Work order processing system based on machine learning

Abstract

One or more embodiments of the invention provide a work order processing system based on machine learning. The system comprises a data layer used for collecting operation and maintenance data of a target system; the intelligent layer is used for obtaining operation and maintenance data features and work order features according to the operation and maintenance data, the to-be-executed work order and a feature extraction model, and predicting the priority of the to-be-executed work order according to the operation and maintenance data features, the work order features and the work order processing model, and the work order processing model is trained according to the executed work order; and the application layer is used for updating the state of the to-be-executed work order according to the priority of the to-be-executed work order. According to the invention, intelligent analysis and processing of the work order are realized through the intelligent layer, the accuracy of judgment of the emergency degree of the work order is improved, resource allocation is optimized, and the operation and maintenance efficiency and the service quality of the system are further improved.

Description

Technical Field

[0001] One or more embodiments of the present application relate to the field of computer technology, and in particular, to a work order processing system based on machine learning. Background Art

[0002] In the current information age, operation and maintenance management is crucial to ensure the stable operation of various systems. The operation and maintenance work order processing system of related technologies mainly relies on manual judgment and decision-making. This processing method has many defects when facing complex and changing operation and maintenance needs. For example, the manual work order processing method makes it difficult for operation and maintenance personnel to quickly and accurately judge the urgency and priority of work orders when facing a large number of work orders, resulting in unreasonable resource allocation and affecting operation and maintenance efficiency and service quality. Summary of the invention

[0003] In view of this, an object of one or more embodiments of the present application is to propose a work order processing system based on machine learning to solve the problems raised by the background technology.

[0004] Based on the above objectives, one or more embodiments of the present application provide a work order processing system based on machine learning. It includes:

[0005] The data layer is used to collect the operation and maintenance data of the target system;

[0006] An intelligent layer, configured to obtain operation and maintenance data features and work order features based on the operation and maintenance data, the work orders to be executed and a feature extraction model, and to predict the priority of the work orders to be executed based on the operation and maintenance data features, the work order features and the work order processing model, wherein the work order processing model is trained based on executed work orders;

[0007] The application layer is used to update the status of the work order to be executed according to the priority of the work order to be executed.

[0008] Optionally, the application layer is further used to obtain user feedback on the work order to be executed, and send the feedback to the intelligent layer; the feedback indicates whether the priority of the work order to be executed is reasonable.

[0009] Optionally, the work order processing model is trained based on the operation and maintenance data, executed work orders and the feedback.

[0010] Optionally, the training of the feature extraction model and the work order processing model includes iteratively executing the following steps until the model loss function value converges:

[0011] Determine the actual priority of the executed work order according to the feedback information;

[0012] Obtaining training work order features and training operation and maintenance data features according to the executed work order, the training operation and maintenance data, and the first machine learning model;

[0013] Establishing a training sample according to the training work order characteristics, the training operation and maintenance data characteristics, and the actual priority of the executed work order;

[0014] Inputting the training work order features and the training operation and maintenance data features into a second machine learning model to obtain a training result;

[0015] Adjust the parameters of the first machine learning model and the second machine learning model according to the training results and the loss function value of the actual priority of the executed work order.

[0016] Optionally, the operation and maintenance data is obtained from different data sources, and the data sources include one or more of server logs and monitoring systems.

[0017] Optionally, it also includes:

[0018] The data fusion layer is used to fuse the operation and maintenance data from different data sources and send the fused operation and maintenance data to the intelligent layer.

[0019] Optionally, the data fusion layer is also used to perform quality assessment on the processed operation and maintenance data, and send the processed operation and maintenance data with a quality score higher than a preset threshold to the intelligent layer.

[0020] Optionally, the application layer is further used to receive instructions from users, and create, edit or delete the to-be-executed work order according to the instructions.

[0021] Optionally, the application layer is also used to store the work orders to be executed.

[0022] Optionally, the display layer is used to display the work orders to be executed according to the user's screening conditions.

[0023] From the above, it can be seen that the work order processing system based on machine learning provided by one or more embodiments of the present application realizes the intelligent upgrade of the work order processing system through the introduction of the intelligent layer. Specifically, the use of the work order processing model can predict the urgency or priority of the operation and maintenance work order, helping the operation and maintenance team to allocate resources more effectively, which is particularly important when facing a large number of work orders. It can ensure that urgent work orders are processed in a timely manner, thereby improving the system operation and maintenance efficiency and service quality. Due to the use of machine learning technology, users can also enjoy more convenient, efficient and personalized services. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate one or more 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, the drawings described below are only one or more embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0025] Figure 1 A schematic diagram of the structure of a work order processing system based on machine learning according to one or more embodiments of the present application;

[0026] Figure 2 A schematic diagram of the structure of a work order processing system based on machine learning according to one or more embodiments of the present application;

[0027] Figure 3 A schematic diagram of the structure of a work order processing system based on machine learning according to one or more embodiments of the present application;

[0028] Figure 4 A schematic diagram of the structure of a work order processing system based on machine learning according to one or more embodiments of the present application. DETAILED DESCRIPTION

[0029] In order to make the objectives, technical solutions and advantages of the present disclosure more clearly understood, the present disclosure is further described in detail below in combination with specific embodiments and with reference to the accompanying drawings.

[0030] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in one or more embodiments of the present application should be understood by people with ordinary skills in the field to which the present disclosure belongs. The "first", "second" and similar words used in one or more embodiments of the present application do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0031] like Figure 1 As shown, the work order processing system based on machine learning of one or more embodiments of the present application includes:

[0032] Data layer 11, used to collect operation and maintenance data of the target system;

[0033] The intelligent layer 12 is used to obtain operation and maintenance data features and work order features based on the operation and maintenance data, the work orders to be executed and the feature extraction model, and predict the priority of the work orders to be executed based on the operation and maintenance data features, the work order features and the work order processing model, and the work order processing model is trained based on the executed work orders;

[0034] The application layer 13 is used to update the status of the above-mentioned work order to be executed according to the priority of the above-mentioned work order to be executed.

[0035] The embodiment of the present application introduces the above-mentioned intelligent layer 12 to realize the intelligent upgrade of the work order processing system. The feature extraction model and the work order processing model can more accurately predict the urgency or priority of the operation and maintenance work order; can help the operation and maintenance team to allocate resources more effectively, especially when faced with a large number of work orders to be executed, can ensure that urgent work orders are processed in a timely manner, thereby improving the operation and maintenance efficiency and service quality of the target system. While improving the user experience, it can also further reduce the workload of the operation and maintenance personnel.

[0036] The target layer in the embodiment of the present application is used to collect the operation and maintenance data of the target system. Among them, the operation and maintenance data (Operational Data) refers to the data generated and collected during the operation and maintenance (operation and maintenance) of the network system. These data can be used to monitor the operating status of the system, analyze performance, troubleshoot, optimize resource usage, prevent potential problems and realize automated operation and maintenance. These data may include one or more of log data, performance data, resource data, event and alarm data, user behavior data and backup and recovery data. As long as the selection of operation and maintenance data can achieve the corresponding purpose, different selections will not affect the protection scope of the present invention.

[0037] In the embodiments of the present application, the above operation and maintenance data may come from different data sources. The data source may include one or more of a server log, a monitoring system, and a user feedback system. The data source may vary depending on the type of operation and maintenance data in the actual scenario. This application does not limit the type of operation and maintenance data, nor does it limit the type of data source.

[0038] The above operation and maintenance data can be collected through pre-configured data interfaces such as API, database connection, file transfer, etc.

[0039] In an embodiment of the present application, the data layer 11 is also used to store the operation and maintenance data, and / or perform data cleaning on the collected operation and maintenance data.

[0040] Specifically, distributed storage technology can be used to implement data storage to ensure high availability and scalability of data. Data cleaning can include one or more of data deduplication, format conversion, and data normalization to provide high-quality data for subsequent prediction work. This application does not limit the selection of data storage technology and data cleaning technology.

[0041] The intelligent layer 12 of the present application deploys at least two modules: a feature extraction model and a work order processing model. Among them, the feature extraction model can be used to extract the features of the above-mentioned operation and maintenance data and the work orders to be executed; the work order processing model can predict the priority of the work order to be executed based on the above-mentioned operation and maintenance data and the features of the work orders to be executed. The features extracted by the feature extraction model are all key features that have a significant impact on the urgency or priority prediction of the work order. The feature extraction model can reduce the dimension of the data and improve the training efficiency and accuracy of the model. The above features represent the key factors that affect the urgency of the work order, such as fault type, impact range, historical processing time, etc. The fault type reflects the specific problem involved in the work order, and different fault types may correspond to different urgency levels; the impact range indicates the number of systems or users affected by the fault, and the wider the range, the higher the urgency; and the occurrence time takes into account the time point when the fault occurs. Faults in certain time periods may be more urgent, such as business peak periods.

[0042] The above feature extraction model can be obtained based on natural language processing technology. Natural language processing (NLP) technology is used to parse and understand the content of the work order, so as to automatically extract key information and assess its urgency. NLP technology can identify entities, emotions and intentions in the text, and provide richer contextual information for work order processing. The above work order processing model can be obtained based on supervised learning algorithms, such as random forests, support vector machines and other deep learning models.

[0043] The feature extraction model can also use a logistic regression algorithm to train the selected features. Logistic regression is a widely used classification algorithm that can establish a mapping relationship between features and target variables (the urgency or priority of the work order in this invention) by learning the distribution law of data. During the training process, the model training submodule will use historical operation and maintenance work order data as a training set, and by continuously adjusting the model parameters, the model can more accurately predict the urgency or priority of new work orders. At the same time, in order to avoid overfitting and underfitting, this submodule will also use cross-validation and other techniques to evaluate and optimize the model.

[0044] According to the obtained priority of the work order to be executed, a recommended processing solution for the work order to be executed, such as an estimated processing time, etc., can be further obtained.

[0045] The application layer 13 of the embodiment of the present application can not only update the status of the work order to be executed according to the priority of the work order to be executed, but also obtain user feedback on the work order to be executed and send the feedback to the intelligent layer 12; the feedback indicates whether the priority of the work order to be executed is reasonable.

[0046] The feature extraction model and work order processing model of the above-mentioned intelligent layer can be trained according to the above-mentioned feedback. Specifically, model training may include iteratively executing the following steps until the model loss function value converges: determining the actual priority of the executed work order according to the feedback information; obtaining the training work order features and the training operation and maintenance data features according to the executed work order, the training operation and maintenance data and the first machine learning model; establishing training samples according to the training work order features, the training operation and maintenance data features and the actual priority of the executed work order; inputting the training work order features and the training operation and maintenance data features into the second machine learning model to obtain training results; adjusting the parameters of the first machine learning model and the second machine learning model according to the training results and the loss function value of the actual priority of the executed work order.

[0047] It can be understood that the present application can update the above two models in real time or periodically according to the acquired operation and maintenance data and feedback to improve the performance of the feature extraction model and work order processing model of the present application.

[0048] In the process of implementing this application, the applicant found that the related technologies still have the problem of insufficient data fusion and processing capabilities when processing work orders. Specifically, the related technologies are often unable to effectively fuse data from different data sources to form a consistent data view. This causes operation and maintenance personnel to spend a lot of time and energy on data integration and analysis when facing multi-source data, which reduces work efficiency.

[0049] Therefore, in the embodiments of the present application, Figure 2 As shown, a data fusion layer 14 may also be included. The data fusion layer 14 is used to fuse the operation and maintenance data. As described above, the operation and maintenance data may be obtained from different data sources. When the data sources are different, the operation and maintenance data from different data sources may be fused through the data fusion layer 14, and then the fused operation and maintenance data may be sent to the intelligent layer.

[0050] In addition, due to the lack of an effective data quality assessment mechanism in the related technology, the quality of the fused data is difficult to guarantee, which affects the accuracy of the work order processing. Therefore, in the embodiment of the present application, the data fusion layer 14 can also be used to perform quality assessment on the processed operation and maintenance data, and send the processed operation and maintenance data with a quality score higher than the preset threshold to the intelligent layer. The quality assessment criteria can be set according to the completeness, accuracy, consistency and timeliness of the data. This function can filter the data and only send high-quality data to the intelligent layer as a prediction basis to improve the basis for priority prediction.

[0051] In the embodiments of the present application, a variety of testing methods including white box testing, black box testing and gray box testing may be used to comprehensively evaluate the data quality in the operation and maintenance work order system.

[0052] White box testing focuses on in-depth inspection of data processing logic and internal structure. In the data quality assessment module, white box testing is mainly used to verify whether the implementation of sub-modules such as data cleaning, conversion and normalization is correct, and whether the internal logic meets expectations. Black box testing regards the data processing logic as an invisible "black box" and only focuses on the correctness of input and output. In the data quality assessment module, black box testing is used to evaluate whether the results after data cleaning, conversion and normalization meet business requirements and quality standards. Gray box testing is between white box testing and black box testing, focusing on both the correctness of input and output and the implementation of some internal logic. In the data quality assessment module, gray box testing is used to evaluate the correctness of key paths and interaction points in the process of data cleaning, conversion and normalization.

[0053] White box testing can be implemented by:

[0054] Code review: Professional testers review the code of the data cleaning, conversion, and normalization sub-modules line by line to check whether the code implementation follows the predetermined logic and data processing rules.

[0055] Unit testing: Write test cases for each independent function or method in the data cleaning, transformation, and normalization submodules to ensure that the output of the function is consistent with the expected result under the given input. For example, for a function that fills missing values ​​with the mean, you can construct a test case containing missing values ​​to verify whether the filling result is correct.

[0056] Coverage analysis: Use code coverage tools to analyze the coverage of test cases on data processing logic to ensure that all critical paths and branches are tested.

[0057] Black box testing can be implemented by:

[0058] Equivalence class division: Divide the input data into different equivalence classes. The data in each equivalence class should have the same behavior in the data processing logic. Then select representative values ​​for each equivalence class for testing to verify the correctness of the output results.

[0059] Boundary value analysis: Test the boundary values ​​of input and output, because many errors often occur at the boundaries. For example, for data normalization operations, you can test the input data near the minimum and maximum values ​​to verify whether the normalization results are within the expected range.

[0060] Random testing: Randomly generate input data and test it to find possible anomalies or uncovered paths. Random testing helps increase the diversity and comprehensiveness of testing.

[0061] Grey box testing can be implemented by:

[0062] Interface testing: Test whether the interfaces between data cleaning, transformation, and normalization submodules correctly pass data and control flow. For example, test cases can be constructed to verify whether data remains complete and accurate when it is passed from one submodule to another.

[0063] State testing: Evaluate whether the state changes of data during processing are in line with expectations. For example, during data cleaning, you can check whether missing values ​​are correctly filled and whether erroneous data are corrected.

[0064] Log analysis: Verify the execution of internal logic by analyzing the log information during the processing. The log should record the results and abnormal conditions of key operations so that testers can verify and debug.

[0065] In the embodiments of the present application, Figure 3 As shown, a display layer 15 may also be included. The display layer 15 may be used to display the to-be-executed work orders according to the user's screening conditions.

[0066] The work order processing system based on machine learning in one or more embodiments of the present application has a structure as follows: Figure 4 shown.

[0067] Specifically, the system includes a data layer 11 , an intelligent layer 12 , an application layer 13 , a data fusion layer 14 and a presentation layer 15 .

[0068] The data layer 11 may include a data acquisition module 111, a data storage module 112 and a data management module 113. The data acquisition module 111 is used to collect operation and maintenance data from at least one preset data source in real time or at a fixed time; the data storage module 112 is used to store the above operation and maintenance data; the data management module 113 is used to clean, convert the format and normalize the above operation and maintenance data.

[0069] In some embodiments, the data management module 113 may further include a data cleaning submodule 1131 , a data conversion submodule 1132 , and a data normalization submodule 1133 .

[0070] The data cleaning submodule 1131 is used to clean the collected operation and maintenance data. This submodule uses deduplication technology to delete duplicate data to ensure the uniqueness of the data set. For missing values, the mean filling method is used to fill them, that is, the average of non-missing values ​​in the data set is calculated and the missing values ​​are replaced with the average to maintain the integrity of the data. For erroneous data, the data items that do not meet the standards are identified and corrected through a preset rule matching method (such as a regular expression or a data validation rule).

[0071] The data conversion submodule 1132 is used to convert data in different formats into a unified format, including data type conversion, such as converting string type data into numeric type for mathematical operations, and data format conversion, such as converting date and time formats from one standard to another to ensure data consistency and comparability.

[0072] The data normalization submodule 1133 uses a minimum-maximum normalization algorithm to normalize the data. The algorithm is implemented by the following formula:

[0073]

[0074] Among them, X is the original data, X min is the minimum value in the data set, X max is the maximum value in the data set, and X′ is the normalized data. Through this process, data of different magnitudes are mapped to the same numerical range, making them comparable, which facilitates subsequent data analysis and processing.

[0075] The data fusion layer 14 may include a fusion module 141 and a data quality assessment module 142. The fusion module 141 is used for data from different data sources; the data quality assessment module 142 is used for performing quality assessment on the fused data.

[0076] The fusion module 141 may further include a data matching submodule 1411 , a data merging submodule 1412 , and a data conflict resolution submodule 1413 .

[0077] The data matching submodule 1411 is used to match the operation and maintenance data from different data sources to identify the same or related data items. This submodule uses two matching algorithms: rule-based matching and similarity-based matching. Rule-based matching uses a preset set of rules, such as data format, the value of a specific field, etc., to filter and match data. Similarity-based matching uses the cosine similarity algorithm, and its calculation formula is:

[0078]

[0079] Among them, A and B are two data vectors to be matched, and Sim(A, B) is their cosine similarity value; the algorithm evaluates their similarity by calculating the cosine value of the angle between the two vectors. The closer the value is to 1, the higher the similarity.

[0080] The data merging submodule 1412 is used to merge the successfully matched data items to form a consistent data view. During the merging process, for numerical data, the weighted average method is used for calculation, that is, the corresponding values ​​are weighted and summed according to the credibility or weight of each data source, so as to obtain a more accurate merging result. For non-numerical data, such as string or enumeration type data, the priority method is used for selection, that is, according to the preset priority order, the data item with the highest credibility or the most consistent with the business logic is selected as the merging result.

[0081] The data conflict resolution submodule 1413 is used to handle conflicts that occur during the data merging process. These conflicts may include inconsistent data values ​​or missing data items. For inconsistent data values, this submodule will determine the final merged value based on preset conflict resolution rules, such as "latest value first", "high-trust data source first", etc. For missing data items, they will be inferred and filled based on the values ​​of other related data items or business logic to ensure the integrity and consistency of the data.

[0082] The intelligent layer 12 may include a feature extraction module 121 and a work order processing module 122. The feature extraction module 121 is used to extract operation and maintenance data features and work order features based on the operation and maintenance data and the work orders to be executed; the work order processing module 122 is used to predict the priority of the work orders to be executed based on the above operation and maintenance data features and work order features.

[0083] The application layer 13 may include a work order management module 131, a knowledge base management module 132 and a user feedback module 133. The work order management module 131 is used to create, edit, delete, query, assign, track and update the status of work orders according to user instructions; the knowledge base management module 132 is used to query and retrieve knowledge base data according to user instructions; and the user feedback module 133 is used to receive the above feedback.

[0084] The data stored in the knowledge base includes operation and maintenance related knowledge, cases and solutions, which can help operation and maintenance personnel solve problems quickly.

[0085] In some embodiments, the knowledge base management module 132 may include a knowledge acquisition submodule 1321 , a knowledge storage submodule 1322 , a knowledge retrieval submodule 1323 and a knowledge update submodule 1324 .

[0086] The knowledge acquisition submodule 1321 is used to collect operation and maintenance related knowledge from various sources. These sources include but are not limited to operation and maintenance manuals, technical documents, expert experience and historical work order records. In order to achieve effective knowledge acquisition, this submodule can use the following technical means:

[0087] Automatic crawling: Using crawler technology or API interfaces, automatically crawling operation and maintenance related knowledge from designated document libraries, websites or databases. This method can efficiently collect a large amount of structured or semi-structured data.

[0088] Manual input: Provides a user-friendly interface that allows operators to manually input or upload operation and maintenance knowledge. This method is suitable for unstructured data or situations that require manual review.

[0089] Interface import: Import existing operation and maintenance knowledge through interfaces with other systems or applications. This method can realize knowledge sharing and reuse and improve knowledge utilization.

[0090] The acquired knowledge is organized and stored in a knowledge base in a structured or semi-structured format to facilitate subsequent retrieval and utilization.

[0091] The knowledge storage submodule 1322 is used to store the acquired knowledge and build an operation and maintenance knowledge base. Considering the diversity and complexity of knowledge, this submodule can adopt a storage method that combines a database and a file system:

[0092] Database storage: For structured data, such as chapters, paragraphs, and tables in the operation and maintenance manual, a relational database is used for storage. The database provides powerful query and indexing functions to quickly retrieve and locate the required knowledge.

[0093] File system storage: For unstructured data, such as technical documents, pictures, and videos, the file system is used for storage. The file system provides flexible storage methods and large storage space, and can accommodate files of various types and sizes.

[0094] Through the combined use of database and file system, the knowledge storage submodule can achieve comprehensive coverage and efficient storage of operation and maintenance knowledge.

[0095] The knowledge search submodule 1323 is used to provide a knowledge search function to help operation and maintenance personnel quickly find the required knowledge. This submodule supports multiple search methods, including keyword search, category browsing, and fuzzy matching.

[0096] Keyword search: allows operators to enter keywords for search. The system will return relevant knowledge items based on the frequency and position of the keywords in the knowledge base.

[0097] Category browsing: According to the classification system of operation and maintenance knowledge, a hierarchical browsing method is provided. Operation and maintenance personnel can gradually narrow the search scope by clicking on the category label until they find the required knowledge.

[0098] Fuzzy matching: When the input is imprecise or contains spelling errors, the system uses a fuzzy matching algorithm for retrieval. This algorithm can identify similar words and phrases and return knowledge items that are close to the input.

[0099] By combining multiple retrieval methods, the knowledge retrieval submodule can improve the work efficiency of operation and maintenance personnel and reduce the time cost of searching for knowledge.

[0100] The knowledge updating submodule 1324 is used to update and maintain the knowledge base regularly or as needed, which may include operations such as adding new knowledge, correcting erroneous knowledge, and deleting obsolete knowledge.

[0101] Adding new knowledge: When operation and maintenance personnel obtain new operation and maintenance knowledge, they can add it to the knowledge base through the knowledge update submodule. The system will automatically classify and index the new knowledge for subsequent retrieval and utilization.

[0102] Correcting incorrect knowledge: For errors or inaccurate information in the knowledge base, operation and maintenance personnel can correct them through the knowledge update submodule. The system will record the correction history to track the changes in knowledge.

[0103] Delete outdated knowledge: As the operation and maintenance environment changes and technology is updated, some knowledge may become outdated or no longer applicable. The knowledge update submodule allows operation and maintenance personnel to delete this knowledge to maintain the timeliness and accuracy of the knowledge base.

[0104] By using the knowledge update submodule, operation and maintenance personnel can ensure the continuous updating and improvement of the knowledge base and improve the quality and efficiency of operation and maintenance work.

[0105] The display layer 15 may include an interface display module 151 and a data display module 152. The interface display module 151 is used to display a preset user operation interface; the data display module 152 is used to display work order data and knowledge base data. The data display module 152 can use tools such as charts and dashboards to provide a visual display function of data, such as work order processing efficiency, fault distribution, resource usage, etc., to help the operation and maintenance team better understand the operation and maintenance status and make more informed decisions.

[0106] Among them, the interface display module 151 can adopt the B / S (Browser / Server) technical architecture, and select the React framework and the Redux state management library for development. The React framework provides an efficient component-based development method that can quickly build a user interface. The Redux state management library is used to manage the state and data flow of the application to ensure the responsiveness and consistency of the interface.

[0107] At the same time, the interface display module 151 also incorporates the Ant Design component library. The component library provides a wealth of UI components and styles, which can quickly realize an aesthetically pleasing and easy-to-use user interface. Through the use of the Ant Design component library, operation and maintenance personnel can easily operate and manage the knowledge base, improving the convenience and comfort of work.

[0108] For the convenience of description, the above devices are described in terms of functions and modules. Of course, when implementing one or more embodiments of the present application, the functions of each module can be implemented in the same or multiple software and / or hardware.

[0109] The apparatus of the above-mentioned embodiment is used to implement the corresponding method in the above-mentioned embodiment, and has the beneficial effects of the corresponding method embodiment, which will not be described in detail here.

[0110] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present disclosure (including the claims) is limited to these examples. Based on the concept of the present disclosure, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of different aspects of one or more embodiments of the present application as described above, which are not provided in detail for the sake of simplicity.

[0111] In addition, to simplify the description and discussion, and in order not to make one or more embodiments of the present application difficult to understand, the known power / ground connections to the integrated circuit (IC) chip and other components may or may not be shown in the provided drawings. In addition, the device can be shown in the form of a block diagram to avoid making one or more embodiments of the present application difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform on which one or more embodiments of the present application will be implemented (that is, these details should be fully within the scope of understanding of those skilled in the art). In the case of elaborating specific details (e.g., circuits) to describe exemplary embodiments of the present disclosure, it is obvious to those skilled in the art that one or more embodiments of the present application can be implemented without these specific details or when these specific details are changed. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0112] Although the present disclosure has been described in conjunction with specific embodiments of the present disclosure, many replacements, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.

[0113] One or more embodiments of the present application are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of one or more embodiments of the present application should be included in the scope of protection of this disclosure.

Claims

1. A work order processing system based on machine learning, characterized in that: include: The data layer is used to collect the operation and maintenance data of the target system; An intelligent layer, configured to obtain operation and maintenance data features and work order features based on the operation and maintenance data, the work orders to be executed and a feature extraction model, and to predict the priority of the work orders to be executed based on the operation and maintenance data features, the work order features and the work order processing model, wherein the work order processing model is trained based on executed work orders; The application layer is used to update the status of the work order to be executed according to the priority of the work order to be executed.

2. The system according to claim 1, characterized in that The application layer is also used to obtain user feedback on the work order to be executed, and send the feedback to the intelligent layer; the feedback indicates whether the priority of the work order to be executed is reasonable.

3. The system according to claim 2, characterized in that The work order processing model is trained based on the operation and maintenance data, the executed work orders and the feedback.

4. The system according to claim 3, characterized in that The training of the feature extraction model and the work order processing model includes iteratively executing the following steps until the model loss function value converges: Determine the actual priority of the executed work order according to the feedback information; Obtaining training work order features and training operation and maintenance data features according to the executed work order, the training operation and maintenance data, and the first machine learning model; Establishing a training sample according to the training work order characteristics, the training operation and maintenance data characteristics, and the actual priority of the executed work order; Inputting the training work order features and the training operation and maintenance data features into a second machine learning model to obtain a training result; Adjust the parameters of the first machine learning model and the second machine learning model according to the training results and the loss function value of the actual priority of the executed work order.

5. The system according to claim 1, characterized in that The operation and maintenance data is obtained from different data sources, including one or more of server logs and monitoring systems.

6. The system according to claim 5, characterized in that Also includes: The data fusion layer is used to fuse the operation and maintenance data from different data sources and send the fused operation and maintenance data to the intelligent layer.

7. The system according to claim 6, characterized in that The data fusion layer is also used to perform quality assessment on the processed operation and maintenance data, and send the processed operation and maintenance data with a quality score higher than a preset threshold to the intelligent layer.

8. The system according to claim 1, characterized in that The application layer is also used to receive instructions from users and create, edit or delete the to-be-executed work order according to the instructions.

9. The system according to claim 8, characterized in that The application layer is also used to store the work orders to be executed.

10. The system according to claim 1, characterized in that Also includes: The display layer is used to display the work orders to be executed according to the user's screening conditions.

Images