Abnormal order repair method, system and device, storage medium and product

Abstract

The invention discloses an abnormal order restoration method, system and device, a storage medium and a product, and relates to the technical field of order management, and the abnormal order restoration method comprises the steps: inputting order data into a preset abnormal order classification model, and obtaining each abnormal order; taking the abnormal orders, which are not matched with a preset abnormal order rule base, in the abnormal orders as to-be-classified abnormal orders; classifying each abnormal order to be classified through a clustering algorithm to obtain each abnormal category; and automatically repairing each abnormal order according to the repairing scheme in the abnormal order rule base and the repairing scheme corresponding to each abnormal category. The technical scheme of the invention aims to solve the technical problem of how to reduce the repair cost of the abnormal order.

Description

Technical Field

[0001] The present application relates to the technical field of order management, and in particular to an abnormal order repair method, system, device, storage medium and product. Background Art

[0002] With the rapid growth of IoT business, the number of orders that need to be processed every day has exceeded 800,000. However, any potential problems such as network fluctuations, equipment failures, or instability of external systems may cause a sharp increase in the number of abnormal orders in the IoT system, resulting in a large number of abnormal orders being in a pending state.

[0003] Nowadays, abnormal orders are usually repaired manually by operation and maintenance personnel. However, with the surge in abnormal orders, the cost of manual repair is also greatly increased. Therefore, how to reduce the cost of repairing abnormal orders is a technical problem that needs to be solved by those skilled in the art. Summary of the invention

[0004] The main purpose of this application is to provide a method, system, device, storage medium and product for repairing abnormal orders, aiming to solve the technical problem of how to reduce the cost of repairing abnormal orders.

[0005] To achieve the above objectives, the present application provides an abnormal order repair method, which includes:

[0006] Input the order data into the preset abnormal order classification model to obtain each abnormal order;

[0007] The abnormal orders that do not match the preset abnormal order rule base among the abnormal orders are regarded as abnormal orders to be classified;

[0008] Classifying each of the abnormal orders to be classified by a clustering algorithm to obtain each abnormal category;

[0009] Each of the abnormal orders is automatically repaired according to the repair solutions in the abnormal order rule base and the repair solutions corresponding to each of the abnormal categories.

[0010] In one embodiment, the solution for automatically repairing each of the abnormal orders according to the repair solutions in the abnormal order rule base and the repair solutions corresponding to each of the abnormal categories includes:

[0011] Determine a target abnormal order among the abnormal orders, and obtain an order execution table of the target abnormal order, wherein the order execution table is an order context table when multiple orders are executed in sequence to implement business logic;

[0012] Taking the abnormal orders belonging to the order execution table among the abnormal orders as the current pending orders, and sorting the current pending orders based on the execution order of the current pending orders in the order execution table to obtain a pending order sequence;

[0013] The first current order to be processed in the sequence of pending orders is taken as an initial abnormal order, and starting from the initial abnormal order, each current order to be processed in the sequence of pending orders is automatically repaired according to the repair solutions in the abnormal order rule base and the repair solutions corresponding to each abnormal category;

[0014] Delete each of the currently pending orders from among the abnormal orders, and return to execute the step of determining the target abnormal order from among the abnormal orders until all of the abnormal orders are repaired.

[0015] In one embodiment, the step of determining a target abnormal order among the abnormal orders includes:

[0016] Randomly determine a target abnormal order from among the abnormal orders;

[0017] or,

[0018] The priorities corresponding to the respective abnormal orders are obtained, and the abnormal order corresponding to the maximum priority among the priorities is taken as the target abnormal order.

[0019] In one embodiment, the step of automatically repairing each of the current pending orders in the pending order sequence based on the repair solutions in the abnormal order rule base and the repair solutions corresponding to each of the abnormal categories, taking the initial abnormal order as the starting point, comprises:

[0020] Taking the initial abnormal order as the traversal starting point, traverse the sequence of orders to be processed;

[0021] Obtain the order processing stage corresponding to the current pending order being traversed, and fill in the preset global variable template corresponding to the order processing stage;

[0022] Determine whether the currently pending order being traversed belongs to an abnormal order to be classified;

[0023] If it is detected that the traversed current pending order belongs to an abnormal order to be classified, the current pending order is automatically repaired based on the repair solution of the abnormal category to which the current pending order belongs and the filled global variable template;

[0024] If it is detected that the current pending order being traversed does not belong to the abnormal order to be classified, the current pending order is automatically repaired based on the repair plan in the abnormal order rule library and the filled global variable template.

[0025] In one embodiment, before the step of inputting the order data into a preset abnormal order classification model to obtain each abnormal order, the method further includes:

[0026] Pre-training of preset neural network models based on common data sets;

[0027] After detecting that the pre-training is completed, the neural network model is spliced ​​with a preset order classification model to obtain an initial classification model;

[0028] The historical order data is used as model input data, and the binary classification labels of whether each order in the historical order data is abnormal are used as model training labels to train the initial classification model;

[0029] The trained initial classification model is used as an abnormal order classification model.

[0030] In one embodiment, after the step of automatically repairing each of the abnormal orders according to the repair solutions in the abnormal order rule base and the repair solutions corresponding to each of the abnormal categories, the method further includes:

[0031] According to each of the abnormal categories and the corresponding repair solutions of each of the abnormal categories, the abnormal order rule library is updated;

[0032] Based on the updated abnormal order rule base, the abnormal order classification model is updated.

[0033] In one embodiment, the step of classifying each of the abnormal orders to be classified by a clustering algorithm to obtain each abnormal category includes:

[0034] Extracting the key entities corresponding to each of the abnormal orders to be classified from the actual scene labels corresponding to each of the abnormal orders to be classified, wherein the key entities are information units that characterize scene features;

[0035] Each of the key entities is used as a data point, and each of the data points is classified by a clustering algorithm to obtain various abnormal categories.

[0036] To achieve the above objectives, the present application also proposes an abnormal order repair system, which includes:

[0037] The order screening module is used to input the order data into the preset abnormal order classification model to obtain each abnormal order;

[0038] A rule matching module, used for treating the abnormal orders that do not match the preset abnormal order rule library among the abnormal orders as abnormal orders to be classified;

[0039] A cluster analysis module, used to classify each of the abnormal orders to be classified by a clustering algorithm to obtain each abnormal category;

[0040] The order repair module is used to automatically repair each abnormal order according to the repair solution in the abnormal order rule base and the repair solution corresponding to each abnormal category.

[0041] To achieve the above objectives, the present application also proposes an electronic device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the abnormal order repair method as described above.

[0042] An embodiment of the present application further proposes a storage medium, which is a computer-readable storage medium. A computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the abnormal order repair method as described above are implemented.

[0043] An embodiment of the present application further proposes a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the abnormal order repair method described above are implemented.

[0044] In an embodiment of the present application, by inputting order data into a preset abnormal order classification model to obtain various abnormal orders, abnormal orders can be identified in the orders obtained in real time; then, the abnormal orders in each abnormal order that do not match the preset abnormal order rule library are taken as abnormal orders to be classified, and the preset abnormal order rule library can be used as a basis to find abnormal orders that cannot be found through the abnormal order rule library among the abnormal orders identified by the model; then, each abnormal order to be classified is classified by a clustering algorithm to obtain various abnormal categories. For abnormal orders that cannot be found through the rule library, the abnormal orders to be classified with similar or similar features can be divided into one category through a clustering algorithm, so that various abnormal categories can be obtained; finally, according to the repair plan in the abnormal order rule library and the repair plan corresponding to each abnormal category, each abnormal order is automatically repaired. For abnormal orders that meet the abnormal order rule library, they can be repaired based on the repair plan specified in the rule library, and for abnormal orders whose abnormal categories have been determined, they can be repaired based on the repair plan of the abnormal category to which they belong. Therefore, for each order in the order system, the present application can achieve the effect of automatically identifying abnormal orders and automatically repairing abnormal orders. Compared with the traditional manual repair method, the present application achieves the purpose of reducing the repair cost of abnormal orders. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the embodiments of the present application, and together with the description are used to explain the principles of the embodiments of the present application.

[0046] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0047] Figure 1 A flowchart of the first embodiment of the abnormal order repair method of the present application is provided;

[0048] Figure 2 This is a module diagram of an embodiment of the abnormal order repair method of the present application;

[0049] Figure 3 This is a processing flow chart of the intelligent AI model of an embodiment of the abnormal order repair method of this application;

[0050] Figure 4 A schematic diagram of data flow for filling data in an embodiment of an abnormal order repair method of the present application;

[0051] Figure 5This is a schematic diagram of the module structure of the abnormal order repair system in the embodiment of the present application;

[0052] Figure 6 This is a schematic diagram of the device structure of the hardware operating environment involved in the abnormal order repair method in the embodiment of the present application.

[0053] The purpose, features and advantages of the embodiments of the present application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0054] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0055] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0056] With the rapid growth of IoT business, the number of orders that need to be processed every day has exceeded 800,000. However, any potential problems such as network fluctuations, equipment failures, or instability of external systems may cause a sharp increase in the number of abnormal orders in the IoT system, resulting in a large number of abnormal orders being in a pending state.

[0057] Nowadays, abnormal orders are usually repaired manually by operation and maintenance personnel. However, with the surge in abnormal orders, the cost of manual repair is also greatly increased. Therefore, how to reduce the cost of repairing abnormal orders is a technical problem that needs to be solved by those skilled in the art.

[0058] To solve the above problems, the present application obtains each abnormal order by inputting order data into a preset abnormal order classification model, and can identify abnormal orders in the orders obtained in real time; then, the abnormal orders in each abnormal order that do not match the preset abnormal order rule library are taken as abnormal orders to be classified, and the preset abnormal order rule library can be used as a basis to find abnormal orders that cannot be found through the abnormal order rule library among the abnormal orders identified by the model; then, each abnormal order to be classified is classified by a clustering algorithm to obtain each abnormal category. For abnormal orders that cannot be found through the rule library, the abnormal orders to be classified with similar or similar features can be divided into one category by a clustering algorithm, so as to obtain each abnormal category; finally, according to the repair plan in the abnormal order rule library and the repair plan corresponding to each abnormal category, each abnormal order is automatically repaired. For abnormal orders that meet the abnormal order rule library, they can be repaired based on the repair plan specified in the rule library, and for abnormal orders whose abnormal categories have been determined, they can be repaired based on the repair plan of the abnormal category to which they belong.

[0059] It can be seen that for each order in the order system, the present application can achieve the effect of automatically identifying abnormal orders and automatically repairing abnormal orders. Compared with the traditional manual repair method, the present application achieves the purpose of reducing the repair cost of abnormal orders.

[0060] It should be noted that the execution subject of the abnormal order repair method of the present application can be an electronic device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a server, etc. The following takes the server as an example to illustrate various embodiments of the abnormal order repair method of the present application.

[0061] Based on this, the embodiment of the present application provides a method for repairing abnormal orders, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the abnormal order repair method of the present application.

[0062] In this embodiment, the abnormal order repair method includes steps S10 to S40:

[0063] Step S10, inputting the order data into a preset abnormal order classification model to obtain each abnormal order;

[0064] It should be noted that the order data refers to the content of the order received in real time. The abnormal order classification model refers to a model that can identify abnormal orders and classify the abnormal orders by scenarios.

[0065] In this embodiment, the present application identifies all orders in the order data through a pre-trained abnormal order classification model, so that each abnormal order can be obtained.

[0066] In a feasible implementation manner, before the above step S10, the method further includes:

[0067] Step S100, pre-training a preset neural network model according to a general data set;

[0068] Understandably, general purpose datasets include a large amount of text, which is used to teach neural network models to understand a wide variety of texts.

[0069] Exemplarily, when the neural network model is a BERT model (Bid irect iona l Encoder Representation from Transformers, a language representation model), the MLM (Masked Language Model l) task and NSP (Next Sentence Pred ict ion) task in the BERT model can be trained based on a general data set, so as to achieve the purpose of pre-training the BERT model.

[0070] Step S200, after detecting that the pre-training is completed, the neural network model is spliced ​​with a preset order classification model to obtain an initial classification model;

[0071] It should be noted that the order classification model refers to one or more fully connected layers.

[0072] It is understandable that after pre-training, the neural network model already has the ability to understand text. Furthermore, in order to facilitate the classification of input data based on the ability to understand text, one or more fully connected layers can be added after the output layer of the neural network model. Then, the model composed of the neural network model and the fully connected layer is used as the initial classification model.

[0073] Step S300, using historical order data as model input data, using binary classification labels of whether each order in the historical order data is abnormal as model training labels, and training the initial classification model;

[0074] It should be noted that historical order data refers to order data in historical records. Binary classification labels are used to characterize whether an order is an abnormal order or whether an order is not an abnormal order.

[0075] After obtaining the initial classification model, the initial classification model can be trained based on known historical orders and the binary classification labels corresponding to the orders, so that the initial classification model has the ability to identify abnormal orders.

[0076] Step S400, using the trained initial classification model as an abnormal order classification model.

[0077] Therefore, the order data received in real time can be processed through the trained initial classification model to obtain various abnormal orders.

[0078] Step S20, taking the abnormal orders that do not match the preset abnormal order rule base among the abnormal orders as abnormal orders to be classified;

[0079] It should be noted that the abnormal order rule base is a collection of abnormal order features and abnormal order processing methods.

[0080] In a feasible implementation, each abnormal order identified by the abnormal order classification model can be compared with the abnormal order features in the abnormal order rule base to determine whether the abnormal order belongs to the type specified by the abnormal order rule base. Abnormal orders that do not belong to the type specified by the abnormal order rule base can be regarded as abnormal orders to be classified.

[0081] Step S30, classifying each of the abnormal orders to be classified by a clustering algorithm to obtain each abnormal category;

[0082] It should be noted that the clustering algorithm may be a k-means algorithm or other clustering algorithms, and the present application does not limit the type of clustering algorithm.

[0083] In this embodiment, after the abnormal orders to be classified are identified through the abnormal order rule library, the abnormal orders to be classified can be divided into multiple abnormal categories through a clustering algorithm, so as to facilitate determining the repair solutions corresponding to each abnormal category based on the abnormal category.

[0084] In a feasible implementation, after obtaining each abnormal category, the abnormal orders corresponding to each abnormal category can also be output, so as to manually check whether there is an abnormal category belonging to the abnormal order rule base. If it is detected that there is an abnormal category belonging to the abnormal order rule base, the abnormal category can be merged into the order abnormal rule base. In addition, after obtaining each abnormal category, the repair solutions corresponding to each abnormal category input by the user can also be received, so as to repair the abnormal orders corresponding to each abnormal category.

[0085] In a feasible implementation manner, the above step S30 includes:

[0086] Step S301, extracting the key entities corresponding to each of the abnormal orders to be classified from the actual scene labels corresponding to each of the abnormal orders to be classified, wherein the key entities are information units that characterize scene features;

[0087] It should be noted that the actual scenario labels include: batch account opening, fulfillment node ID, exception code, etc. Each key entity can be obtained by keyword extraction.

[0088] Step S302: taking each of the key entities as a data point, and classifying each of the data points using a clustering algorithm to obtain each abnormal category.

[0089] Exemplarily, when performing cluster analysis using the k-means algorithm, the order dimension and the chunk dimension can be used as the initial cluster centers, respectively. That is, the abnormal order is regarded as a separate cluster center (order dimension), or the data processing unit corresponding to the abnormal order is regarded as a cluster center (chunk dimension). Then, the K-means clustering algorithm is used to perform cluster analysis on the key entities, and after multiple clustering iterative calculations, the optimal key entity is obtained (that is, the K-means algorithm is used for iterative calculations, and the position of the cluster center is continuously adjusted until the convergence condition is reached (such as the cluster center no longer changes or changes very little)). Among them, the optimal key entity refers to the optimal key entity in each cluster cluster (the key entity with the highest frequency of occurrence in the cluster cluster, or the key entity most relevant to the cluster cluster). Then, based on the correspondence between the optimal key entity in each cluster and the preset category, the abnormal category represented by each optimal key entity can be determined in the preset category.

[0090] Therefore, this application can integrate various abnormal orders to be classified and improve the processing efficiency of abnormal orders.

[0091] Step S40, automatically repairing each of the abnormal orders according to the repair solutions in the abnormal order rule base and the repair solutions corresponding to each of the abnormal categories.

[0092] It can be understood that after the abnormal orders identified by the abnormal order classification model are divided into those that comply with the abnormal order rule base and those that do not comply with the abnormal order rule base, the abnormal orders that comply with the abnormal order rule base can be repaired based on the repair solutions in the rule base; and the orders that do not comply with the abnormal order rule base can be repaired based on the preset repair solutions corresponding to the abnormal categories of each order. Thus, the purpose of automatically repairing the abnormal orders can be achieved.

[0093] In the present application, by inputting order data into a preset abnormal order classification model to obtain various abnormal orders, abnormal orders can be identified in the orders obtained in real time; then, the abnormal orders in each abnormal order that do not match the preset abnormal order rule library are used as abnormal orders to be classified, and the preset abnormal order rule library can be used as the basis to find abnormal orders that cannot be found through the abnormal order rule library among the abnormal orders identified by the model; then, each abnormal order to be classified is classified by a clustering algorithm to obtain various abnormal categories. For abnormal orders that cannot be found through the rule library, the abnormal orders to be classified with similar or similar features can be divided into one category through a clustering algorithm, so that various abnormal categories can be obtained; finally, according to the repair plan in the abnormal order rule library and the repair plan corresponding to each abnormal category, each abnormal order is automatically repaired. For abnormal orders that meet the abnormal order rule library, they can be repaired based on the repair plan specified in the rule library, and for abnormal orders whose abnormal categories have been determined, they can be repaired based on the repair plan of the abnormal category to which they belong. Therefore, for each order in the order system, the present application can achieve the effect of automatically identifying abnormal orders and automatically repairing abnormal orders. Compared with the traditional manual repair method, the present application achieves the purpose of reducing the repair cost of abnormal orders.

[0094] Furthermore, based on the above-mentioned first embodiment of the abnormal order repair method of the present application, a second embodiment of the abnormal order repair method of the present application is proposed.

[0095] In this embodiment, the above step S40 includes:

[0096] Step S401, determining a target abnormal order from among the abnormal orders, and obtaining an order execution table of the target abnormal order, wherein the order execution table is an order context table when multiple orders are executed in sequence to implement business logic;

[0097] It should be noted that the target abnormal order refers to the selected abnormal order.

[0098] It is understandable that in the process of implementing a business logic, multiple orders may be involved. Moreover, multiple orders must be executed in a specific order to implement the business logic. However, when an order in the business logic chain is abnormal, other orders executed after the order may also be abnormal. Therefore, when processing multiple abnormal orders with business logic, it is also necessary to consider the order relationship between the orders, that is, the blood relationship between the orders.

[0099] Exemplarily, assuming that there are 1,000 abnormal orders in total, when repairing the abnormal orders, you can first select a target abnormal order from the 1,000 abnormal orders. Then, you can obtain the order execution table corresponding to the target abnormal order based on the id of the target abnormal order, the scenario in which it is located, and other information. For example, if the unique identifier of the target abnormal order is 12, you can use the unique identifier as the query keyword to find the order execution table corresponding to the target abnormal order. For example, the order execution table may include 20 orders (order 1-order 20, where the order number is the unique identifier of the order).

[0100] Step S402, taking the abnormal orders in the order execution table as the current pending orders, and sorting the current pending orders based on the execution order of the current pending orders in the order execution table to obtain a pending order sequence;

[0101] In this embodiment, after obtaining the order execution table, it is possible to detect whether the orders in the order execution table are abnormal orders identified by the model. If an order belonging to the order execution table is detected among the abnormal orders identified by the model, the abnormal orders belonging to the order execution table can be used as the current abnormal orders to be processed, and the current abnormal orders to be processed are sorted according to the order in the order execution table.

[0102] For example, assuming that there are order 5, order 20, and order 12 among the 1000 abnormal orders, then order 5, order 12, and order 20 are all currently pending abnormal orders. The sequence of pending orders is: order 5, order 12, and order 20.

[0103] Step S403, taking the first current order to be processed in the sequence of pending orders as an initial abnormal order, and taking the initial abnormal order as a starting point, automatically repairing each current order to be processed in the sequence of pending orders according to the repair solutions in the abnormal order rule base and the repair solutions corresponding to each abnormal category;

[0104] For example, when the sequence of orders to be processed is: order 5, order 12, order 20, order 5 is the initial abnormal order. Therefore, order 5 can be repaired first, and then order 12 and order 20 can be repaired in sequence.

[0105] In a feasible implementation, after order 5 is repaired, when preparing to repair order 12, it is also necessary to obtain the execution prerequisite of order 12 (for example, whether the data is currently valid, whether the execution of order 12 requires the execution result of order 11), and then repair order 12 based on the execution prerequisite of order 12.

[0106] For example, when executing order 12, if the timeliness of data is required to be high, the orders before order 12 can be executed on demand to facilitate repairing order 12.

[0107] Step S404, deleting each of the currently pending orders from among the abnormal orders, and returning to execute the step of determining the target abnormal order from among the abnormal orders, until all of the abnormal orders are repaired.

[0108] It is understandable that after all orders in the pending order sequence are repaired, the abnormal orders that have been repaired can be deleted from the abnormal orders. Then, the steps of determining the target abnormal order, obtaining the pending order sequence, and repairing the current pending order in the pending order sequence can be continued among the remaining abnormal orders until all abnormal orders are repaired.

[0109] In this embodiment, the present application improves the orderliness in the abnormal order repair process based on the data lineage capability between orders, thereby improving the efficiency of repairing abnormal orders.

[0110] In a feasible implementation manner, the above step S401 includes:

[0111] Step S4011, randomly determining a target abnormal order from each of the abnormal orders;

[0112] It can be understood that, in a feasible implementation, a random selection may be made from each abnormal order, and the selected abnormal order may be used as a target abnormal order.

[0113] Step S4012, obtaining the priorities corresponding to each of the abnormal orders, and taking the abnormal order corresponding to the maximum priority among the priorities as the target abnormal order.

[0114] In this embodiment, when each abnormal order has its own priority, in order to facilitate the response to the business logic with a high priority, the abnormal order corresponding to the maximum priority among the priorities can be used as the target abnormal order.

[0115] In this embodiment, the present application can determine the target abnormal orders by random sampling and priority comparison, so that in any scenario, the present application can repair the abnormal orders in an orderly manner.

[0116] In a feasible implementation manner, the above step S403 includes:

[0117] Step S4031, taking the initial abnormal order as the traversal starting point, traversing the sequence of orders to be processed;

[0118] Step S4032, obtaining the order processing stage corresponding to the current pending order traversed, and filling the preset global variable template corresponding to the order processing stage;

[0119] It is understandable that when executing an order, there are requirements for the parameters in the program execution environment. Therefore, when repairing abnormal orders, different parameters need to be determined according to the processing stage of the abnormal order. It should be noted that the order processing stage includes initialization, formatting, order transmission, order fulfillment, etc. The global variable template refers to a template that includes parameters required to execute orders such as global variables.

[0120] In this embodiment, different global variable templates may be preset for different order processing stages. Therefore, after obtaining the order processing stage of the current order to be processed, the global variable template corresponding to the order processing stage may be filled.

[0121] Step S4033, determining whether the current pending order being traversed is an abnormal order to be classified;

[0122] In this embodiment, it can be determined through the abnormal order rule library whether the currently processed order being traversed is an abnormal order to be classified.

[0123] Step S4034: if it is detected that the current pending order being traversed belongs to an abnormal order to be classified, then based on the repair solution of the abnormal category to which the current pending order belongs and the filled global variable template, the current pending order is automatically repaired;

[0124] Step S4035: If it is detected that the current pending order being traversed does not belong to the abnormal order to be classified, the current pending order is automatically repaired based on the repair solution in the abnormal order rule library and the filled global variable template.

[0125] In this embodiment, for abnormal orders that fall within the provisions of the abnormal order rule base and abnormal orders that do not fall within the provisions of the abnormal order rule base, the present application can repair them according to the order in the sequence of orders to be processed, so that the abnormal orders can be repaired in units of the order execution table.

[0126] Furthermore, based on the above-mentioned embodiments of the abnormal order repair method of the present application, a third embodiment of the abnormal order repair method of the present application is proposed.

[0127] In this embodiment, after the above step S40, the method further includes:

[0128] Step S50, updating the abnormal order rule base according to each of the abnormal categories and the repair solutions corresponding to each of the abnormal categories;

[0129] Step S60: updating the abnormal order classification model based on the updated abnormal order rule base.

[0130] In this embodiment, after obtaining the abnormal category, if it is detected that the abnormal category does not belong to the abnormal order rule base, the abnormal order rule base can be updated based on the abnormal category, so that the abnormal order rule base can be continuously updated, thereby improving the efficiency of determining abnormal orders based solely on the abnormal order rule base. At the same time, in order to improve the abnormal order recognition ability of the model, the abnormal order rule base can also be used as prior knowledge to update the abnormal order classification model to improve the abnormal order recognition ability of the model.

[0131] Furthermore, based on the above-mentioned various embodiments of the abnormal order repair method of the present application, a specific embodiment of the abnormal order repair method of the present application is proposed.

[0132] Please refer to Figure 2 In this embodiment, the abnormal order repair method may include an order repair module and an intelligent AI module. For the order repair module, data collection can be used to collect all order data in the business system, and the order data can be synchronized to the intelligent AI module; order management can realize the ability to manage abnormal orders identified by the intelligent AI module; abnormal scenario management can realize the labeling management capability of abnormal scenarios based on AI+manual; repair strategy management can realize the establishment of business order repair strategies based on abnormal scenarios and system capabilities, and can be called by repair tasks to repair orders; repair tasks can provide automated order repair task execution capabilities based on different businesses and scenarios; repair logs can be used to provide repair record logs for abnormal orders, and at the same time, repair suggestions provided by intelligent AI can be displayed for abnormal orders that the system cannot automatically repair.

[0133] For the intelligent AI module, log data, business summary, operation and maintenance summary, historical problem tickets and other data can be processed through data initialization; then, the initialized data can be used for training to identify abnormal orders and generate an abnormal scenario division model; then, the rule library can be used to manage rule models such as abnormal order identification, abnormal scenario division, and repair strategy; intelligent repair can be used to achieve rule matching, strategy matching, content completion and other capabilities in the abnormal order repair process; repair file management can be used to manage repair files for various scenarios; and repair files, abnormal scenarios, order repair strategy and other files can be synchronized to the order repair module through file transfer for the repair task to perform order repair operations.

[0134] This application uses an intelligent AI module + order repair module to repair abnormal orders. It can use historical order data and knowledge data to create an order repair model and identify and mark abnormal orders. It also builds automatic order repair capabilities to achieve automated repair capabilities for batch orders in the IoT centralized support system, thereby improving operation and maintenance efficiency.

[0135] Specifically, the processing flow of the intelligent AI model is as follows Figure 3 As shown. Specifically, the initialization capability of the intelligent AI module can be used to collect historical problem data to form the training set data of the BERT abnormal order scenario classification model; then the training set can be used to complete the first stage of pre-training using the MLM and NSP tasks in the BERT model; after the pre-training is completed, the pre-trained model is loaded, and the text classification model is fine-tuned, and the text classifier and abnormal order scenario labels are loaded into the text classification model to form an abnormal order scenario classification model; then the model training learning and accuracy verification are completed by loading the training set. Finally, the abnormal order data can be initialized as a formal data set and loaded into the scenario classification model, and the scenario labels can be marked in the abnormal order data to form a list of abnormal order scenario label data. Then, the initial data can be fully classified and labeled according to the abnormal scenario judgment criteria in the existing abnormal rule library, and the model can be generated and saved.

[0136] Specifically, the order anomaly rule library integrates order anomaly verification rules and order repair solutions corresponding to abnormal scenarios, which are used to pre-classify the abnormal scenarios of the original data when initializing the model; for the subsequent AI-identified anomalies that are not in the rule library, it is manually determined whether they belong to new abnormal scenarios, and their abnormal labels and processing solutions are determined, and the abnormal rule library is incrementally updated. Then, based on the model established based on the abnormal rule library, the abnormal scenarios are classified to identify the abnormal scenarios and their order repair solutions. The abnormal orders identified by the model include: the identified abnormal scenarios and order repair solutions that already exist in the rule library, and the unknown abnormal scenarios that do not exist in the rule library.

[0137] Furthermore, for abnormal scenarios that already exist in the rule base, the corresponding repair order solution is used to repair the order; for unknown abnormal scenarios that do not exist in the identified rule base, they can be processed through a clustering algorithm. Specifically, based on the abnormal order scenario classification model, the abnormal order scenario label data list and historical repair orders, operation and maintenance summaries, repair strategies and other data generated can be converted into word vectors through the BERT model and key entities in the corresponding data can be extracted, such as batch account opening, fulfillment node ID, abnormal coding, etc. The key entities extracted from the abnormal order scenario label data are used as data points, and the order dimension and chunk dimension are used as the initial clustering centers respectively. The K-means clustering algorithm is used to perform cluster analysis on the key entities extracted from historical repair orders, operation and maintenance summaries and other empirical data. After multiple clustering iterative calculations, the optimal key entity is obtained. According to the reverse correlation of the key entity with the corresponding historical repair order data, operation and maintenance summary data, repair order strategy and other data, the addition of abnormal order scenarios is realized.

[0138] For unknown abnormal scenarios that do not exist in the identified rule base, unsupervised learning based on kmeans is used for abnormal identification and classification. The abnormality is the result of AI modeling calculation, and business personnel are required to judge whether it belongs to a new abnormal scenario and determine its abnormal scenario label and processing plan.

[0139] If it is a new abnormal scenario: update the abnormal rule library and BERT model, and add a new abnormal scenario. If it is not a new abnormal scenario: determine it to be normal, and handle it according to the normal order scenario.

[0140] During the order repair process, it is necessary to automatically create an order repair task to regularly scan the abnormal order repair file directory. If a repair file is found, it will automatically pull up the creation of an automatic order repair task; then create an automatic order repair task to load and parse the repair file and create an order repair task; then execute the order repair task call according to the created order repair task, generate the corresponding detailed task according to the original file, and modify the main order repair task to the execution state, and pull up the order repair subtask; then the order repair execution subtask modifies the detailed task to being executed, and executes the order repair. The overall order intelligent repair process includes the following modules:

[0141] 1) Data filling

[0142] The data flow diagram of data filling is as follows Figure 4As shown. Due to a series of special situations such as the interface needs to be initialized before calling, there is a correlation between the parameters of the message class interface, the parameters of some interfaces cannot be constructed directly, and they need to rely on the parameter construction method, etc., the business scenario can be used as the driver, and the global information supplement tool for batch orders can be implemented based on the MTV framework to implement the supplementary information template definition filling according to the business scenario. Specifically, the parser can be used to uniformly parse and output the filling request and the filling group return information; the template suite can be used to implement scenario-based parameter template definition; the scheduler can be used to implement process scheduling and public capability scheduling, and unified scheduling and control capabilities such as function-level template control, parameter name template control, and type template control can be implemented; the built-in code generator is implemented through the data processing plug-in to realize data acquisition logic code generation; the private constructor is used to call public or private methods in the business scenario template to obtain code blocks, so as to realize the parameter content completion of business scenarios.

[0143] 2) Dynamic data verification

[0144] The business acceptance scenarios of IoT are complex. Basic businesses involve hundreds of business data, and complex business data often involve thousands of data parameters. For the data filling parameters of the repair task, the data lineage capability is applied to realize data link association according to the order fulfillment process, track the data processing process, and ensure the compliance verification of data and business. Specifically, it can realize the lineage analysis between abnormal orders, identify the business scope, application name, data source, processing component, data flow direction and other information, and calculate the lineage relationship at the same time; obtain the rule name, rule type, and rule content information of the order fulfillment node through the interface, and verify the data rules according to the application name and processing component calculated by the lineage relationship, to ensure the accuracy of business data and realize the automatic verification of orders.

[0145] 3) Repair order

[0146] Create a list of executed actions. Before executing each action, set the action ID to the list and append it to the action track in the task list to track the action execution track in case of an exception. Finally, execute the matched repair strategy.

[0147] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the abnormal order repair method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0148] Please refer to Figure 5 The present application also proposes an abnormal order repair system, the abnormal order repair system comprising:

[0149] The order screening module 10 is used to input the order data into a preset abnormal order classification model to obtain each abnormal order;

[0150] A rule matching module 20, used for taking the abnormal orders that do not match the preset abnormal order rule base among the abnormal orders as abnormal orders to be classified;

[0151] A cluster analysis module 30, used to classify each of the abnormal orders to be classified by a clustering algorithm to obtain each abnormal category;

[0152] The order repair module 40 is used to automatically repair each abnormal order according to the repair solutions in the abnormal order rule base and the repair solutions corresponding to each abnormal category.

[0153] In one embodiment, the order repair module 40 is further configured to:

[0154] Determine a target abnormal order among the abnormal orders, and obtain an order execution table of the target abnormal order, wherein the order execution table is an order context table when multiple orders are executed in sequence to implement business logic;

[0155] Taking the abnormal orders belonging to the order execution table among the abnormal orders as the current pending orders, and sorting the current pending orders based on the execution order of the current pending orders in the order execution table to obtain a pending order sequence;

[0156] The first current order to be processed in the sequence of pending orders is taken as an initial abnormal order, and starting from the initial abnormal order, each current order to be processed in the sequence of pending orders is automatically repaired according to the repair solutions in the abnormal order rule base and the repair solutions corresponding to each abnormal category;

[0157] Delete each of the currently pending orders from among the abnormal orders, and return to execute the step of determining the target abnormal order from among the abnormal orders until all of the abnormal orders are repaired.

[0158] In one embodiment, the order repair module 40 is further configured to:

[0159] Randomly determine a target abnormal order from among the abnormal orders;

[0160] The priorities corresponding to the respective abnormal orders are obtained, and the abnormal order corresponding to the maximum priority among the priorities is taken as the target abnormal order.

[0161] In one embodiment, the order repair module 40 is further configured to:

[0162] Taking the initial abnormal order as the traversal starting point, traverse the sequence of orders to be processed;

[0163] Obtain the order processing stage corresponding to the currently pending order being traversed, and fill in the preset global variable template corresponding to the order processing stage;

[0164] Determine whether the currently pending order being traversed belongs to an abnormal order to be classified;

[0165] If it is detected that the traversed current pending order belongs to an abnormal order to be classified, the current pending order is automatically repaired based on the repair solution of the abnormal category to which the current pending order belongs and the filled global variable template;

[0166] If it is detected that the current pending order being traversed does not belong to the abnormal order to be classified, the current pending order is automatically repaired based on the repair plan in the abnormal order rule library and the filled global variable template.

[0167] In one embodiment, the abnormal order repair system further includes:

[0168] A pre-training module is used to pre-train a preset neural network model based on a common data set;

[0169] A splicing module, used to splice the neural network model with a preset order classification model to obtain an initial classification model after detecting that the pre-training is completed;

[0170] A training module, used to train the initial classification model using historical order data as model input data and using a binary classification label of whether each order in the historical order data is abnormal as a model training label;

[0171] The model acquisition module is used to use the trained initial classification model as an abnormal order classification model.

[0172] In one embodiment, the abnormal order repair system further includes:

[0173] A first updating module, configured to update the abnormal order rule base according to each of the abnormal categories and the repair solutions corresponding to each of the abnormal categories;

[0174] The second updating module is used to update the abnormal order classification model based on the updated abnormal order rule base.

[0175] In one embodiment, the cluster analysis module 30 is further used to:

[0176] Extracting the key entities corresponding to each of the abnormal orders to be classified from the actual scene labels corresponding to each of the abnormal orders to be classified, wherein the key entities are information units that characterize scene features;

[0177] Each of the key entities is used as a data point, and each of the data points is classified by a clustering algorithm to obtain various abnormal categories.

[0178] The abnormal order repair system provided in the embodiment of the present application adopts the abnormal order repair method in the above embodiment, which can solve the technical problem of how to reduce the repair cost of abnormal orders. Compared with the prior art, the beneficial effects of the abnormal order repair system provided in the embodiment of the present application are the same as the beneficial effects of the abnormal order repair method provided in the above embodiment, and other technical features of the abnormal order repair system are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0179] An embodiment of the present application provides an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the abnormal order repair method in the above-mentioned embodiment one.

[0180] Reference below Figure 6 , which shows a schematic diagram of the structure of an electronic device suitable for implementing an embodiment of the present application. Figure 6 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0181] like Figure 6As shown, the electronic device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read On ly Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the electronic device are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows an electronic device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0182] In particular, according to the embodiments disclosed in the embodiments of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the embodiments of the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through a communication device, or installed from the storage device 1003, or installed from the ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the embodiment of the present application are executed.

[0183] The electronic device provided in the embodiment of the present application adopts the abnormal order repair method in the above embodiment, which can solve the technical problem of how to reduce the repair cost of abnormal orders. Compared with the prior art, the beneficial effects of the electronic device provided in the embodiment of the present application are the same as the beneficial effects of the abnormal order repair method provided in the above embodiment, and the other technical features in the electronic device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0184] It should be understood that the various parts disclosed in the embodiments of the present application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0185] The above is only a specific implementation of the embodiment of the present application, but the protection scope of the embodiment of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or replacements within the technical scope disclosed in the embodiment of the present application, which should be included in the protection scope of the embodiment of the present application. Therefore, the protection scope of the embodiment of the present application should be based on the protection scope of the claims.

[0186] An embodiment of the present application provides a computer-readable storage medium having computer-readable program instructions (ie, a computer program) stored thereon, wherein the computer-readable program instructions are used to execute the abnormal order repair method in the above-mentioned embodiment.

[0187] The computer-readable storage medium provided in the embodiment of the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read On ly Memory), an erasable programmable read-only memory (EPROM: Erasab le Programmab le Read On ly Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read On ly Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0188] The computer-readable storage medium may be included in the electronic device, or may exist independently without being installed in the electronic device.

[0189] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by an electronic device, the electronic device: inputs order data into a preset abnormal order classification model to obtain each abnormal order; takes the abnormal orders among the abnormal orders that do not match the preset abnormal order rule library as abnormal orders to be classified; classifies each abnormal order to be classified through a clustering algorithm to obtain each abnormal category; and automatically repairs each abnormal order according to the repair solutions in the abnormal order rule library and the repair solutions corresponding to each abnormal category.

[0190] The computer program code for performing the operation of the embodiment of the present application can be written in one or more programming languages ​​or a combination thereof, and the above-mentioned programming languages ​​include object-oriented programming languages, such as Java, Smalltalk, C++, and also include conventional procedural programming languages, such as "C" language or similar programming languages. The program code can be executed completely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on the remote computer, or completely on the remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., using an Internet service provider to connect through the Internet).

[0191] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the function marked in the box can also occur in a different order from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0192] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0193] The readable storage medium provided in the embodiment of the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned abnormal order repair method, and can solve the technical problem of how to reduce the repair cost of abnormal orders. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in the embodiment of the present application are the same as the beneficial effects of the abnormal order repair method provided in the above-mentioned embodiment, and will not be repeated here.

[0194] An embodiment of the present application also provides a computer program product, including a computer program, which implements the steps of the abnormal order repair method as described above when executed by a processor.

[0195] The computer program product provided in the embodiment of the present application can solve the technical problem of how to reduce the repair cost of abnormal orders. Compared with the prior art, the beneficial effects of the computer program product provided in the embodiment of the present application are the same as the beneficial effects of the abnormal order repair method provided in the above embodiment, which will not be repeated here.

[0196] The above descriptions are only some embodiments of the embodiments of the present application, and do not limit the patent scope of the embodiments of the present application. All equivalent structural transformations made by using the description and drawings of the embodiments of the present application under the technical concept of the embodiments of the present application, or directly / indirectly applied in other related technical fields are included in the patent protection scope of the embodiments of the present application.

Claims

1. A method for repairing abnormal orders, characterized in that: The abnormal order repair method comprises: Input the order data into the preset abnormal order classification model to obtain each abnormal order; The abnormal orders that do not match the preset abnormal order rule base among the abnormal orders are regarded as abnormal orders to be classified; Classifying each of the abnormal orders to be classified by a clustering algorithm to obtain each abnormal category; Each of the abnormal orders is automatically repaired according to the repair solutions in the abnormal order rule base and the repair solutions corresponding to each of the abnormal categories.

2. The abnormal order repair method according to claim 1, characterized in that: The solution for automatically repairing each abnormal order according to the repair solution in the abnormal order rule base and the repair solution corresponding to each abnormal category includes: Determine a target abnormal order among the abnormal orders, and obtain an order execution table of the target abnormal order, wherein the order execution table is an order context table when multiple orders are executed in sequence to implement business logic; Taking the abnormal orders belonging to the order execution table among the abnormal orders as the current pending orders, and sorting the current pending orders based on the execution order of the current pending orders in the order execution table to obtain a pending order sequence; The first current order to be processed in the sequence of pending orders is taken as an initial abnormal order, and starting from the initial abnormal order, each current order to be processed in the sequence of pending orders is automatically repaired according to the repair solutions in the abnormal order rule base and the repair solutions corresponding to each abnormal category; Delete each of the currently pending orders from among the abnormal orders, and return to execute the step of determining the target abnormal order from among the abnormal orders until all of the abnormal orders are repaired.

3. The abnormal order repair method according to claim 2, characterized in that: The step of determining a target abnormal order from among the abnormal orders comprises: Randomly determine a target abnormal order from among the abnormal orders; or, The priorities corresponding to the respective abnormal orders are obtained, and the abnormal order corresponding to the maximum priority among the priorities is taken as the target abnormal order.

4. The abnormal order repair method according to claim 2, characterized in that: The step of automatically repairing each of the current pending orders in the pending order sequence based on the repair solutions in the abnormal order rule base and the repair solutions corresponding to each of the abnormal categories, taking the initial abnormal order as the starting point, comprises: Taking the initial abnormal order as the traversal starting point, traverse the sequence of orders to be processed; Obtain the order processing stage corresponding to the current pending order being traversed, and fill in the preset global variable template corresponding to the order processing stage; Determine whether the currently pending order being traversed belongs to an abnormal order to be classified; If it is detected that the traversed current pending order belongs to an abnormal order to be classified, the current pending order is automatically repaired based on the repair solution of the abnormal category to which the current pending order belongs and the filled global variable template; If it is detected that the current pending order being traversed does not belong to the abnormal order to be classified, the current pending order is automatically repaired based on the repair plan in the abnormal order rule library and the filled global variable template.

5. The abnormal order repair method according to claim 1, characterized in that: Before the step of inputting the order data into a preset abnormal order classification model to obtain each abnormal order, the method further includes: Pre-training of preset neural network models based on common data sets; After detecting that the pre-training is completed, the neural network model is spliced ​​with a preset order classification model to obtain an initial classification model; The historical order data is used as model input data, and the binary classification labels of whether each order in the historical order data is abnormal are used as model training labels to train the initial classification model; The trained initial classification model is used as an abnormal order classification model.

6. The abnormal order repair method according to claim 1, characterized in that: After the step of automatically repairing each of the abnormal orders according to the repair solutions in the abnormal order rule base and the repair solutions corresponding to each of the abnormal categories, the method further includes: According to each of the abnormal categories and the corresponding repair solutions of each of the abnormal categories, the abnormal order rule library is updated; Based on the updated abnormal order rule base, the abnormal order classification model is updated.

7. The abnormal order repair method according to claim 1, characterized in that: The step of classifying each abnormal order to be classified by a clustering algorithm to obtain each abnormal category includes: Extracting the key entities corresponding to each of the abnormal orders to be classified from the actual scene labels corresponding to each of the abnormal orders to be classified, wherein the key entities are information units that characterize scene features; Each of the key entities is used as a data point, and each of the data points is classified by a clustering algorithm to obtain various abnormal categories.

8. An abnormal order repair system, characterized in that: The abnormal order repair system comprises: The order screening module is used to input the order data into the preset abnormal order classification model to obtain each abnormal order; A rule matching module, used for treating the abnormal orders that do not match the preset abnormal order rule library among the abnormal orders as abnormal orders to be classified; A cluster analysis module, used to classify each of the abnormal orders to be classified by a clustering algorithm to obtain each abnormal category; The order repair module is used to automatically repair each abnormal order according to the repair solution in the abnormal order rule base and the repair solution corresponding to each abnormal category.

9. An electronic device, characterized in that: The electronic device comprises: a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the abnormal order repair method as claimed in any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the abnormal order repair method as described in any one of claims 1 to 7 are implemented.

11. 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 steps of the abnormal order repairing method as claimed in any one of claims 1 to 7 are implemented.

Images