Method and device for predicting network fault processing duration, equipment and medium
By training the gradient enhancement tree model, analyzing historical fault data, predicting the network fault processing time and generating early warning information, the problem of inaccurate prediction and identification of timeout nodes in the existing technology is solved, and the fault response efficiency is improved.
Patent Information
- Application Number
- CN202411147018.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-20
- Publication Date
- 2024-11-26
AI Technical Summary
The prior art is difficult to locate and analyze the main features that affect the network fault processing time, and it is impossible to accurately predict the fault processing time and identify timeout nodes, resulting in a lack of effective reference for operators when handling network faults.
Using a machine learning-based method, by acquiring and processing the data of historical fault nodes, training the gradient boost tree model, predicting the fault processing time, and generating early warning information to indicate the timeout node.
It realizes accurate prediction of network fault handling time and identification of timeout nodes, helps operators quickly locate bottleneck points and improves fault response efficiency.
Smart Images

Figure CN119030887A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of fault handling time prediction, and specifically to a method, device, equipment and medium for predicting network fault handling time. Background Art
[0002] In the network fault handling work order process of communication operators, many factors have an impact on the processing time of the fault node. It is very important to predict the fault handling time based on the fault node information, find out the bottleneck point of fault handling timeout, and provide a reference for operators to handle network faults.
[0003] In the prior art, a single factor for handling a faulty node is analyzed through sorting statistics to provide a reference for operators to handle network failures.
[0004] However, this simple ranking statistical method is difficult to locate the main features that affect the fault handling time, cannot quantitatively analyze the importance of the features that affect the processing time, and cannot predict the node processing time based on the newly received fault handling work orders. Summary of the invention
[0005] The embodiments of the present application provide a method, apparatus, device and medium for predicting the duration of network fault processing, which are used to analyze the main influencing features that affect the duration of fault node processing and predict the duration of fault node processing.
[0006] In a first aspect, an embodiment of the present application provides a method for predicting a network fault processing time, comprising:
[0007] Obtain the fault data corresponding to the faulty node;
[0008] Inputting the fault data into a fault handling time prediction model to obtain a prediction result output by the fault handling time prediction model;
[0009] The prediction result is used to indicate the predicted fault processing time of the fault node, and the fault processing time prediction model is trained based on historical fault data corresponding to historical fault nodes and historical fault processing times.
[0010] In a possible implementation manner, before acquiring the fault data corresponding to the faulty node, the method further includes:
[0011] Obtain fault node data corresponding to multiple historical fault nodes;
[0012] Performing data processing on multiple groups of fault node data to obtain historical digital data corresponding to each group of fault node data;
[0013] The gradient boosting tree model is trained based on multiple groups of historical digital data to obtain the fault handling time prediction model.
[0014] In a possible implementation, the fault node data includes historical fault data and historical fault processing time, wherein the fault data includes: multiple types of fault parameters, and multiple groups of fault node data are processed to obtain historical digital data corresponding to each group of fault node data, including:
[0015] Classify and convert the historical fault data to obtain first digital data, wherein different types of fault parameters correspond to different digital data;
[0016] Standardizing the historical fault processing time to obtain second digital data, where the second digital data is used to indicate a preset number of processing time ranges;
[0017] The first digital data and the second digital data are aggregated to obtain the historical digital data.
[0018] In a possible implementation, a gradient boosting tree model is trained based on multiple groups of historical digital data to obtain the fault handling time prediction model, including:
[0019] Using the first digital data as input data and the second digital data as output data, iteratively processing the gradient boosting tree model to obtain hyperparameters of the gradient boosting tree model;
[0020] When the hyperparameters meet the preset conditions or the number of iterations reaches the preset number, determining that the training is completed;
[0021] According to the hyperparameters after training, the parameter range of the model parameters of the gradient boosting tree model after training is determined, and the model parameters of the gradient boosting tree model after training are optimized based on the parameter range to obtain the fault handling time prediction model.
[0022] In a possible implementation, the model parameters of the trained gradient boosting tree model are optimized based on the parameter range to obtain the fault handling time prediction model, including:
[0023] Determining a plurality of parameter combinations and the parameter combinations according to the parameter range;
[0024] According to the parameter combination, determining the prediction accuracy corresponding to each hyperparameter combination;
[0025] Determining an optimal prediction accuracy rate according to the prediction accuracy rate;
[0026] A fault handling time prediction model is determined according to the parameter combination corresponding to the optimal prediction accuracy.
[0027] In a possible implementation manner, before processing the multiple groups of fault node data to obtain the historical digital data corresponding to each group of fault node data, the method further includes:
[0028] For any set of faulty node data among the multiple sets of faulty node data, if the faulty node data has missing values, the faulty node data is discarded, and the discarded faulty node data with missing values are taken as multiple sets of faulty node data.
[0029] In a possible implementation, after inputting the fault data into the fault handling time prediction model to obtain a prediction result output by the fault handling time prediction model, the method further includes:
[0030] Determine whether the predicted duration of the fault handling is greater than a preset duration;
[0031] When the predicted fault processing time is greater than the preset time, a warning message is generated, where the warning message is used to indicate that the fault processing time of the fault node has timed out.
[0032] In a second aspect, an embodiment of the present application provides a device for predicting a network fault processing time, comprising:
[0033] An acquisition module, used to acquire fault data corresponding to a faulty node;
[0034] The prediction module is used to input the fault data into the fault processing time prediction model to obtain the prediction result output by the fault processing time prediction model.
[0035] In a possible implementation, the device also includes: a training module; used to obtain fault node data corresponding to multiple historical fault nodes; perform data processing on multiple groups of fault node data to obtain historical digital data corresponding to each group of fault node data; and perform training processing on a gradient boosting tree model based on multiple groups of historical digital data to obtain the fault processing time prediction model.
[0036] In a possible implementation manner, the fault node data includes: historical fault data and historical fault processing duration, wherein the fault data includes: multiple types of fault parameters, and the training module is further used to classify and convert the historical fault data to obtain first digital data, wherein different types of fault parameters correspond to different digital data;
[0037] Standardizing the historical fault processing time to obtain second digital data, where the second digital data is used to indicate a preset number of processing time ranges;
[0038] The first digital data and the second digital data are aggregated to obtain the historical digital data.
[0039] In a possible implementation, the training module is also used to use the first digital data as input data and the second digital data as output data, to iteratively process the gradient boosting tree model to obtain hyperparameters of the gradient boosting tree model; when the hyperparameters meet preset conditions or the number of iterations reaches a preset number, the training is determined to be completed; based on the hyperparameters after the training is completed, the parameter range of the model parameters of the gradient boosting tree model after the training is completed is determined, and based on the parameter range, the model parameters of the gradient boosting tree model after the training are optimized to obtain the fault handling time prediction model.
[0040] In a possible implementation manner, the training module is further used to determine a plurality of parameter combinations and the parameter combination according to the parameter range;
[0041] According to the parameter combination, determining the prediction accuracy corresponding to each hyperparameter combination;
[0042] Determining an optimal prediction accuracy rate according to the prediction accuracy rate;
[0043] A fault handling time prediction model is determined according to the parameter combination corresponding to the optimal prediction accuracy.
[0044] In a possible implementation, the device further includes: a processing module; used for discarding the faulty node data for any group of the multiple groups of faulty node data if the faulty node data has missing values, and treating the discarded faulty node data with missing values as multiple groups of faulty node data.
[0045] In a possible implementation, the device also includes: an early warning module; used to determine whether the predicted fault handling time is greater than a preset time; when the predicted fault handling time is greater than the preset time, generating early warning information, wherein the early warning information is used to indicate that the fault handling time of the fault node has timed out.
[0046] In a third aspect, an embodiment of the present application provides a device for predicting the duration of network fault processing, including: a memory, a processor;
[0047] The memory stores computer-executable instructions;
[0048] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the above first aspect and / or various possible implementations of the first aspect.
[0049] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the first aspect above and / or various possible implementations of the first aspect.
[0050] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the above first aspect and / or various possible implementation methods of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS
[0051] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.
[0052] Figure 1 Schematic diagram of the process of predicting the network fault processing time provided in the embodiment of the present application Figure 1 ;
[0053] Figure 2 Schematic diagram of the process of predicting the network fault processing time provided in the embodiment of the present application Figure 2 ;
[0054] Figure 3 A schematic diagram of the structure of a device for predicting network fault processing time provided in an embodiment of the present application;
[0055] Figure 4 A schematic diagram of the structure of a device for predicting network fault processing time provided in an embodiment of the present application. DETAILED DESCRIPTION
[0056] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0057] "Multiple" means two or more, and other quantifiers are similar. "And / or" describes the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects are in an "or" relationship.
[0058] In the embodiments of the present application, the words "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.
[0059] In the network fault handling work order process of communication operators, there are many factors that affect the processing time of network fault nodes. It is very important to predict the fault handling time based on the fault node information, find the bottleneck point of fault handling timeout, and provide reference for operators to handle network faults.
[0060] XGBoost (Extreme Gradient Boosting) model: is a gradient boosting model based on gradient boosting decision trees. It gradually reduces the prediction error of the model by continuously adding new trees to fit the residuals of the previous trees. In each iteration, XGBoost uses a technique called "gradient boosting" to optimize the objective function and minimize the loss function by adding new trees.
[0061] In the prior art, a single factor is analyzed through sorting statistics to provide a reference for operators to handle network failures.
[0062] However, the existing technology makes it difficult to locate the main features that affect the fault processing time, and it is impossible to quantitatively analyze the importance of the features that affect the processing time. It is also impossible to predict the node processing time based on the new fault processing work order and generate early warnings for timed nodes.
[0063] The network fault processing time prediction method provided in the present application is based on a machine learning method, a fault processing time prediction model is trained, and the fault data corresponding to the fault node information is input into the fault processing time prediction model to realize the prediction of the network fault processing time.
[0064] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.
[0065] Figure 1 Schematic diagram of the process of predicting the network fault processing time provided in the embodiment of the present application Figure 1 ,like Figure 1 As shown, the method includes:
[0066] S101. Obtain fault data corresponding to a faulty node.
[0067] The fault node is a fault handling work order corresponding to the network fault, and the data related to the fault handling work order is the fault data. The fault data may be, for example, a fault handling identifier, a fault area section, a fault cause, a fault handling node, and the like.
[0068] Obtain the fault handling ID, fault area section, fault cause, fault handling node and other fault data corresponding to the fault handling work order.
[0069] The fault handling identifier may be, for example, the fault handling province or the fault handling department.
[0070] The cause of the failure may be, for example, equipment failure, unknown cause, termination of troubleshooting, business consultation, incorrect operation, equipment power failure, cable interruption, etc.
[0071] Fault handling nodes may include, for example, troubleshooting, return order elimination, acceptance, upgrade, work order dispatch, order cancellation application, unblocking, urging, stage reply, and information appending.
[0072] Exemplarily, the fault data corresponding to the faulty node A is obtained, namely: Ax Province - Headquarters - User within the province - Equipment failure - Fault elimination.
[0073] S102: Input the fault data into a fault handling time prediction model to obtain a prediction result output by the fault handling time prediction model.
[0074] Among them, the fault handling time prediction model is pre-trained through historical fault data corresponding to historical fault nodes, and the prediction result includes the fault handling time.
[0075] The fault data corresponding to the fault node is taken as input and input into the fault processing time prediction model pre-trained by the historical fault data corresponding to the historical fault nodes to obtain the fault processing prediction time corresponding to the fault node. The fault data is analyzed for importance according to the fault processing time prediction model to obtain the main influencing features of the fault processing time corresponding to the fault node.
[0076] In a possible implementation, the method further includes:
[0077] After obtaining the prediction result output by the fault handling time prediction model, determine whether the fault handling prediction time is greater than the preset time; if the fault handling prediction time is greater than the preset time, generate an early warning message, which is used to indicate that the fault handling time of the fault node has exceeded.
[0078] Among them, the fault node whose predicted fault processing time is longer than the preset time is a timeout node, and the preset time is the maximum fault processing time of the fault node.
[0079] After obtaining the prediction result output by the fault processing time prediction model, according to the relationship between the fault processing time and the preset time in the prediction result, if the fault processing prediction time is greater than the preset time, that is, the fault node cannot complete the fault processing within the maximum fault processing time, then it indicates that the fault node is a timeout node, and for the timeout node, a warning message is generated; if the fault processing prediction time is not greater than the preset time, then it indicates that the fault node is not a timeout node, that is, the fault node can complete the fault processing within the maximum fault processing time, and therefore, there is no need to generate a warning message.
[0080] For a timeout node, the operator finds out the processing bottleneck point according to the main influencing characteristics of the fault processing time corresponding to the timeout node, and plans the processing of the timeout node in advance according to the processing bottleneck point to realize the rapid processing of the timeout node.
[0081] The network fault processing time prediction method provided in the embodiment of the present application obtains fault data corresponding to the fault node, inputs the fault data corresponding to the fault node into a pre-trained fault processing time prediction model, and the fault processing prediction model outputs a prediction result corresponding to the fault node. The prediction result provides a reference basis for operators to handle the fault node. For a timeout node whose fault processing time has exceeded, the processing bottleneck point corresponding to the timeout node can be determined more accurately and quickly according to the main influencing characteristics of the fault processing prediction time of the timeout node. The processing bottleneck point provides an important reference basis for operators when dealing with network faults of timeout nodes, so as to plan the processing of timeout nodes in advance, so as to realize rapid processing of timeout nodes and improve the efficiency of fault response.
[0082] Figure 2 Schematic diagram of the process of predicting the network fault processing time provided in the embodiment of the present application Figure 2 ,like Figure 2 As shown, in this embodiment Figure 2 Based on the embodiment, the training process of the fault handling time prediction model is supplemented, and the method includes:
[0083] S201. Obtain fault node data corresponding to a plurality of historical fault nodes, wherein the fault node data includes: historical fault data and historical fault processing duration.
[0084] The historical fault node is a node that has completed fault processing, and its corresponding fault node data may be historical fault data and historical fault processing duration.
[0085] The fault data includes: multiple types of fault parameters, and different types of fault parameters correspond to different digital data.
[0086] The historical fault data includes the historical fault processing identifier corresponding to the historical fault node, the historical fault area section, the historical fault cause, the historical fault processing node, etc. The historical fault processing identifier includes the historical fault processing province and the historical fault processing department.
[0087] The historical fault processing time is the processing time corresponding to the completion of the historical fault node.
[0088] Obtain fault node data of multiple nodes that have completed fault processing, including historical fault processing identifiers, historical fault area sections, historical fault causes, historical fault processing nodes, etc.
[0089] Exemplarily, the fault node data corresponding to the historical fault node B is Bx province-headquarters-overseas user-unknown reason-excluding return order-78717 minutes.
[0090] S202: Classify and convert the historical fault data to obtain first digital data.
[0091] Among them, the process of digitally converting the historical fault data according to the category of the historical fault data is classified conversion.
[0092] The historical fault data is converted according to categories such as historical fault processing province, historical fault processing department, historical fault area section, historical fault cause, historical fault processing node, etc. to obtain the first digital data.
[0093] For example, historical fault causes include: equipment failure, unknown cause, suspension of fault investigation, and business consultation; historical fault processing nodes include: troubleshooting, elimination of return orders, acceptance, and transfer.
[0094] The above historical fault causes are digitally converted to obtain the corresponding first digital data as follows:
[0095] 1- Equipment failure;
[0096] 2-Unknown cause;
[0097] 3-Stop troubleshooting;
[0098] 4-Business consultation.
[0099] The above historical fault processing nodes are digitally converted to obtain the corresponding first digital data as follows:
[0100] 1- Eliminate obstacles;
[0101] 2-Exclude return orders;
[0102] 3-Acceptance;
[0103] 4- Transfer.
[0104] S203: Standardize the historical fault processing time to obtain second digital data.
[0105] Among them, the process of converting historical fault handling time into dimensionless indicator data is standardization processing.
[0106] The standardization process can be a normalization process or a segmentation process.
[0107] Convert historical fault handling times into normalized data to facilitate subsequent training of the fault handling time prediction model.
[0108] The extreme values in multiple historical fault processing times are removed, and the historical fault processing times after the extreme values are removed are segmented according to the preset number of segments, and the multiple historical fault processing times are divided into the range of the preset number of segments in ascending order.
[0109] Exemplarily, the preset number of segments is 5. After de-extreme value operation is performed on multiple historical fault processing times, the remaining multiple historical fault processing times are divided into 5 ranges in ascending order. Within the corresponding ranges, the second digital data corresponding to the multiple historical fault processing times are 0, 1, 2, 3, and 4, respectively.
[0110] S204: Summarize the first digital data and the second digital data to obtain the historical digital data.
[0111] The process of aggregating the first digital data and the second digital data corresponding to the historical fault node is aggregation processing.
[0112] The first digital data and the second digital data corresponding to the historical fault node are aggregated together to obtain aggregated data, and the aggregated data is used as the historical digital data.
[0113] S205. Use the first digital data as input data and the second digital data as output data, iteratively process the gradient boosting tree model, and obtain hyperparameters of the gradient boosting tree model.
[0114] The gradient boosting tree model is an XGBoost gradient boosting tree model, and the corresponding hyperparameters may be, for example, the learning rate, tree depth, and number of iterations of the gradient boosting tree model.
[0115] According to the first digital data and the second digital data, an objective function is constructed, and the XGBoost gradient boosting tree is iteratively processed to find the optimal parameters based on the principle of minimizing the objective function. According to the optimal parameters, the gradient boosting tree after training is obtained.
[0116] Among them, the objective function O bjIt can be divided into the error function term L and the model complexity function term Ω.
[0117] The objective function can be expressed as:
[0118] O bj =L+Ω
[0119] According to the first digital data and the XGBoost gradient boosting tree model after each iterative processing, the corresponding objective function is obtained, wherein the error function term L represents the difference between the predicted result and the actual value, and the difference between the predicted value and the historical fault processing time is continuously reduced through multiple iterative processing.
[0120] The error function term L is expressed by the following formula:
[0121]
[0122] Where n is the number of iterations, y i is the historical fault handling time, y i ^ is the predicted value.
[0123] It is understandable that before iteration, the error function term L needs to be initialized, and the result of the initialization depends on the selected initialization method. For example, the initialization method may be to take the mean square error of the historical fault duration; or to take the mean absolute value error of the historical fault duration.
[0124] The model complexity function term Ω is used to limit the complexity of the XGBoost gradient boosting tree model and is expressed in the following formula:
[0125]
[0126] Among them, γT is the L1 regularization term, is the L2 regular term, T is the number of leaf nodes, and ω is the real number fraction of leaf nodes.
[0127] According to the above formula, the error function term is minimized and the objective function is optimized by adding new trees.
[0128] S206: Determine whether the hyperparameters meet the preset conditions or whether the number of iterations reaches the preset number. If yes, execute step S207; if not, execute step S205.
[0129] According to the hyperparameters corresponding to the gradient boosting tree obtained after each iteration, it is determined whether the preset adjustment is satisfied or whether the number of iterations reaches the preset number. If the hyperparameters satisfy the preset conditions or the number of iterations reaches the preset number, it indicates that the gradient boosting tree training is completed, and step S207 is executed. If the hyperparameters do not satisfy the preset conditions or the number of iterations does not reach the preset number, it indicates that the gradient boosting tree training is not completed, and the gradient boosting tree training is continued until the training is completed.
[0130] S207. Determine the parameter range of the model parameters of the gradient boosting tree model after training according to the hyperparameters after training, and optimize the model parameters of the gradient boosting tree model after training based on the parameter range to obtain the fault handling time prediction model.
[0131] Among them, the parameter range includes the hyperparameters that need to be optimized corresponding to the gradient boosting tree and the corresponding value range.
[0132] In one possible implementation, the model parameters of the gradient boosting tree model after training are optimized based on the parameter range to obtain the fault handling time prediction model. The specific implementation method is as follows: according to the parameter range, multiple parameter combinations and parameter combinations are determined; according to the parameter combination, the prediction accuracy corresponding to each hyperparameter combination is determined; according to the prediction accuracy, the best prediction accuracy is determined; according to the parameter combination corresponding to the best prediction accuracy, the fault handling time prediction model is determined.
[0133] Among them, the parameter range is the parameter that needs to be optimized and the corresponding value range; the prediction accuracy is the ratio between the predicted value of the fault handling time output by the gradient boosting tree model and the actual value, which is used to evaluate the performance of the gradient boosting tree model.
[0134] According to the model parameters of the gradient boosting tree model after training, the parameters that need to be optimized and the corresponding value range are determined to obtain a data set; the sliding window or rolling window method is used to split the data set so that each iteration uses a different data subset, that is, each iteration uses a different parameter combination for training. After training, the prediction accuracy of the corresponding gradient boosting tree model simulated according to each parameter combination is obtained, and the performance of each simulated gradient boosting tree model is evaluated based on the prediction accuracy.
[0135] Among multiple prediction accuracy rates, the prediction accuracy rate with the highest prediction accuracy rate is determined, that is, the gradient boosting tree with the best performance is determined. The gradient boosting tree corresponding to the best parameter combination is used as the fault handling time prediction model.
[0136] Exemplarily, the best parameter combination can be found through strategies such as grid search, random search or Bayesian optimization.
[0137] For example, parameter combination A: learning rate 0.005, tree depth: 3, number of iterations: 100, the corresponding prediction accuracy is: 87.21%;
[0138] Parameter combination B: learning rate 0.005, tree depth: 5, number of iterations: 100, corresponding prediction accuracy: 87.16%. Therefore, the gradient boosting tree corresponding to parameter combination A is used as the fault handling time prediction model.
[0139] In one possible implementation, before processing multiple groups of fault node data to obtain historical digital data corresponding to each group of fault node data, the fault node data can also be processed: for any group of fault node data among the multiple groups of fault node data, if the fault node data has missing values, the fault node data is discarded, and the discarded fault node data with missing values is used as multiple groups of fault node data.
[0140] If there are missing values in the fault node data, XGBoost will treat it as a special case in the subsequent training of the fault handling time prediction model, which will affect the training of the fault handling time prediction model. To ensure the stability and interpretability of the model, the fault node data needs to be processed first.
[0141] For any set of faulty node data among the multiple sets of faulty node data, if there are missing values in the set of faulty node data, it indicates that the set of data is incomplete, which may affect the model training. Therefore, the faulty node data with missing values is discarded, and the multiple sets of faulty node data remaining after discarding the incomplete data are used as the faulty node data; if there are no missing values in the set of faulty node data, it indicates that the set of data is complete, and the faulty node data is directly obtained.
[0142] The embodiment of the present application provides a method for predicting the network fault processing time. The method processes fault node data corresponding to multiple historical fault nodes to obtain corresponding multiple groups of historical data. After the multiple groups of historical data are processed according to classification conversion or standardization, corresponding historical digital data are obtained. The historical digital data is input into a gradient boosting tree model to obtain a trained gradient boosting tree. According to the model parameters of the trained gradient boosting tree model, optimizable hyperparameters and corresponding value ranges are determined. The trained gradient boosting tree model is optimized according to the parameter range to obtain a fault processing time prediction model for predicting the fault processing time corresponding to the fault node.
[0143] Figure 3 A schematic diagram of the structure of a device for predicting network fault processing time provided in an embodiment of the present application, such as Figure 3 The network fault processing time prediction device 30 provided in this embodiment includes:
[0144] An acquisition module 301 is used to acquire fault data corresponding to a faulty node;
[0145] The prediction module 302 is used to input the fault data into the fault processing time prediction model to obtain the prediction result output by the fault processing time prediction model.
[0146] In a possible implementation, the device also includes: a training module 303; used to obtain fault node data corresponding to multiple historical fault nodes; perform data processing on multiple groups of fault node data to obtain historical digital data corresponding to each group of fault node data; and perform training processing on a gradient boosting tree model based on multiple groups of historical digital data to obtain a fault processing time prediction model.
[0147] In a possible implementation manner, the fault node data includes: historical fault data and historical fault processing duration, wherein the fault data includes: multiple types of fault parameters, and the training module 303 is further used to classify and convert the historical fault data to obtain first digital data, wherein different types of fault parameters correspond to different digital data;
[0148] Performing standardization processing on the historical fault processing time to obtain second digital data, where the second digital data is used to indicate a preset number of processing time ranges;
[0149] The first digital data and the second digital data are aggregated to obtain historical digital data.
[0150] In a possible implementation, the training module 303 is also used to use the first digital data as input data and the second digital data as output data to iteratively process the gradient boosting tree model to obtain hyperparameters of the gradient boosting tree model; when the hyperparameters meet preset conditions or the number of iterations reaches a preset number, the training is determined to be completed; based on the hyperparameters after the training is completed, the parameter range of the model parameters of the gradient boosting tree model after the training is completed is determined, and based on the parameter range, the model parameters of the gradient boosting tree model after the training are optimized to obtain a fault handling time prediction model.
[0151] In a possible implementation, the training module 303 is further used to determine a plurality of parameter combinations and parameter combinations according to the parameter range;
[0152] According to the parameter combination, determine the prediction accuracy corresponding to each hyperparameter combination;
[0153] According to the prediction accuracy, determine the best prediction accuracy;
[0154] The fault handling time prediction model is determined based on the parameter combination corresponding to the best prediction accuracy.
[0155] In a possible implementation, the device also includes: a processing module; used to discard the faulty node data for any group of faulty node data among the multiple groups of faulty node data if the faulty node data has missing values, and to treat the discarded faulty node data with missing values as multiple groups of faulty node data.
[0156] In a possible implementation, the device also includes: an early warning module 304; used to determine whether the predicted duration of fault processing is greater than the preset duration; when the predicted duration of fault processing is greater than the preset duration, generate early warning information, and the early warning information is used to indicate that the fault processing duration of the fault node has timed out.
[0157] Figure 4 A schematic diagram of a network fault processing time prediction device provided in an embodiment of the present application. Figure 4 As shown, the network fault processing time prediction device 40 provided in this embodiment includes: at least one processor 401 and a memory 402. Optionally, the device 40 also includes a communication component 403. The processor 401, the memory 402 and the communication component 403 are connected via a bus 404.
[0158] In a specific implementation process, at least one processor 401 executes the computer-executable instructions stored in the memory 402, so that at least one processor 401 executes the prediction method based on the network fault processing time shown in the above embodiment.
[0159] The specific implementation process of the processor 401 can be found in the above method embodiment, and its implementation principle and technical effect are similar, so this embodiment will not be repeated here.
[0160] In the above embodiments, it should be understood that the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the invention may be directly implemented as being executed by a hardware processor, or may be executed by a combination of hardware and software modules in the processor.
[0161] The memory may include a high-speed memory (Random Access Memory, RAM), and may also include a non-volatile memory (Non-volatile Memory, NVM), such as at least one disk memory.
[0162] The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, the bus in the drawings of the present application is not limited to only one bus or one type of bus.
[0163] An embodiment of the present application further provides a computer program product, including a computer program, which, when executed by a processor, implements the method for predicting the network fault processing duration shown in the above embodiment.
[0164] The embodiment of the present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the method for predicting the network fault processing time shown in the above embodiment is implemented.
[0165] The above-mentioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general or special-purpose computer.
[0166] An exemplary readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (Application Specific Integrated Circuits, referred to as: ASIC). Of course, the processor and the readable storage medium can also exist in the device as discrete components.
[0167] The division of units is only a logical function division, and there may be other divisions in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interface, device or unit, which can be electrical, mechanical or other forms.
[0168] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0169] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
[0170] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention can essentially or in other words, the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.
[0171] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, disk or optical disk and other media that can store program codes.
[0172] Finally, it should be noted that those skilled in the art will readily conceive of other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses or adaptations of the present invention, which follow the general principles of the present invention and include common knowledge or customary technical means in the art not disclosed by the present invention, are not limited to the precise structure described above and shown in the drawings, and may be modified and changed in various ways without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.
Claims
1. A method for predicting the duration of network fault processing, characterized in that: include: Obtain the fault data corresponding to the faulty node; Inputting the fault data into a fault handling time prediction model to obtain a prediction result output by the fault handling time prediction model; The prediction result is used to indicate the predicted fault processing time of the fault node, and the fault processing time prediction model is trained based on historical fault data corresponding to historical fault nodes and historical fault processing times.
2. The method according to claim 1, characterized in that Before obtaining the fault data corresponding to the fault node, the method further includes: Obtain fault node data corresponding to multiple historical fault nodes; Performing data processing on multiple groups of fault node data to obtain historical digital data corresponding to each group of fault node data; The gradient boosting tree model is trained based on multiple groups of historical digital data to obtain the fault handling time prediction model.
3. The method according to claim 2, characterized in that The fault node data includes: historical fault data and historical fault processing time, wherein the fault data includes: multiple types of fault parameters, and the data processing of multiple groups of fault node data to obtain historical digital data corresponding to each group of fault node data includes: Classify and convert the historical fault data to obtain first digital data, wherein different types of fault parameters correspond to different digital data; Standardizing the historical fault processing time to obtain second digital data, where the second digital data is used to indicate a preset number of processing time ranges; The first digital data and the second digital data are aggregated to obtain the historical digital data.
4. The method according to claim 3, characterized in that The training process of the gradient boosting tree model according to the multiple groups of historical digital data to obtain the fault handling time prediction model includes: Using the first digital data as input data and the second digital data as output data, iteratively processing the gradient boosting tree model to obtain hyperparameters of the gradient boosting tree model; When the hyperparameters meet the preset conditions or the number of iterations reaches the preset number, determining that the training is completed; According to the hyperparameters after training, the parameter range of the model parameters of the gradient boosting tree model after training is determined, and the model parameters of the gradient boosting tree model after training are optimized based on the parameter range to obtain the fault handling time prediction model.
5. The method according to claim 4, characterized in that The optimizing process of the model parameters of the trained gradient boosting tree model based on the parameter range to obtain the fault handling time prediction model includes: Determining a plurality of parameter combinations and the parameter combinations according to the parameter range; According to the parameter combination, determining the prediction accuracy corresponding to each hyperparameter combination; Determining an optimal prediction accuracy rate according to the prediction accuracy rate; A fault handling time prediction model is determined according to the parameter combination corresponding to the optimal prediction accuracy.
6. The method according to claim 2, characterized in that Before performing data processing on the multiple groups of fault node data to obtain historical digital data corresponding to each group of fault node data, the method further includes: For any set of faulty node data among the multiple sets of faulty node data, if the faulty node data has missing values, the faulty node data is discarded, and the discarded faulty node data with missing values are taken as multiple sets of faulty node data.
7. The method according to claim 1, characterized in that After inputting the fault data into the fault handling time prediction model to obtain the prediction result output by the fault handling time prediction model, the method further includes: Determine whether the predicted duration of the fault handling is greater than a preset duration; When the predicted fault processing time is greater than the preset time, a warning message is generated, where the warning message is used to indicate that the fault processing time of the fault node has timed out.
8. A device for predicting the duration of network fault processing, characterized in that: include: An acquisition module, used to acquire fault data corresponding to a faulty node; The prediction module is used to input the fault data into the fault processing time prediction model to obtain the prediction result output by the fault processing time prediction model.
9. A device for predicting the duration of network fault processing, characterized in that: include: Memory, processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor executes the method for predicting the network fault processing duration according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method for predicting the network fault processing duration according to any one of claims 1 to 7.
Citation Information
Cited By
Method and device for processing read-write request, storage medium and electronic equipment
CN119473403A
