service work orders included in the service work order set can be filled into the third prompt template to obtain prompt word information, and then the prompt word information is input into the fourth largest language model, so that the third largest language model merges the work order problem scenarios corresponding to at least one batch of sample service work orders included in the service work order set to obtain the merged work order problem scenario corresponding to the service work order set. Among them, the third prompt template is used to instruct the third largest language model to merge the same or similar work order problem scenarios in the work order problem scenarios corresponding to at least one batch of sample service work orders included in any service work order set. The content of the third prompt template can be: #Role Now that you are a scenario analysis expert and are very good at analyzing ticket scenarios, I will give you a few simple ticket scenarios. The scenarios contain the scenario ID and the problem.

[0063] #Task You need to merge similar work order scenarios, and your output should tell me the problem phenomenon after the merger, as well as the included scenario IDs.

[0064] #Require The answer format is: New problem phenomenon: ****** Included scenario id: ***** Note: Only the information of the merged scenario is given, no other information is needed. The merged ticket scenario should contain as many similar simple ticket scenarios as possible, but the simple ticket scenarios cannot be arranged in order. The new problem description should be as simple as possible, no more than 20 characters. Here are a few simple ticket scenarios I give you #data Scenario id: {}; Problem phenomenon: {}.

[0065] Accordingly, when determining the field value of the troubleshooting solution field and the field value of the handling solution field corresponding to the target work order problem scenario based on the sample service work orders belonging to the target work order problem scenario in any of the service work order sets, the field value of the troubleshooting solution field and the field value of the handling solution field corresponding to the same merged work order problem scenario can be determined based on the sample service work orders belonging to the same merged work order problem scenario in any of the service work order sets.

[0066] Furthermore, since the at least one work order problem scenario generated by calling the first largest language model for each service work order set only includes the field value of the problem phenomenon field and the field value of the key error field, after obtaining the field value of the troubleshooting solution field and the field value of the processing solution field corresponding to each service work order set in at least one work order problem scenario, the troubleshooting solution field and its field value and the processing solution field and its field value corresponding to each work order problem scenario can be mounted to the corresponding work order problem scenario, thereby obtaining at least one work order problem scenario corresponding to each service work order set. The work order problem scenario is a 4K scene, including the problem phenomenon field, the key error field, the troubleshooting solution field and the processing solution field, and also includes the field values ​​of these fields.

[0067] In another embodiment of the present application, each service work order set obtained by the clustering operation is labeled with an N-level work order category label, and there may be at least two service work order sets with the same N-level label in the at least one service work order set. In order to distinguish different service work order sets, the embodiment of the present application can also generate a corresponding N+1-level work order category label for each of the at least one service work order sets, and then build a work order knowledge base based on the corresponding N+1-level work order category label generated for each of the at least one service work order sets, and at least one work order problem scenario corresponding to the at least one service work order set. Among them, when generating a corresponding N+1-level work order category label for each of the at least one service work order sets, it can be generated with the help of the fourth largest language model, and it can also be labeled manually. Taking the use of the fourth largest language model to generate a corresponding N+1-level work order category label for each of the at least one service work order sets as an example, the process includes the following steps: In the first step, second prompt word information corresponding to at least one service work order set is generated according to work order problem information described in a sample service work order included in at least one service work order set.

[0068] Based on the work order problem information described in the sample service work orders included in each service work order set, according to the requirements of the fourth prompt template, the information that meets the requirements in the sample service work orders included in each service work order set is filled into the corresponding position of the fourth prompt template to obtain the second prompt word information corresponding to each service work order set. Among them, the fourth largest language model can be the Qwen2.5 large model, etc. The fourth prompt template is used to instruct the fourth largest language model to process the sample service work orders included in each filled service work order set, and summarize the work order category label corresponding to each service work order set. The content of the fourth prompt template can be: #Role Now that you are an expert in problem summarization and are very good at summarizing problems, I will give you the problem descriptions of the tickets that have been categorized. Please name these categories.

[0069] #data There are {} classes below The {}th class: {} #Task Please name each class based on the description of the ticket issues in this class.

[0070] The name cannot be the same as the name of other classes and must be clearly distinguishable. Numeric suffixes are not allowed.

[0071] #Require The output format is: Class name: ***; Explanation: *** The second step is to input at least one second prompt word information into the fourth language model, so that the fourth language model generates a corresponding work order category label for at least one service work order set based on the at least one second prompt word information.

[0072] After obtaining the second prompt word information corresponding to each service ticket set, the second prompt word information is input into the second largest language model respectively. The second largest language model identifies the information filled in the corresponding position of the second prompt information based on the indication of the second prompt word information, thereby generating the corresponding N+1-th level ticket category label for each service ticket set respectively.

[0073] Based on the N+1-level work order category label generated for at least one service work order set, the N+1-level work order category label of each service work order set can be used as an index, and different work order problem scenarios and sample service work orders included under the same N+1-level work order category label can be used as index content to build a work order knowledge base. The work order knowledge base can be used to review service work orders, and can also determine the work order problem scenario to which the service work order belongs after obtaining a new service work order, so as to provide processing suggestions based on the determined work order problem scenario. The constructed work order knowledge base presents a tree structure according to the label level. The tree structure is shown in FIG. Figure 3 .

[0074] See also Figure 3For a certain product, primary classification, secondary classification and tertiary classification can be performed. Through primary classification, first-level work order category label 1, first-level work order category label 2, first-level work order category label 3, ... can be obtained. For first-level work order category label 1, secondary classification can be performed to obtain second-level work order category label 1, second-level work order category label 2, second-level work order category label 3, ... under first-level work order category label 1. Among them, the labels obtained by primary classification and the labels obtained by secondary classification can be maintained manually. For the second-level work order category label 1 under the first-level work order category label 1, by performing three-level classification, we can obtain the third-level work order category label 1, third-level work order category label 2, third-level work order category label 3, ... included in the second-level work order category label 1 under the first-level work order category label 1, wherein the third-level work order category label 1 includes 4K scene 1, 4K scene 2, 4K scene 3, ..., each 4K scene includes fields such as problem phenomenon, key error, troubleshooting plan and processing plan and the field values ​​of these fields, and each 4K scene includes multiple service work orders. After obtaining the three-level work order category label of each service work order set, different work order problem scenarios under the same third-level work order category label and the three-level work order category labels of the sample service work orders included, the three-level work order category label can be used as the index, and the different work order problem scenarios under the same third-level work order category label and the sample service work orders included are used as the index content to construct. Figure 3 The ticket knowledge base shown.

[0075] Furthermore, in order to improve the accuracy of the constructed work order knowledge base, after obtaining at least one work order problem scenario corresponding to each service work order set, the at least one work order problem scenario corresponding to each service work order set can be manually calibrated to obtain at least one calibrated work order problem scenario corresponding to each service work order set. Accordingly, when constructing the work order knowledge base, the N+1-level work order category label of each service work order set can be used as an index, and the different calibrated work order problem scenarios and the sample service work orders included under the N+1-level work order category label are used as index content to construct the work order knowledge base. By calibrating the constructed service work order database, the accuracy of the constructed work order knowledge base is improved.

[0076] For the above service ticket classification process, the following Figure 4 See Figure 4 , the process includes the following steps: Step 1. Data cleaning.

[0077] Data cleaning is performed on the raw data provided, and tools such as regular expressions are used to remove http links, scheme codes, etc. in the raw data, and the data format is unified as input for the 4key large model.

[0078] Step 2. Vectorize the model.

[0079] Two vectorization processes are designed for historical data and new data. For historical data with labels, the BGE model is first used to initialize the work order data for vectorization, and then the feature converter ResNet is trained using the initialization vector, so that the work order data under the same category have similarities in the high-dimensional space, and the work order data of different categories have differences in the high-dimensional space. For new data without labels, the BGE model is first used to initialize the work order data for vectorization, and then the trained feature converter ResNet neural network is used to transform its features.

[0080] Step 3, clustering model.

[0081] With the help of the property of "similar within a class and different between classes" of vector data after feature transformation in high-dimensional space, the K-means clustering algorithm is used to cluster the data, and the class is obtained by minimizing the distance between each data and the cluster center. Then the large language model is called to generate a category name for each class.

[0082] Step 4: Summarize and refine the model.

[0083] 1) For 4key scene generation: Use prompt project to generate 4key scene for data; 2) Scenario merging: Due to the limitation of input tokens of the big model, work orders under the same three-level classification need to generate scenarios in batches according to the number of work orders, which will generate the same or similar scenarios, resulting in duplication of 4key scenarios. With the language understanding ability of the big model, the same or similar scenarios are merged.

[0084] 3) Solution mounting: Match the solution used in the work order processing process and mount it to the corresponding 4key scene; 4) Generation of troubleshooting suggestions: For missing 4key scenarios and work order data, with the help of the language generation capability of the big model, troubleshooting suggestions are generated using work order summaries, solution processes and other information.

[0085] Step 5. Evaluation and calibration: For the generated 4key scene, manual error correction and calibration are supported to ensure data accuracy.

[0086] The embodiment of the present application utilizes the computing and evaluation capabilities of the large model to re-learn historical service tickets in actual scenarios, automatically generate problem phenomena and root causes, and associate the troubleshooting methods and solutions for the corresponding phenomena, so that the coverage of scenarios is gradually improved, avoiding misunderstandings caused by manual maintenance, and reducing the labor cost of summarizing and reviewing, gradually moving towards the direction of intelligent operation and maintenance with automated association, automated labeling, and automated push.

[0087] All the above optional technical solutions can be arbitrarily combined to form optional embodiments of the present application, which will not be described one by one here.

[0088] The present application embodiment provides a service work order classification method, which applies the work order knowledge base constructed in the above embodiment, and takes the electronic device executing the present application embodiment as an example, see Figure 5 , the method flow provided in the embodiment of the present application includes: 501. Get the target service work order to be processed.

[0089] The target service work order is an unprocessed service work order obtained by the user in an actual application scenario. After obtaining the target service work order, the user can label the target service work order with a first-level work order category label and a second-level work order category label, and then provide the target service work order to the electronic device, so that the electronic device can determine the target work order problem scenario for the target service work order, and then provide the operation and maintenance personnel with processing guidance information based on the problem phenomenon field, key error field, troubleshooting solution field, and processing solution field included in the target work order problem scenario and the field values ​​of these fields.

[0090] 502. Determine a target feature vector corresponding to the target service work order.

[0091] Optionally, for the acquired target service work order, a regularization formula written by the user may be used to perform data cleansing on the target service work order to obtain a cleaned target service work order, and then a preset format may be used to format the cleaned target service work order.

[0092] In this step, the vector initialization model can be called to extract features of the target service work order to obtain the initial feature vector corresponding to the target service work order, and then the vector conversion model can be called to convert the initial feature vector of the target service work order to obtain the target feature vector corresponding to the target service work order.

[0093] 503. Based on the target feature vector corresponding to the target service work order and the work order knowledge base, determine the target work order problem scenario to which the target service work order belongs.

[0094] Taking the work order knowledge base including three-level work order category labels as an example, after obtaining the target feature vector corresponding to the target service work order, the first-level work order category label index of the target service work order can be used to query the first-level work order category label that is the same as the first-level work order category label from the work order knowledge base, and then query the second-level work order category label that is the same as the second-level work order category label corresponding to the target service work order under the first-level work order category label, obtain each third-level work order category label included in the second-level work order category label in the work order knowledge base, and then calculate the similarity between the target feature vector corresponding to the target service work order and the target feature vector of each service work order corresponding to each third-level work order category label. According to the similarity calculation result, the work order problem scenario corresponding to the third-level work order category label with the greatest similarity is determined as the target work order problem scenario, and the target work order problem scenario is used to provide processing suggestions for the target service work order. The above example uses the work order knowledge base including three levels of work order category labels. Of course, if the work order knowledge base includes more than three levels of labels, the labels can be matched layer by layer according to the above method to find the index with the longest match with the work order category label marked with the target service work order, and then based on the target feature vector corresponding to the target service work order, the target work order problem scenario can be further found through similarity calculation.

[0095] Furthermore, after determining the target work order problem scenario, the problem phenomenon field, key error field, troubleshooting solution field and processing solution field and the field values ​​of these fields included in the target work order problem scenario can be obtained, and then these fields and their field values ​​can be provided to the user, thereby providing guidance and suggestions for the processing of the target service work order.

[0096] All the above optional technical solutions can be arbitrarily combined to form optional embodiments of the present application, which will not be described one by one here.

[0097] Please refer to Figure 6 , which shows a schematic diagram of the structure of a service ticket classification device provided in an embodiment of the present application. The device can be implemented by software, hardware, or a combination of both, and becomes all or part of an electronic device. The device includes: A clustering module 601 is used to perform a clustering operation on at least one sample service work order to obtain at least one service work order set, wherein the sample service work orders included in each service work order set in the at least one service work order set have the same work order category; Identification module 602 is used to perform scenario recognition processing on the sample service work orders included in any service work order set based on the first language model and the first prompt word information, to obtain at least one work order problem scenario and the sample service work orders corresponding to each work order problem scenario in any service work order set, each work order problem scenario includes a problem scenario description field and its field value, and the problem scenario description field includes a problem phenomenon field, a key error field, a troubleshooting solution field and a processing solution field.

[0098] In another embodiment of the present application, the device further comprises: A data cleaning module, used to clean the data of the at least one sample service work order to obtain at least one cleaned sample service work order; The formatting module is used to format the at least one cleaned sample service work order using a preset format.

[0099] In another embodiment of the present application, the clustering module 601 is used to determine the target feature vector corresponding to each of the at least one service work orders, and the target feature vector can make the sample service work orders belonging to the same work order category closer in the vector space; based on the target feature vector corresponding to each of the at least one service sample work orders, the at least one sample service work order is clustered to obtain the at least one service work order set.

[0100] In another embodiment of the present application, a determination module is used to call a vector initialization model to perform feature extraction on the at least one sample service work order to obtain an initial feature vector corresponding to the at least one sample service work order; and call a vector conversion model to convert the initial feature vector of the at least one sample service work order to obtain a target feature vector corresponding to the at least one sample service work order.

[0101] In another embodiment of the present application, the at least one sample service work order includes a plurality of historical service work orders, and the plurality of historical service work orders are respectively marked with N-level work order category labels, where N is greater than or equal to 1; the apparatus further includes: The determination module is further used to call the vector initialization model to determine the initial feature vectors corresponding to the multiple historical service work orders; A conversion module, used to call the vector conversion model to be trained, convert the initial feature vectors corresponding to the multiple historical service work orders, and obtain the target feature vectors corresponding to the multiple historical service work orders; A determination module, configured to determine N loss function values ​​corresponding to the N-level work order category labels and the target feature vectors corresponding to the multiple historical service work orders, respectively; wherein the N loss function values ​​respectively reflect that the target feature vectors corresponding to the historical service work orders with the same-level work order category labels have a closer distance; The determination module is also used to determine a target loss function value based on the N loss function values, so as to train the vector conversion model to be trained according to the target loss function value.

[0102] In another embodiment of the present application, a clustering module is used to use a preset clustering algorithm to perform a clustering operation on at least one target feature vector corresponding to the at least one sample service ticket to obtain at least one cluster; the sample service tickets corresponding to the target feature vector in any cluster are grouped into an initial service ticket set to obtain at least one initial service ticket set; and in response to a user's adjustment operation, the at least one initial service ticket set is adjusted to obtain the at least one service ticket set.

[0103] In another embodiment of the present application, an identification module is used to process the sample service work orders included in any one of the service work order sets in batches to obtain at least one batch of sample service work orders; sequentially generate first prompt word information corresponding to the at least one batch of sample service work orders; respectively input the first prompt word information corresponding to the at least one batch of sample service work orders into the first large language model, so that the first large language model generates at least one work order problem scenario corresponding to the at least one batch of sample service work orders; wherein any work order problem scenario includes the field value of the problem phenomenon field and the field value of the key error field corresponding to the corresponding batch of sample service work orders; based on the sample service work orders in any one of the service work order sets that belong to the target work order problem scenario, determine the field value of the troubleshooting plan field and the field value of the processing plan field corresponding to the target work order problem scenario, and the target work order problem scenario is any one of the at least one work order problem scenario.

[0104] In another embodiment of the present application, the identification module is used to extract from the corpus related to the troubleshooting solution and the corpus related to the processing solution if the sample service work order belonging to the target work order problem scenario contains corpus related to the troubleshooting solution and corpus related to the processing solution; if the sample service work order belonging to the target work order problem scenario does not contain corpus related to the troubleshooting solution and corpus related to the processing solution, then call the second largest language model to generate the field value of the troubleshooting solution field and the field value of the processing solution field corresponding to the target work order problem scenario.

[0105] In another embodiment of the present application, the device further comprises: a merging module, configured to call the third language model to merge at least one work order problem scenario corresponding to the at least one batch of sample service work orders, to obtain a merged work order problem scenario corresponding to any one set of service work orders; The identification module is used to determine the field value of the troubleshooting solution field and the field value of the processing solution field corresponding to the same merged work order problem scenario based on the sample service work orders in any service work order set that belong to the same merged work order problem scenario.

[0106] In another embodiment of the present application, each service work order set in the at least one service work order set is marked with an N-level work order category label, and the device further includes: A generating module, configured to generate a corresponding N+1-th level work order category label for each of the at least one service work order set; A construction module is used to construct a work order knowledge base using the N+1-level work order category label of at least one service work order set as an index and different work order problem scenarios and sample service work orders included under the same N+1-level work order category label as index content.

[0107] In another embodiment of the present application, a generation module is used to generate at least one second prompt word information corresponding to at least one service work order set based on the work order problem information described in the sample service work order included in the at least one service work order set; and input the at least one second prompt word information into the fourth largest language model respectively, so that the fourth largest language model generates the corresponding N+1-th level work order category label for the at least one service work order set based on the at least one second prompt word information.

[0108] In another embodiment of the present application, the device further comprises: The acquisition module is used to obtain the target service work order to be processed; A determination module, used to determine a target feature vector corresponding to the target service work order; The determination module is also used to determine the target work order problem scenario to which the target service work order belongs based on the target feature vector corresponding to the target service work order and the work order knowledge base, and the target work order problem scenario is used to provide processing suggestions for the target service work order.

[0109] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0110] Figure 7The structure block diagram of an electronic device 700 provided by an exemplary embodiment of the present application is shown. Generally, the electronic device 700 includes: a processor 701 and a memory 702 .

[0111] The processor 701 can be implemented in at least one of the following hardware forms: DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 701 may also include a main processor and a coprocessor, wherein the main processor is a processor for processing data in an awake state; and the coprocessor is a low-power processor for processing data in a standby state. In some embodiments, the processor 701 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 701 may also include an artificial intelligence processor, which is used to process computing operations related to machine learning.

[0112] The memory 702 may include one or more computer-readable storage media, which may be non-temporary computer-readable storage media, for example, the non-temporary computer-readable storage media may be CD-ROM (Compact Disc Read-Only Memory), ROM, RAM (Random Access Memory), magnetic tape, floppy disk and optical data storage device, etc. The computer-readable storage medium stores at least one computer program, which, when executed, can implement the method for constructing the work order knowledge base or the method for processing the service work order. Of course, the above electronic device may also include other components, such as input / output interface, communication component, etc. The input / output interface provides an interface between the processor and the peripheral interface module, and the above peripheral interface module may be an output device, an input device, etc. The communication component is configured to facilitate wired or wireless communication between the electronic device and other devices.

[0113] Those skilled in the art will understand that Figure 7 The structure shown in the figure does not constitute a limitation on the electronic device 700, and may include more or less components than those shown in the figure, or combine some components, or adopt a different component arrangement.

[0114] An embodiment of the present application provides a computer-readable storage medium, in which at least one computer program is stored. When the at least one computer program is executed by a processor, it can implement the above-mentioned work order knowledge base construction method, or the above-mentioned service work order processing method.

[0115] An embodiment of the present application provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it can implement the above-mentioned work order knowledge base construction method or the above-mentioned service work order processing method.

[0116] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A service work order classification method, characterized in that: The method comprises: Performing a clustering operation on at least one sample service work order to obtain at least one service work order set, wherein the sample service work orders included in each service work order set in the at least one service work order set have the same work order category; For any service ticket set, based on the first language model and the first prompt word information, scenario recognition processing is performed on the sample service tickets included in the any service ticket set to obtain at least one ticket problem scenario and the sample service tickets corresponding to each ticket problem scenario in the any service ticket set. Each ticket problem scenario includes a problem scenario description field and its field value. The problem scenario description field includes a problem phenomenon field, a key error field, a troubleshooting solution field and a processing solution field.

2. The method according to claim 1, characterized in that The clustering operation is performed on at least one sample service work order to obtain at least one service work order set, including: Determine a target feature vector corresponding to each of the at least one service work order, wherein the target feature vector can make sample service work orders belonging to the same work order category closer in the vector space; Based on the target feature vector corresponding to each of the at least one sample service work orders, a clustering operation is performed on the at least one sample service work order to obtain the at least one service work order set.

3. The method according to claim 2, characterized in that The determining of the target feature vector corresponding to each of the at least one sample service work orders includes: Calling a vector initialization model to perform feature extraction on the at least one sample service work order to obtain an initial feature vector corresponding to the at least one sample service work order; The vector conversion model is called to convert the initial feature vector of the at least one sample service work order to obtain a target feature vector corresponding to the at least one sample service work order.

4. The method according to claim 3, characterized in that The at least one sample service work order includes a plurality of historical service work orders, each of which is marked with an N-level work order category label, where N is greater than or equal to 1; the method further includes: Calling a vector initialization model to determine initial feature vectors corresponding to the multiple historical service work orders; Calling the vector conversion model to be trained, converting the initial feature vectors corresponding to the multiple historical service work orders, and obtaining target feature vectors corresponding to the multiple historical service work orders; According to the N-level work order category labels and target feature vectors corresponding to the multiple historical service work orders, N loss function values ​​corresponding to the N-level work order category labels are determined; wherein the N loss function values ​​respectively reflect that the target feature vectors corresponding to the historical service work orders with the same level of work order category labels have a closer distance; Based on the N loss function values, a target loss function value is determined to train the vector conversion model to be trained according to the target loss function value.

5. The method according to claim 2, characterized in that: The step of performing a clustering operation on the at least one sample service work order based on the target feature vector corresponding to each of the at least one sample service work order to obtain the at least one service work order set includes: Using a preset clustering algorithm, clustering the at least one target feature vector corresponding to the at least one sample service work order to obtain at least one cluster; The sample service orders corresponding to the target feature vector in each cluster are combined into an initial service order set to obtain at least one initial service order set; In response to the user's adjustment operation, the at least one initial service ticket set is adjusted to obtain the at least one service ticket set.

6. The method according to claim 1, characterized in that The method of performing scene recognition processing on the sample service work orders included in any one of the service work order sets based on the first language model and the first prompt word information to obtain at least one work order problem scene includes: Processing the sample service work orders included in any one of the service work order sets in batches to obtain at least one batch of sample service work orders; Sequentially generate first prompt word information corresponding to the at least one batch of sample service work orders; Inputting the first prompt word information corresponding to the at least one batch of sample service work orders into the first large language model respectively, so that the first large language model generates at least one work order problem scenario corresponding to the at least one batch of sample service work orders; wherein each work order problem scenario includes a field value of a problem phenomenon field and a field value of a key error field corresponding to the corresponding batch of sample service work orders; Based on a sample service order in any of the service order sets that belongs to a target order problem scenario, determine a field value of a troubleshooting solution field and a field value of a processing solution field corresponding to the target order problem scenario, wherein the target order problem scenario is any one of the at least one order problem scenario.

7. The method according to claim 6, characterized in that The determining, based on a sample service ticket belonging to a target service ticket problem scenario in any of the service ticket sets, a field value of a troubleshooting solution field and a field value of a processing solution field corresponding to the target service ticket problem scenario, comprises: If the sample service ticket belonging to the target ticket problem scenario contains corpus related to the troubleshooting solution and corpus related to the processing solution, extract the field value of the troubleshooting solution field and the field value of the processing solution field corresponding to the target ticket problem scenario from the corpus related to the troubleshooting solution and the corpus related to the processing solution; If the sample service work order belonging to the target work order problem scenario does not contain corpus related to the troubleshooting solution and corpus related to the processing solution, the second largest language model is called to generate the field value of the troubleshooting solution field and the field value of the processing solution field corresponding to the target work order problem scenario.

8. The method according to claim 6, characterized in that Before determining the field value of the troubleshooting solution field and the field value of the processing solution field corresponding to the target work order problem scenario based on the sample service work order belonging to the target work order problem scenario in any of the service work order sets, the method further includes: Invoking the third language model to merge at least one work order problem scenario corresponding to the at least one batch of sample service work orders to obtain a merged work order problem scenario corresponding to any one of the service work order sets; The determining, based on a sample service ticket belonging to a target service ticket problem scenario in any of the service ticket sets, a field value of a troubleshooting solution field and a field value of a processing solution field corresponding to the target service ticket problem scenario, comprises: Based on the sample service tickets in any one of the service ticket sets that belong to the same merged ticket problem scenario, the field value of the troubleshooting solution field and the field value of the processing solution field corresponding to the same merged ticket problem scenario are determined.

9. The method according to any one of claims 1 to 8, characterized in that Each service ticket set in the at least one service ticket set is marked with an N-level ticket category label, and the method further includes: Generating a corresponding N+1th level work order category label for each of the at least one service work order set; A work order knowledge base is constructed by taking the N+1-level work order category label of at least one service work order set as an index and taking different work order problem scenarios and the included sample service work orders under the same N+1-level work order category label as index content.

10. The method according to claim 9, characterized in that Generating a corresponding N+1th level work order category label for each of the at least one service work order set includes: Generate at least one second prompt word information corresponding to the at least one service work order set according to the work order problem information described in the sample service work order included in the at least one service work order set; The at least one second prompt word information is respectively input into the fourth largest language model, so that the fourth largest language model generates a corresponding N+1th level work order category label for the at least one service work order set based on the at least one second prompt word information.

11. The method according to claim 9, characterized in that The method further comprises: Get the target service work order to be processed; Determine a target feature vector corresponding to the target service work order; Based on the target feature vector corresponding to the target service work order and the work order knowledge base, a target work order problem scenario to which the target service work order belongs is determined.

12. An electronic device, characterized in that: It comprises a processor and a memory; the memory stores at least one program code; the at least one program code is used to be called and executed by the processor to implement the service work order classification method as described in any one of claims 1 to 11.

13. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one computer program, and when the at least one computer program is executed by a processor, it can implement the service ticket classification method described in any one of claims 1 to 11.

14. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the service ticket classification method according to any one of claims 1 to 11 can be implemented.

Images