A fault locating method, device and electronic equipment

Abstract

Embodiments of the present application provide a fault positioning method and device and electronic equipment, and relate to the field of data processing. The method comprises: firstly obtaining fault alarm information; then extracting first language information and second language information in the fault alarm information, wherein the language categories of the first language information and the second language information are different; and inputting the first language information into a word segmentation model to obtain fault service type information output by the word segmentation model; then combining the second language information and the fault service type to obtain alarm comprehensive information; and finally inputting the alarm comprehensive information into a decision tree model to obtain fault position information output by the decision tree model. According to the embodiments of the present application, the fault position can be quickly positioned according to the fault alarm information, and the maintenance efficiency of the fault is improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a fault location method, apparatus, and electronic device. Background Technology

[0002] When a client encounters a fault during operation, such as an authentication attempt exceeding the limit and requesting a retry, the client will generate a fault alarm message including the location of the fault and send the fault alarm message to the server. The server will then send the fault log, which consists of fault alarm messages sent by the client or fault alarm messages sent by the server itself, to the development and operations department through the fault alarm notification system to determine the location of the fault and carry out repairs.

[0003] In existing technologies, the fault location provided in fault alarm information is not clear enough. For example, a fault alarm message submitted to the development and operations department through the fault alarm system may be: "Authentication attempts exceeded, please try again." Based on the fault alarm message, it can only be determined that the fault location belongs to the authentication service. However, since the authentication service also includes many other items such as authentication certificate information and security key information, it is difficult for the development and operations department to directly determine which type of item in the authentication service the fault location belongs to. As a result, the development and operations department needs to spend a lot of time locating the fault location, leading to low efficiency in fault repair. Summary of the Invention

[0004] The purpose of this invention is to provide a fault location method to quickly locate the fault based on fault alarm information, thereby improving fault repair efficiency. The specific technical solution is as follows:

[0005] In a first aspect, embodiments of the present invention provide a fault location method, the method comprising:

[0006] Obtain fault alarm information;

[0007] Extract the first language information and the second language information from the fault alarm information; wherein the first language information and the second language information are of different language categories;

[0008] The first language information is input into the word segmentation model to obtain the fault service type information output by the word segmentation model; wherein, the word segmentation model is obtained by training the first initial model with the first language sample information of the sample fault information as samples and the fault service type information as labels.

[0009] The second language information and the fault service type are combined to obtain comprehensive alarm information;

[0010] The alarm summary information is input into the decision tree model to obtain the fault location information output by the decision tree model; wherein, the decision tree model is trained on the second initial model using the alarm summary sample information as samples and the fault location information as labels.

[0011] Optionally, in one specific implementation, the training method of the word segmentation model includes: for each sample fault information, determining the first language sample information in the sample fault information as a sample, and determining the fault service type information corresponding to the first language sample information as a label, to obtain each first training sample; training the first initial model using each first training sample; and stopping the training when the first initial model meets the preset conditions, to obtain the word segmentation model.

[0012] Optionally, in one specific implementation, the training method of the decision tree model includes: for each sample fault information, determining the alarm comprehensive sample information obtained by combining the second language sample information and the fault service type in the sample fault information as a sample, and determining the fault location information corresponding to the alarm comprehensive sample information as a label, to obtain each second training sample; wherein, the alarm comprehensive sample information includes the fault service type information in the first initial model; training the second initial model using each second training sample; when the second initial model meets the preset conditions, stopping the training to obtain the decision tree model.

[0013] Optionally, in one specific implementation, the step of inputting the alarm comprehensive information into a decision tree model to obtain the fault location information output by the decision tree model includes: extracting keyword information from the second language information and combining it with the fault service type information to obtain alarm key information; the keyword information in the second language information includes one or more of fault service type information, fault occurrence time, fault code stream information, and fault content; inputting the alarm key information into the decision tree model, and determining the fault location information according to the annotation of the fault service type information; wherein, the annotation of the fault service type information includes the hierarchical relationship and calling relationship of the service type to which the fault service type information belongs.

[0014] Optionally, in one specific implementation, the method further includes: outputting the fault location information to a designated device.

[0015] Secondly, embodiments of the present invention provide a fault location device, the device comprising:

[0016] The alarm information acquisition module is used to acquire fault alarm information;

[0017] The language information extraction module is used to extract the first language information and the second language information from the fault alarm information; wherein the first language information and the second language information are of different language categories.

[0018] The first information input module is used to input the first language information into the word segmentation model to obtain the fault service type information output by the word segmentation model; wherein, the word segmentation model is obtained by training the first initial model with the first language sample information of the sample fault information as samples and the fault service type information as labels.

[0019] The alarm information combination module is used to combine the second language information and the fault service type to obtain comprehensive alarm information;

[0020] The comprehensive information input module is used to input the comprehensive alarm information into the decision tree model to obtain the fault location information output by the decision tree model; wherein, the comprehensive alarm sample information is a sample, and the fault location information is used as a label to train the second initial model.

[0021] Optionally, in one specific implementation, the training method of the word segmentation model includes: for each sample fault information, determining the first language sample information in the sample fault information as a sample, and determining the fault service type information corresponding to the first language sample information as a label, to obtain each first training sample; training the first initial model using each first training sample; and stopping the training when the first initial model meets the preset conditions, to obtain the word segmentation model.

[0022] Optionally, in one specific implementation, the training method of the decision tree model includes: for each sample fault information, determining the alarm comprehensive sample information obtained by combining the second language sample information and the fault service type in the sample fault information as a sample, and determining the fault location information corresponding to the alarm comprehensive sample information as a label, to obtain each second training sample; wherein, the alarm comprehensive sample information includes the fault service type information in the first initial model; training the second initial model using each second training sample; when the second initial model meets the preset conditions, stopping the training to obtain the decision tree model.

[0023] Optionally, in one specific implementation, the comprehensive information input module is specifically used to: extract keyword information from the second language information and combine it with the fault service type information to obtain alarm key information; the keyword information in the second language information includes one or more of fault service type information, fault occurrence time, fault code stream information, and fault content; input the alarm key information into the decision tree model, and determine the fault location information according to the annotation of the fault service type information; wherein, the annotation of the fault service type information includes the hierarchical relationship and calling relationship of the service type to which the fault service type information belongs.

[0024] Optionally, in one specific implementation, the device further includes: a location information output module, used to output the fault location information to a designated device.

[0025] Thirdly, embodiments of the present invention provide an electronic device, including: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0026] Memory, used to store computer programs;

[0027] When a processor executes a program stored in memory, it implements the steps of any of the fault location methods provided in the first aspect above.

[0028] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements any of the fault location method steps provided in the first aspect above.

[0029] Beneficial effects of the embodiments of the present invention:

[0030] As can be seen from the above, when determining the fault location using the technical solution provided in the embodiments of the present invention, fault alarm information can be obtained first; then, the first language information and the second language information in the fault alarm information can be extracted, wherein the first language information and the second language information have different language categories; and the first language information can be input into a word segmentation model to obtain fault service type information output by the word segmentation model; wherein, the word segmentation model is trained on a first initial model using the first language sample information of the sample fault information as samples and the fault service type information as labels; then, the second language information and the fault service type information are combined to obtain comprehensive alarm information, and finally, the comprehensive alarm information is input into a decision tree model to obtain fault location information output by the decision tree model; wherein, the decision tree model is trained on a second initial model using the second language sample information and sample service information of the sample fault information as samples and the fault location information as labels.

[0031] Based on this, the solution provided by the embodiments of the present invention, compared to the layer-by-layer analysis and algorithmic classification of fault alarm information by development and maintenance personnel to determine fault location information, utilizes a pre-trained word segmentation model and decision tree model. By leveraging the first and second language information extracted from the fault alarm information, the fault location information can be directly and quickly located. Therefore, the embodiments of the present invention can quickly locate fault locations based on fault alarm information, thereby improving fault repair efficiency. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0033] Figure 1 This is a flowchart illustrating the first fault location method in an embodiment of the present invention.

[0034] Figure 2 This is a flowchart illustrating the second fault location method in an embodiment of the present invention.

[0035] Figure 3 This is a flowchart illustrating the third fault location method in this embodiment of the invention.

[0036] Figure 4 This is a flowchart illustrating the fourth fault location method in this embodiment of the invention.

[0037] Figure 5 This is a schematic diagram of the structure of a fault location device according to an embodiment of the present invention.

[0038] Figure 6 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0039] The technical solutions of the present invention will now be described with reference to the accompanying drawings in the embodiments of the present invention.

[0040] In related technologies, the fault alarms obtained are often vague. Based solely on the alarm information, only the type of service experiencing the fault can be determined, not the specific location of the anomaly. Furthermore, since multiple service types may be calling each other at the fault location, fault localization is difficult and time-consuming. Therefore, how to quickly locate the fault based on alarm information, thereby improving fault repair efficiency, has become an urgent problem to be solved.

[0041] To address the aforementioned technical problems, this application provides a fault location method.

[0042] This method is applicable to various application scenarios requiring fault localization, such as identity authentication errors during online withdrawals or webpage loading failures. Furthermore, it can be applied to various types of electronic devices, including servers, desktop computers, and mobile phones (hereinafter referred to as electronic devices). Therefore, this embodiment of the invention does not limit the application scenarios or the executing entity of the method.

[0043] The electronic device can execute the fault location method through an installed client with fault location functionality, or it can execute the fault location method through a module configured with fault location functionality. Furthermore, when the electronic device executes the fault location method through a client, the module used to implement the fault location method can be a plugin for the client; the client can also be a dedicated client for executing the fault location method. All of these are reasonable.

[0044] This application provides a fault location method that may include the following steps:

[0045] Obtain fault alarm information;

[0046] Extract the first language information and the second language information from the fault alarm information; wherein the first language information and the second language information are of different language categories;

[0047] The first language information is input into the word segmentation model to obtain the fault service type information output by the word segmentation model; wherein, the word segmentation model is obtained by training the first initial model with the first language sample information of the sample fault information as samples and the fault service type information as labels.

[0048] The second language information and the fault service type are combined to obtain comprehensive alarm information;

[0049] The alarm summary information is input into the decision tree model to obtain the fault location information output by the decision tree model; wherein, the decision tree model is trained on the second initial model using the alarm summary sample information as samples and the fault location information as labels.

[0050] As can be seen from the above, when determining the fault location using the technical solution provided in the embodiments of the present invention, fault alarm information can be obtained first; then, the first language information and the second language information in the fault alarm information can be extracted, wherein the first language information and the second language information have different language categories; and the first language information can be input into a word segmentation model to obtain the fault service type information output by the word segmentation model; wherein, the word segmentation model is trained on a first initial model using the first language sample information of the sample fault information as samples and the fault service type information as labels; then, the second language information and the fault service type information are combined to obtain comprehensive alarm information, and finally, the comprehensive alarm information is input into a decision tree model to obtain the fault location information output by the decision tree model; wherein, the decision tree model is trained on a second initial model using the second language sample information and sample service information of the sample fault information as samples and the fault location information as labels.

[0051] Based on this, the solution provided by the embodiments of the present invention, compared with the development and maintenance personnel performing layer-by-layer analysis and algorithm classification of fault alarm information to determine fault location information, can obtain fault location information by inputting the first language information and second language information extracted from the fault alarm information into the pre-trained word segmentation model and decision tree model. Therefore, the embodiments of the present invention can quickly locate the fault location based on the fault alarm information, thereby improving the fault repair efficiency.

[0052] The following is a detailed description of a fault location method provided in an embodiment of this application, with reference to the accompanying drawings.

[0053] Figure 1 This is a flowchart illustrating a fault location method provided in an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes steps S101-S105:

[0054] Step S101: Obtain fault alarm information;

[0055] Typically, when various server-side components, such as clients and servers, encounter failures during operation, such as a requested page not existing, an internal server error, or a server refusing access, they will generate fault alarm information related to the failure, such as the time of the failure and the type of failure. The client-side and server-side components will then deliver the fault alarm information to the electronic devices executing this solution.

[0056] Optionally, for various service terminals such as clients and servers connected to the fault notification system, when the fault notification system detects a fault in the service terminal, the fault notification system generates a fault alarm message and delivers it to the electronic device executing this solution.

[0057] Optionally, for various service terminals such as clients and servers connected to the fault notification system, when the fault notification system detects that the server has generated a fault alarm, the fault notification system receives the fault alarm information and delivers it to the electronic device executing this solution.

[0058] Step S102: Extract the first language information and the second language information from the above fault alarm information;

[0059] Among them, the language categories of the first language information and the second language information mentioned above are different;

[0060] According to the technical requirements for generating fault alarm information, the aforementioned fault alarm information includes two language information, namely, first language information and second language information. Furthermore, the first language information and the second language information belong to different language categories; for example, the first language information is in Chinese, and the second language information is in English. The first language information is used to determine the fault service type; the second language information and the fault service type are used to determine the fault location information. Therefore, after obtaining the fault alarm information, the first language information and the second language information can be extracted from it.

[0061] For example, the first language information can be Chinese, and the second language information can be English. The first language information and the second language information are used as inputs to the subsequent word segmentation model and decision tree model, respectively. The specific methods will be explained in detail in the following content. In addition, the embodiments of the present invention do not limit the language categories to which the first language information and the second language information belong.

[0062] Step S103: Input the above first language information into the word segmentation model to obtain the fault service type information output by the word segmentation model;

[0063] The word segmentation model described above is obtained by training the first initial model with the first language sample information of the sample fault information as the sample and the fault service type information as the label.

[0064] The first language information extracted from the fault alarm information is used as the input to the word segmentation model. The word segmentation algorithm in the word segmentation model is used to obtain the fault service type.

[0065] For example, when the first language information is Chinese, the word segmentation model can use NLP (Neuro-Linguistic Programming) word segmentation algorithms. The output fault service type can include various types such as contract service, withdrawal service, authentication service, and query service. The fault service type output by the word segmentation model is also different depending on the different fault alarm information.

[0066] During the training of the preset initial model using the aforementioned first language sample information and labels, the initial model can learn the correspondence between each set of samples and labels. Furthermore, through learning the correspondence between a large number of samples and labels, the correspondence between samples and labels is gradually established.

[0067] In this way, after extracting the first language information from the fault alarm information, the first language information can be input into the word segmentation model. The word segmentation model can determine the tag corresponding to the input first language information based on the correspondence determined above, and output it as the fault service type related to the first language information.

[0068] Step S104: Combine the above-mentioned second language information and the above-mentioned fault service type to obtain comprehensive alarm information;

[0069] After obtaining the second language information and fault service type through the above steps, the second language information and fault service type can be combined to obtain comprehensive alarm information.

[0070] The alarm summary information obtained by combining the second language information and fault service type can include all the language information in the second language information; alternatively, some second language information can be selected to be combined with fault service type according to needs, such as extracting the required keyword information from the second language information: fault occurrence time, specific fault type, etc., or one or more of these are all reasonable.

[0071] Optionally, the alarm summary information obtained by combining the aforementioned second language information and fault service type can be obtained by concatenating the second language information and fault service type, wherein there is no limitation on the concatenation order and method of the second language information and fault service type; in addition, the two can also be combined as independent parallel information and the combined information can be used as the alarm summary information. Both are reasonable.

[0072] Step S105: Input the above alarm information into the decision tree model to obtain the fault location information output by the decision tree model;

[0073] The alarm information includes the second language information and the fault service type information output by the word segmentation model. The decision tree model is trained on the second initial model using the alarm information as samples and the fault location information as labels.

[0074] The alarm composite sample information mentioned above is obtained by combining second language sample information and fault service type.

[0075] During the training of the preset second initial model using the above alarm comprehensive sample information as samples and fault location information as labels, the second initial model can learn the correspondence between each set of samples and labels. Furthermore, through learning a large number of sample and label correspondences, the correspondence between samples and labels is gradually established.

[0076] In this way, the alarm summary information is input into the preset decision tree model. Based on the correspondence determined above, the decision tree model can determine the label corresponding to the input alarm summary information and output it as the fault location information of the alarm summary information.

[0077] The present invention will now be described through a specific embodiment:

[0078] When a request error displayed on a webpage is submitted to the development team via a fault notification system, after receiving the aforementioned fault alarm information, the first and second language information in the fault alarm information are extracted. For example, if the first language information is Chinese and the second language information is English, the Chinese alarm information in the fault alarm information is input into a word segmentation model. The fault service type output by the NLP (Neuro-Linguistic Programming) word segmentation algorithm is authentication service. The second language information and the fault service type are then combined to obtain comprehensive alarm information. This comprehensive alarm information is then input into a decision tree model, and the root cause analysis method is used to determine the output fault location information, which is that the client program requesting authentication service has been corrupted.

[0079] As can be seen from the above, when determining the fault location using the technical solution provided in the embodiments of the present invention, fault alarm information can be obtained first; then, the first language information and the second language information in the fault alarm information can be extracted, wherein the first language information and the second language information have different language categories; and the first language information can be input into a word segmentation model to obtain fault service type information output by the word segmentation model; wherein, the word segmentation model is trained on a first initial model using the first language sample information of the sample fault information as samples and the fault service type information as labels; then, the service information corresponding to the fault service type information is determined according to the second language information; finally, the second language information and the service information are input into a decision tree model to obtain fault location information output by the decision tree model; wherein, the service information includes the fault service type information output by the word segmentation model; and the decision tree model is trained on a second initial model using the second language sample information and sample service information of the sample fault information as samples and the fault location information as labels.

[0080] Based on this, the solution provided by the embodiments of the present invention, compared with the development and maintenance personnel performing layer-by-layer analysis and algorithm classification of fault alarm information to determine fault location information, can obtain fault location information by inputting the first language information and second language information extracted from the fault alarm information into the pre-trained word segmentation model and decision tree model. Therefore, the embodiments of the present invention can quickly locate the fault location based on the fault alarm information, thereby improving the fault repair efficiency.

[0081] Alternatively, in one specific implementation, such as Figure 4 As shown, step S104 above: combining the above second language information and the above fault service type to obtain alarm comprehensive information may include S401:

[0082] Step S401: Extract the keyword information from the above second language information and combine it with the above fault service type information to obtain alarm key information; wherein, the keyword information in the above second language information includes one or more of the following: fault service type information, fault occurrence time, fault code stream information, and fault content;

[0083] Optionally, the aforementioned key alarm information can be obtained by concatenating keyword information from the second language information with the fault service type.

[0084] Step S105 above: Inputting the above alarm integrated information into the decision tree model to obtain the fault location information output by the decision tree model may include S402:

[0085] Step S402: Input the above alarm key information into the decision tree model, and determine the fault location information based on the labeling of the fault service type information;

[0086] The labeling of the aforementioned fault service type information includes the hierarchical relationship and invocation relationship of the service type to which the fault service type information belongs.

[0087] After extracting the keyword information, it is combined with the fault service type as the input to the decision tree model; similarly, when training the decision tree model in advance, the alarm comprehensive sample information in the second training sample is composed of the keyword sample information and the fault service type.

[0088] After extracting the second language information from the fault alarm information, in order to determine the fault location information more quickly, keyword information can be extracted from the second language information and used as input to the decision tree model. The keyword information includes one or more of the following: fault service type information, fault occurrence time, fault code stream information, and fault content. Using the keyword information as input to the decision tree model narrows the scope of determining the fault location information.

[0089] Furthermore, since each service type includes different specific project information, such as authentication services including login database, security information authentication, authentication service creation, and authentication service feedback; and presentation services including account information review, login database, and encrypted information authentication; and since different service types may also have project information calls, it is possible to add annotations to the faulty service type. These annotations include the hierarchical relationship and call relationship of the service type to which the faulty service type information belongs, that is, the different specific project information included in the faulty service type and the record of each project information being called by each service type.

[0090] Alternatively, in one specific implementation, such as Figure 2 As shown, the training method of the above word segmentation model can include the following steps S201-S203:

[0091] Step S201: For each sample fault information, the first language sample information in the sample fault information is determined as a sample, and the fault service type information corresponding to the first language sample information is determined as a label, so as to obtain each first training sample;

[0092] When training the word segmentation model, the first language sample information in the sample fault information is collected as the sample, and the fault service type corresponding to each first language sample information is determined as the label to form each first training sample, which is used for the training of the subsequent first initial model; wherein, the first language sample information can be Chinese sample information or sample information of other language types.

[0093] Step S202: Train the first initial model using each of the first training samples;

[0094] Step S203: When the first initial model meets the preset conditions, stop training and obtain the word segmentation model.

[0095] After obtaining each first training sample, the first initial model can be trained using each first training sample. Training is stopped when the first initial model meets the preset conditions, thus obtaining the word segmentation model.

[0096] During the training process of the preset first initial model using the above training samples, the first initial model can learn the correspondence between each first language sample information and the fault service type label. Furthermore, through learning the correspondence between a large number of samples and labels, the correspondence between the first language sample information and the fault service type label is gradually established.

[0097] In this way, after obtaining the first language information in the fault alarm information, the first language sample information can be input into the pre-trained word segmentation model. The word segmentation model can determine the tag corresponding to the input first language information based on the correspondence determined above, and output it as the fault service type of the first language information.

[0098] The electronic device used for model training and the fault location device for executing the fault location method provided in this application embodiment may be the same electronic device or different electronic devices.

[0099] The first initial model mentioned above can be a word segmentation model constructed based on one or more of the following neural networks: CNN (Convolutional Neural Network), GRU (Gated Recurrent Unit) neural network, LSTM (Long Short-Term Memory). This application embodiment does not impose specific limitations on the construction of the first initial model.

[0100] Optionally, the above preset condition can be that the number of iterations for each first training sample reaches a preset number.

[0101] Optionally, the first training samples can be divided into training set samples and test set samples. Furthermore, the preset condition can be that the error between the accuracy of the label and the predicted value of each test set sample is less than a preset error. The true value of the sample label of the test set sample can be the fault service type corresponding to the label of the test set sample, and the predicted value of the label of the test set sample can be the fault service type obtained by inputting the test set sample into the word segmentation model.

[0102] Alternatively, in one specific implementation, such as Figure 3 As shown, the training method of the above decision tree model includes steps S301-S303:

[0103] Step S301: For each sample fault information, the above alarm comprehensive sample information is determined as a sample, and the fault location information corresponding to the alarm comprehensive sample information is determined as a label to obtain each second training sample; wherein, the alarm comprehensive sample information is obtained by combining the second language sample information extracted from the sample fault information and the fault service type.

[0104] When training the decision tree model, the alarm comprehensive sample information in the sample fault information is taken as the sample, and the fault location information corresponding to each alarm comprehensive sample information is determined as the label to form each second training sample for subsequent training of the second initial model; wherein, the above-mentioned second language sample information can be English sample information or other language type sample information; wherein, the above-mentioned sample fault information and the sample fault information collected from the first language sample information are the same sample fault information.

[0105] Step S302: Train the second initial model using each of the second training samples;

[0106] Step S303: When the second initial model meets the preset conditions, stop training and obtain the decision tree model.

[0107] After obtaining each second training sample, the second initial model can be trained using each second training sample. Training is stopped when the second initial model meets the preset conditions, thus obtaining the decision tree model.

[0108] During the training process of the preset second initial model using the above training samples, the second initial model can learn the correspondence between each alarm comprehensive sample information and the fault location information label. Furthermore, through learning the correspondence between a large number of samples and labels, the correspondence between alarm comprehensive sample information and fault location information labels is gradually established. The alarm comprehensive sample information is obtained by combining the second language sample information and the fault service type information corresponding to the first language sample information in the sample fault information where the second language sample information is located.

[0109] In this way, after obtaining the second language information from the fault alarm information, the second language sample information is combined with the fault service type to obtain the alarm comprehensive sample information. The alarm comprehensive sample information can then be input into the pre-trained decision tree model. The decision tree model can determine the label corresponding to the input alarm comprehensive information based on the above-determined correspondence and output it as the fault service type of the alarm comprehensive information.

[0110] The electronic device used for model training and the fault location device for executing the fault location method provided in this application embodiment may be the same electronic device or different electronic devices.

[0111] The aforementioned second initial model can be a decision tree model constructed based on one or more of the following neural networks: CNN (Convolutional Neural Network), GRU (Gated Recurrent Unit) neural network, LSTM (Long Short-Term Memory). This application embodiment does not impose specific limitations on the construction of the second initial model.

[0112] Optionally, the above preset condition can be that the number of iterations for each second training sample reaches a preset number.

[0113] Optionally, the aforementioned second training samples can be divided into training set samples and test set samples. Furthermore, the aforementioned preset condition can be that the error between the accuracy of the label and the predicted value of each test set sample is less than a preset error. Here, the true value of the sample label of the aforementioned test set sample can be the fault location information corresponding to the label of the test set sample, and the predicted value of the label of the aforementioned test set sample can be the fault location information obtained by inputting the test set sample into the decision tree model.

[0114] Optionally, in one specific implementation, the fault location method provided in this embodiment of the invention further includes the following step A1:

[0115] Step A1: Output the above fault location information to the designated device.

[0116] After receiving the fault alarm information, the fault location information can be output to a designated device. The designated device can be a terminal device used for fault maintenance, a terminal device used for system development, or the terminal device that obtained the fault location information. All of these are reasonable and the present invention does not impose any specific restrictions on them.

[0117] In this way, after the designated device receives the fault location information, it can promptly address the fault location to improve the user experience.

[0118] Corresponding to the fault location method provided in the above embodiments of this application, the embodiments of this application also provide a fault location device, such as... Figure 5 As shown, the device includes:

[0119] The alarm information acquisition module 501 is used to acquire fault alarm information;

[0120] The language information extraction module 502 is used to extract the first language information and the second language information from the fault alarm information; wherein the first language information and the second language information are of different language categories.

[0121] The first information input module 503 is used to input the first language information into the word segmentation model to obtain the fault service type information output by the word segmentation model; wherein, the word segmentation model is obtained by training the first initial model with the first language sample information of the sample fault information as samples and the fault service type information as labels.

[0122] The alarm information combination module 504 is used to combine the second language information and the fault service type to obtain comprehensive alarm information;

[0123] The comprehensive information input module 505 is used to input the comprehensive alarm information into the decision tree model to obtain the fault location information output by the decision tree model; wherein, the comprehensive alarm sample information is a sample, and the fault location information is used as a label to train the second initial model.

[0124] As can be seen from the above, when determining the fault location using the technical solution provided in the embodiments of the present invention, fault alarm information can be obtained first; then, the first language information and the second language information in the fault alarm information can be extracted, wherein the first language information and the second language information have different language categories; and the first language information can be input into a word segmentation model to obtain fault service type information output by the word segmentation model; wherein, the word segmentation model is trained on a first initial model using the first language sample information of the sample fault information as samples and the fault service type information as labels; then, the second language information and the fault service type information are combined to obtain comprehensive alarm information, and finally, the comprehensive alarm information is input into a decision tree model to obtain fault location information output by the decision tree model; wherein, the decision tree model is trained on a second initial model using the second language sample information and sample service information of the sample fault information as samples and the fault location information as labels.

[0125] Based on this, the solution provided by the embodiments of the present invention, compared with the development and maintenance personnel performing layer-by-layer analysis and algorithm classification of fault alarm information to determine fault location information, can obtain fault location information by inputting the first language information and second language information extracted from the fault alarm information into the pre-trained word segmentation model and decision tree model. Therefore, the embodiments of the present invention can quickly locate the fault location based on the fault alarm information, thereby improving the fault repair efficiency.

[0126] Optionally, in one specific implementation, the comprehensive information input module 505 is specifically used for:

[0127] Keyword information is extracted from the second language information and combined with the fault service type information to obtain alarm key information; the keyword information in the second language information includes one or more of the following: fault service type information, fault occurrence time, fault code stream information, and fault content;

[0128] The alarm key information is input into the decision tree model, and the fault location information is determined based on the labeling of the fault service type information; wherein, the labeling of the fault service type information includes the hierarchical relationship and calling relationship of the service type to which the fault service type information belongs.

[0129] Optionally, in one specific implementation, the training method of the word segmentation model includes:

[0130] For each sample fault information, the first language sample information in the sample fault information is determined as a sample, and the fault service type information corresponding to the first language sample information is determined as a label, so as to obtain each first training sample;

[0131] The first initial model is trained using each of the first training samples;

[0132] When the first initial model meets the preset conditions, training stops, and the word segmentation model is obtained.

[0133] Optionally, in one specific implementation, the training method of the decision tree model includes:

[0134] For each sample fault information, the alarm comprehensive sample information obtained by combining the second language sample information and the fault service type in the sample fault information is determined as a sample, and the fault location information corresponding to the alarm comprehensive sample information is determined as a label, thus obtaining each second training sample; wherein, the alarm comprehensive sample information includes the fault service type information in the first initial model;

[0135] The second initial model is trained using each of the second training samples;

[0136] When the second initial model meets the preset conditions, training stops, and the decision tree model is obtained.

[0137] Optionally, in one specific implementation, the apparatus further includes:

[0138] The location information output module is used to output the fault location information to a designated device.

[0139] This invention also provides an electronic device, such as... Figure 6As shown, it includes a processor 601, a communication interface 602, a memory 603, and a communication bus 604, wherein the processor 601, the communication interface 602, and the memory 603 communicate with each other through the communication bus 604.

[0140] Memory 603 is used to store computer programs;

[0141] When processor 601 executes a program stored in memory 603, it performs the following steps:

[0142] Obtain fault alarm information;

[0143] Extract the first language information and the second language information from the fault alarm information; wherein the first language information and the second language information are of different language categories;

[0144] The first language information is input into the word segmentation model to obtain the fault service type information output by the word segmentation model; wherein, the word segmentation model is obtained by training the first initial model with the first language sample information of the sample fault information as samples and the fault service type information as labels.

[0145] The second language information and the fault service type are combined to obtain comprehensive alarm information;

[0146] The alarm summary information is input into the decision tree model to obtain the fault location information output by the decision tree model; wherein, the decision tree model is trained on the second initial model using the alarm summary sample information as samples and the fault location information as labels.

[0147] The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0148] The communication interface is used for communication between the aforementioned terminal and other devices.

[0149] The memory may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0150] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0151] In another embodiment of the present invention, a computer-readable storage medium is also provided, wherein a computer program is stored therein, and when the computer program is executed by a processor, it implements any of the fault location methods described in the above embodiments.

[0152] In another embodiment of the present invention, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute any of the fault location methods described in the above embodiments.

[0153] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk (SSD)).

[0154] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0155] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0156] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A fault location method, characterized in that, Applied to electronic devices, the method includes: Obtain fault alarm information; wherein the fault alarm information comes from the client or the server and is presented in natural language. Extract the first language information and the second language information from the fault alarm information; wherein the first language information and the second language information have different language categories, the first language information is used to determine the fault service type, and the second language information and the fault service type are used to determine the fault location information. The first language information is input into the word segmentation model to obtain the fault service type information output by the word segmentation model; wherein, the word segmentation model is obtained by training the first initial model with the first language sample information of the sample fault information as samples and the fault service type information as labels. Keyword information is extracted from the second language information and combined with the fault service type information to obtain alarm key information; the keyword information in the second language information includes one or more of the following: fault service type information, fault occurrence time, fault code stream information, and fault content; The alarm key information is input into the decision tree model, and the fault location information is determined according to the labeling of the fault service type information; wherein, the labeling of the fault service type information includes the hierarchical relationship and calling relationship of the service type to which the fault service type information belongs, and the decision tree model is trained on the second initial model using alarm key sample information as samples and fault location information as labels; The training method of the decision tree model includes: for each sample fault information, determining the alarm key sample information obtained by combining the keyword information in the second language sample information of the sample fault information and the fault service type as a sample, and determining the fault location information corresponding to the alarm key sample information as a label, to obtain each second training sample; wherein, the alarm key sample information includes the fault service type information in the first initial model; training the second initial model using each second training sample; when the second initial model meets the preset conditions, stopping the training, and obtaining the decision tree model.

2. The method according to claim 1, characterized in that, The training methods for the word segmentation model include: For each sample fault information, the first language sample information in the sample fault information is determined as a sample, and the fault service type information corresponding to the first language sample information is determined as a label, so as to obtain each first training sample; The first initial model is trained using each of the first training samples; When the first initial model meets the preset conditions, training stops, and the word segmentation model is obtained.

3. The method according to claim 1, characterized in that, The method further includes: Output the fault location information to the designated device.

4. A fault location device, characterized in that, The device includes: An alarm information acquisition module is used to acquire fault alarm information; wherein, the fault alarm information comes from the client or the server and is presented in natural language form; The language information extraction module is used to extract first language information and second language information from the fault alarm information; wherein the first language information and the second language information have different language categories, the first language information is used to determine the fault service type, and the second language information and the fault service type are used to determine the fault location information. The first information input module is used to input the first language information into the word segmentation model to obtain the fault service type information output by the word segmentation model; wherein, the word segmentation model is obtained by training the first initial model with the first language sample information of the sample fault information as samples and the fault service type information as labels. The alarm information combination module is used to combine the second language information and the fault service type to obtain comprehensive alarm information; The comprehensive information input module is used to input the comprehensive alarm information into the decision tree model to obtain the fault location information output by the decision tree model; wherein, the decision tree model is trained on the second initial model using the comprehensive alarm sample information as samples and the fault location information as labels; The comprehensive information input module is specifically used for: extracting keyword information from the second language information and combining it with the fault service type information to obtain alarm key information; the keyword information in the second language information includes one or more of fault service type information, fault occurrence time, fault code stream information, and fault content; inputting the alarm key information into the decision tree model, and determining the fault location information based on the annotation of the fault service type information; wherein, the annotation of the fault service type information includes the hierarchical relationship and calling relationship of the service type to which the fault service type information belongs; The training methods for the decision tree model include: For each sample fault information, the alarm comprehensive sample information obtained by combining the second language sample information and the fault service type in the sample fault information is determined as a sample, and the fault location information corresponding to the alarm comprehensive sample information is determined as a label, thus obtaining each second training sample; wherein, the alarm comprehensive sample information includes the fault service type information in the first initial model; The second initial model is trained using each of the second training samples; When the second initial model meets the preset conditions, training stops, and the decision tree model is obtained.

5. The apparatus according to claim 4, characterized in that, The training methods for the word segmentation model include: For each sample fault information, the first language sample information in the sample fault information is determined as a sample, and the fault service type information corresponding to the first language sample information is determined as a label, so as to obtain each first training sample; The first initial model is trained using each of the first training samples; When the first initial model meets the preset conditions, training stops, and the word segmentation model is obtained.

6. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, when executing a program stored in memory, implements the steps of the method described in any one of claims 1-3.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method described in any one of claims 1-3.

Images