CN115828090A - Network fault prediction method, device, equipment and storage medium - Google Patents
Network fault prediction method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN115828090A CN115828090A CN202111092331.9A CN202111092331A CN115828090A CN 115828090 A CN115828090 A CN 115828090A CN 202111092331 A CN202111092331 A CN 202111092331A CN 115828090 A CN115828090 A CN 115828090A
- Authority
- CN
- China
- Prior art keywords
- sample data
- data set
- marked
- network
- network fault
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The invention discloses a network fault prediction method, a device, equipment and a storage medium. Wherein the method comprises the following steps: acquiring a first sample data set and a second sample data set; the first sample data set is sample data marked with network fault categories; the second sample data set is sample data which is not marked with a network fault category; marking all data in the second sample data set by using the first sample data set and combining a plurality of preset classifiers to obtain a third sample data set marked with network fault categories; taking the third sample data set as training data; inputting the training data into a prediction model for training to obtain a neural network model for predicting the network fault category; and predicting the sample set to be tested by using the neural network model to obtain the network fault category.
Description
Technical Field
The present invention relates to the field of wireless technologies, and in particular, to a method, an apparatus, a device, and a storage medium for predicting a network failure.
Background
With the development of communication technology, networks and related devices have become increasingly complex and complex. In such a complex network, various types of failures inevitably occur. After the fault occurs, the fault reason needs to be checked, the specific reason causing the fault occurrence is determined, and then the corresponding reasonable and effective means is adopted for maintenance. Because the communication network has a large number of network elements in a complex and heterogeneous network structure, when a fault occurs in the network, the fault causes are complex, so that great difficulty is caused to a manager to find the problem and investigate the problem, and a large amount of manpower and material resources are consumed.
Disclosure of Invention
In view of this, embodiments of the present invention are intended to provide a method, an apparatus, a device, and a storage medium for predicting a network failure.
The technical scheme of the embodiment of the invention is realized as follows:
at least one embodiment of the present invention provides a network failure prediction method, including:
acquiring a first sample data set and a second sample data set; the first sample data set is sample data marked with network fault categories; the second sample data set is sample data which is not marked with a network fault category;
marking all data in the second sample data set by using the first sample data set and combining a plurality of preset classifiers to obtain a third sample data set marked with network fault categories;
taking the third sample data set as training data; inputting the training data into a prediction model for training to obtain a neural network model for predicting the network fault category;
and predicting the sample set to be tested by using the neural network model to obtain the network fault category.
Furthermore, according to at least one embodiment of the present invention, the tagging all data in the second sample data set with the first sample data set in combination with a preset classifier includes:
inputting the first sample data set and the second sample data set into a first classifier and a second classifier respectively to obtain a first classification result and a second classification result;
determining sample data from the second sample data set that can be marked as a network failure category based on the first classification result and the second classification result;
updating the first sample data set by using the determined sample data; and excluding the determined sample data from the second sample data set;
inputting the updated first sample data set and the second sample data set excluding the determined sample data into the first classifier and the second classifier again;
and repeating the steps until all the data in the second sample data set are marked.
Furthermore, according to at least one embodiment of the present invention, the determining, from the second sample data set, sample data that can be marked as a network failure class based on the first classification result and the second classification result, comprises:
performing fusion processing on the first classification result and the second classification result to obtain a classification result after the fusion processing;
and determining sample data which can be marked as a network fault category from the second sample data set based on the classification result after the fusion processing.
Furthermore, according to at least one embodiment of the present invention, the determining, from the second sample data set, sample data that can be marked as a network fault category based on the classification result after the merging process includes:
calculating an evidence entropy corresponding to corresponding data aiming at each data in the classification result after the fusion processing to obtain a plurality of evidence entropies;
and marking the data corresponding to the minimum evidence entropy in the plurality of evidence entropies as sample data of the network fault category.
Furthermore, according to at least one embodiment of the present invention, the determining, from the second sample data set, sample data that can be marked as a network failure class based on the first classification result and the second classification result, includes:
calculating an evidence entropy corresponding to corresponding data for each data in the first classification result to obtain a plurality of evidence entropies; marking the data corresponding to the minimum evidence entropy in the plurality of evidence entropies as sample data with network fault categories;
calculating an evidence entropy corresponding to corresponding data for each data in the second classification result to obtain a plurality of evidence entropies; and marking the data corresponding to the minimum evidence entropy in the plurality of evidence entropies as sample data of the network fault category.
Further, in accordance with at least one embodiment of the present invention, the method further comprises:
and after marking all data in the second sample data set, taking the updated first sample data set as third sample data marked with a network fault category.
At least one embodiment of the present invention provides a network failure prediction apparatus, including:
an obtaining unit, configured to obtain a first sample data set and a second sample data set; the first sample data set is sample data marked with network fault categories; the second sample data set is sample data which is not marked with a network fault category;
the first processing unit is used for marking all data in the second sample data set by utilizing the first sample data set and combining a plurality of preset classifiers to obtain a third sample data set marked with network fault categories;
a second processing unit, configured to use the third sample data set as training data; inputting the training data into a prediction model for training to obtain a neural network model for predicting the network fault category;
and the third processing unit is used for predicting the sample set to be tested by utilizing the neural network model to obtain the network fault category.
At least one embodiment of the present invention provides a network device, including:
a communication interface for acquiring a first sample data set and a second sample data set; the first sample data set is sample data marked with network fault categories; the second sample data set is sample data which is not marked with a network fault category;
the processor is used for marking all data in the second sample data set by utilizing the first sample data set and combining a plurality of preset classifiers to obtain a third sample data set marked with network fault categories; taking the third sample data set as training data; inputting the training data into a prediction model for training to obtain a neural network model for predicting the network fault category; and predicting the sample set to be tested by using the neural network model to obtain the network fault category.
At least one embodiment of the invention provides a network device comprising a processor and a memory storing a computer program capable of running on the processor,
wherein the processor is configured to execute the steps of any one of the methods of the network device side when running the computer program.
At least one embodiment of the invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of any of the methods described above.
The network fault prediction method, the device, the equipment and the storage medium provided by the embodiment of the invention acquire a first sample data set and a second sample data set; the first sample data set is sample data marked with network fault categories; the second sample data set is sample data which is not marked with a network fault category; marking all data in the second sample data set by using the first sample data set and combining a plurality of preset classifiers to obtain a third sample data set marked with network fault categories; taking the third sample data set as training data; inputting the training data into a prediction model for training to obtain a neural network model for predicting the network fault category; and predicting the sample set to be tested by using the neural network model to obtain the network fault category. By adopting the technical scheme of the embodiment of the invention, the first sample data set marked with the network fault category and the second sample data set not marked with the network fault category are used for marking the sample data not marked with the fault category in the second sample data set, so that the first sample data set is expanded to obtain the third sample data set. Therefore, by utilizing the obtained more sufficient third sample data set, a more reliable neural network model for predicting the network fault category can be obtained through training, the sample data to be tested can be predicted subsequently, and more accurate and reliable network fault reasons can be output.
Drawings
FIG. 1 is a first flowchart illustrating a method for predicting a network failure according to the related art;
FIG. 2 is a schematic diagram of a flow chart of implementing network failure prediction in the related art;
FIG. 3 is a schematic diagram of an implementation flow of a network fault prediction method according to an embodiment of the present invention;
FIG. 4 is a flowchart illustrating a specific implementation of a network failure prediction method according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of an implementation process of using two classifiers to label sample data in a second sample data set according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of a network failure prediction apparatus according to an embodiment of the present invention;
fig. 7 is a schematic structural diagram of a network device according to an embodiment of the present invention.
Detailed Description
Before the technical solution of the embodiment of the present invention is introduced, a description is given of a related art.
In the related art, with the development of communication technology, networks and related devices become originally more numerous and complex. In such a complex network, some performance degradation may affect the normal use of the user. Various types of failures inevitably occur during operation of the network. After the fault occurs, the fault reason needs to be checked, the specific reason causing the fault occurrence is determined, and then the corresponding reasonable and effective means is adopted for maintenance. Because the number of network elements of a communication network is huge in a complex and heterogeneous network structure, when a fault occurs in the network, the cause of the fault is complex, so that great difficulty is caused for a manager to find and troubleshoot the problem, and the troubleshooting of the fault cause usually consumes a large amount of manpower and material resources, some technicians provide a classifier model by using a machine learning algorithm to find out data (marked sample data) of the fault cause through training history, and the classifier model is used for predicting unknown network faults to be detected. The method has a certain effect, however, in many cases, a large amount of data (unmarked sample data) which does not find the cause of the network fault in time often exists. This type of approach fails to model with this portion of data information. Therefore, the classifier model with better effect is obtained by learning by comprehensively utilizing the sample data with the marks and a large amount of unmarked data, so that the more accurate classification prediction of the fault reasons is obtained.
Fig. 1 is a schematic view of an implementation process for predicting a network fault in the related art, and as shown in fig. 1, in a patent with a patent publication number of CN110868731A and a name of the invention of a Long Term Evolution Voice over Long-Term Evolution (VoLTE) network fault detection method and system, the following technical features are disclosed: inputting a data source to be detected of the VoLTE network into a trained Support Vector Machine (SVM) model; and analyzing the characteristic value of the data source to be detected based on the trained SVM model, and outputting the reasons of the VoLTE network fault.
Fig. 2 is a schematic diagram illustrating an implementation process of predicting a network fault in the related art, and as shown in fig. 2, the patent with patent publication No. CN 1104786A and invented name of the method and apparatus for analyzing a cause of a VoLTE network fault based on random forest, discloses the following technical features: establishing a data sample by using a plurality of Key Performance Indexes (KPI) and Key Quality indexes (KQI, key Quality Index) of a VoLTE network, training the sample data based on a random forest method to obtain a classification model, analyzing newly input network characteristics through the classification model, and outputting wireless fault classification corresponding to the characteristics. Therefore, the fault classification corresponding to the unknown network characteristics is identified based on the known network characteristics.
However, in the related art, the scheme for classifying the network data has the following technical defects: the accuracy of the predicted fault category cannot be better ensured only by using a single classifier to predict the fault. The network fault causes are complex, and the accuracy is low when classification prediction is carried out by only using the existing marked training samples. The feature information contained in the unmarked sample data cannot be utilized, which causes a certain degree of information loss and waste of data resources.
Based on this, in the embodiment of the present invention, a first sample data set and a second sample data set are obtained; the first sample data set is sample data marked with network fault categories; the second sample data set is sample data which is not marked with a network fault category; marking all data in the second sample data set by using the first sample data set and combining a plurality of preset classifiers to obtain a third sample data set marked with network fault categories; taking the third sample data set as training data; inputting the training data into a prediction model for training to obtain a neural network model for predicting the network fault category; and predicting the sample set to be tested by using the neural network model to obtain the network fault category.
Fig. 3 is a schematic flow chart of an implementation of a network failure prediction method according to an embodiment of the present invention, as shown in fig. 3, the method includes steps 301 to 304:
step 301: acquiring a first sample data set and a second sample data set; the first sample data set is sample data marked with network fault categories; the second sample data set is sample data which is not marked with a network fault category.
It will be appreciated that sample data in the first sample data set may be referred to as marked data, i.e. sample data marked with a network failure category. Sample data in the second sample data set may be referred to as unlabeled data, i.e. sample data not labeled with a network failure category.
It is understood that the process of obtaining the first sample data set and the second sample data set may include:
firstly, collecting historical network data by using KPI and KQI indexes; wherein, KPI and KQI index include: reference Signal Receiving Power (RSRP), reference Signal Receiving Quality (RSRQ), radio Resource Control (RRC) establishment success rate, radio Access Bearer (ERAB, E-UTRAN Rad1 Access Bearer) establishment success rate, call drop rate, handover success rate, time delay, packet loss rate and jitter.
Secondly, preprocessing the acquired historical network data to determine tag data of the historical network data.
Specifically, the field information of the historical network data may be converted into vector data using one-hot (onehot) encoding, resulting in a tag of the historical network data.
For example, the field information for representing the failure cause in the historical network data is stored and converted into a network failure category vector, so that the label data of the historical network data can be obtained.
Thirdly, historical network data and corresponding tag data are formed into the first sample data set and stored locally.
Fourthly, forming the second sample data set by the historical network data without the field information representing the failure reason, and storing the second sample data set locally.
Step 302: and marking all data in the second sample data set by utilizing the first sample data set and combining a plurality of preset classifiers to obtain a third sample data set marked with network fault categories.
It can be understood that, in practical application, in order to make the failure cause of the analysis more accurate, the second sample data set not marked with the network failure category may be marked to obtain sample data marked with the network failure category, so as to expand the first sample data set, and thus, subsequently, the accuracy of the prediction of the obtained neural network model is higher by performing prediction model training using the expanded first sample data set.
Based on this, in an embodiment, the marking, by using the first sample data set and combining a preset classifier, all data in the second sample data set includes:
inputting the first sample data set and the second sample data set into a first classifier and a second classifier respectively to obtain a first classification result and a second classification result;
determining sample data from the second sample data set that can be marked as a network failure category based on the first classification result and the second classification result;
updating the first sample data set by using the determined sample data; and excluding the determined sample data from the second sample data set;
inputting the updated first sample data set and the second sample data set excluding the determined sample data into the first classifier and the second classifier again;
and repeating the steps until all the data in the second sample data set are marked.
Step 303: taking the third sample data set as training data; and inputting the training data into a prediction model for training to obtain a neural network model for predicting the network fault category.
Step 304: and predicting the sample set to be tested by using the neural network model to obtain the network fault category.
In the embodiment of the invention, the neural network model for predicting the network fault category is obtained by using the first sample data set and the third sample data set obtained by using the second sample data set, and the method has the following advantages:
(1) In the semi-supervised training process, the first sample data set marked with the fault reason and some second sample data sets which are not found out of the fault reason are used for collaborative semi-supervised learning, so that data which are not marked with the fault reason in the second sample data set are marked, the marked first sample data set is expanded to a certain extent, and a third sample data set is obtained.
(2) In the classification prediction process, a more reliable neural network model for predicting the network fault category is obtained by training by utilizing a more sufficient third sample data set obtained by semi-supervised training, so that the sample data to be tested can be predicted subsequently, and more accurate and reliable fault reasons are output.
Fig. 4 is a schematic flow chart of a specific implementation of the network failure prediction method according to the embodiment of the present invention, as shown in fig. 4, including steps 401 to 408:
step 401: acquiring a first sample data set and a second sample data set; the first sample data set is sample data marked with network fault categories; the second sample data set is sample data which is not marked with a network fault category.
Step 402: and inputting the first sample data set and the second sample data set into a first classifier and a second classifier respectively to obtain a first classification result and a second classification result.
Step 403: and determining sample data capable of being marked as a network fault class from the second sample data set based on the first classification result and the second classification result.
It is understood that, as an embodiment, considering that the classification results obtained by the first classifier and the second classifier may be different, that is, the sample data of the network fault category marked by the first classifier and the sample data of the network fault category marked by the second classifier are different, the labeling of the network fault category may be performed on the sample data in the second sample data set by combining the first classification result obtained by the first classifier and the second classification result obtained by the second classifier.
Based on this, in an embodiment, the determining, from the second sample data set, sample data that can be marked as a network failure class based on the first classification result and the second classification result includes:
performing fusion processing on the first classification result and the second classification result to obtain a classification result after the fusion processing;
and determining sample data which can be marked as a network fault category from the second sample data set based on the classification result after the fusion processing.
Specifically, the determining, from the second sample data set, sample data that can be marked as a network fault category based on the classification result after the fusion processing includes:
calculating an evidence entropy corresponding to corresponding data aiming at each data in the classification result after the fusion processing to obtain a plurality of evidence entropies;
and marking the data corresponding to the minimum evidence entropy in the plurality of evidence entropies as sample data of the network fault category.
For example, assume that there are five categories of causes of failure for sample data markers in the first sample dataset, with Θ = { A = { 1 ,A 2 ,A 3 ,A 4 ,A 5 Represents it.
Firstly, inputting sample data marked with a network fault category in the first sample data set and sample data not marked with a network fault category in the second sample data set into an evidence K neighbor algorithm classifier for training; and simultaneously, inputting the sample data marked with the network fault category in the first sample data set and the sample data not marked with the network fault category in the second sample data set into an extreme learning machine classifier for training.
Secondly, for the sample data which is not marked with the network fault category in the second sample data set, the evidence K neighbor algorithm classifier outputs a group of classification results, and the classification results are expressed by m 1. And for the sample data which is not marked with the network fault category in the second sample data set, the extreme learning machine classifier outputs a group of classification results which are expressed by m 2.
Each group of classification results may contain 6 focal elements including the full set: { A 1 ,A 2 ,A 3 ,A 4 ,A 5 Θ }. Wherein the corpus { Θ } is used to characterize the ambiguity of the ambiguity contained in the class feature.
Thirdly, fusing the two groups of classification results by using a Dempster combination rule according to the following formula.
Fourthly, calculating the evidence entropy corresponding to the corresponding data according to the following formula aiming at each data in the classification result after the fusion processing.
Evidence entropy AM = -sigma θ∈Θ BetP(θ)log 2 (BetP(θ)) (2)
And fifthly, selecting the sample data with the minimum evidence entropy value for marking, adding the sample data into the original first sample data set, and removing the sample data from the second sample data set.
It should be noted that the sample data with the minimum evidence entropy shows that the network fault category of the sample data is most clear and reliable.
And sixthly, repeatedly executing the steps by using the new first sample data set and the new second sample data set until the sample data in the second sample data set is marked.
It is to be understood that, as another embodiment, considering that the classification results obtained by the first classifier and the second classifier may be different, that is, the sample data of the network fault category marked by the first classifier and the sample data of the network fault category marked by the second classifier are different, the labeling of the network fault category may be performed on the sample data in the second sample data set by using the first classification result obtained by the first classifier and the second classification result obtained by the second classifier, respectively.
Based on this, in an embodiment, the determining, from the second sample data set, sample data that can be marked as a network failure class based on the first classification result and the second classification result includes:
calculating an evidence entropy corresponding to corresponding data for each data in the first classification result to obtain a plurality of evidence entropies; marking the data corresponding to the minimum evidence entropy in the multiple evidence entropies as sample data with network fault categories;
calculating an evidence entropy corresponding to corresponding data for each data in the second classification result to obtain a plurality of evidence entropies; and marking the data corresponding to the minimum evidence entropy in the plurality of evidence entropies as sample data of the network fault category.
For example, assume that there are five categories of causes of failure for sample data markers in the first sample dataset, with Θ = { A = { 1 ,A 2 ,A 3 ,A 4 ,A 5 Represents it.
Fig. 5 is a schematic flow chart of an implementation of marking sample data in the second sample data set by using two classifiers, and as shown in fig. 5, the method specifically includes the following steps:
firstly, inputting sample data marked with a network fault category in the first sample data set and sample data not marked with a network fault category in the second sample data set into an evidence K neighbor algorithm classifier for training; and simultaneously, inputting the sample data marked with the network fault category in the first sample data set and the sample data not marked with the network fault category in the second sample data set into an extreme learning machine classifier for training.
Secondly, for the sample data which is not marked with the network fault category in the second sample data set, the evidence K nearest neighbor algorithm classifier outputs a group of classification results, and the classification results are expressed by m 1. And for the sample data which is not marked with the network fault category in the second sample data set, the extreme learning machine classifier outputs a group of classification results which are expressed by m 2.
Each group of classification results may contain 6 focal elements including the full set: { A 1 ,A 2 ,A 3 ,A 4 ,A 5 Θ }. Wherein the corpus { Θ } is used to characterize the ambiguity of the ambiguity contained in the class feature.
Thirdly, calculating the evidence entropy corresponding to the corresponding data according to the formula (2) for each data in each group of classification results.
And fourthly, selecting the sample data with the minimum evidence entropy value for marking according to each group of classification results, adding the sample data into the original first sample data set, and removing the sample data from the second sample data set.
It should be noted that the sample data with the minimum evidence entropy shows that the network fault category of the sample data is most clear and reliable.
Fifthly, the new first sample data set and the new second sample data set are utilized, and the steps are repeatedly executed until the sample data in the second sample data set is marked.
Step 404: updating the first sample data set by using the determined sample data; and excluding the determined sample data from the second set of sample data.
Step 405: inputting the updated first sample data set and the second sample data set excluding the determined sample data into the first classifier and the second classifier again; and repeating the steps until all the data in the second sample data set are marked.
Step 406: and after marking all data in the second sample data set, taking the updated first sample data set as third sample data marked with a network fault category.
Step 407: taking the third sample data set as training data; and inputting the training data into a prediction model for training to obtain a neural network model for predicting the network fault category.
Specifically, a gcForest classifier is trained by using a labeled new training sample set, namely a third sample set, obtained through collaborative semi-supervised learning, so as to obtain a gcForest classifier model.
The deep forest is a new integrated learning method based on decision trees. The purpose of performing characterization learning on the classifier is achieved by integrating and connecting forests formed by decision trees in series, and the method has the advantages of easy training and good performance; high efficiency, expandability, support of small-scale training data and the like.
Step 408: and predicting the sample set to be tested by using the neural network model to obtain the network fault category.
In the training and predicting process, the sample data set to be tested is input into a trained gcForest classifier model, and the prediction of the network fault reason category is completed according to the obtained classification output result.
In this example, in combination with the first classifier and the second classifier, the following advantages are provided for tagging sample data in the second sample data set:
(1) The semi-supervised learning is introduced into the network fault analysis problem, so that the accuracy of network fault reason analysis can be improved.
(2) In the collaborative semi-supervised learning process, an evidence K-nearest neighbor algorithm and an extreme learning machine are utilized to model training samples, uncertainty contained in the training samples (such as ambiguity of data fault categories) is modeled, and the uncertainty is preserved in the modeling process.
(3) And the uncertainty in the model is resolved by fusing different evidence functions by using an evidence theory, so that the accuracy of model prediction is improved. In addition, the difference between different evidence functions is better comprehensively utilized through fusion, so that the unlabeled sample is more accurately predicted, more reliable labeled training sample data information is provided for subsequent classifier training, and a more reliable classifier model is obtained.
In order to implement the network failure prediction method according to the embodiment of the present invention, an embodiment of the present invention further provides a network failure prediction apparatus, which is disposed on a network device. Fig. 6 is a schematic structural diagram of a network fault prediction apparatus according to an embodiment of the present invention, and as shown in fig. 6, the apparatus includes:
an obtaining unit 61, configured to obtain a first sample data set and a second sample data set; the first sample data set is sample data marked with network fault categories; the second sample data set is sample data which is not marked with a network fault category;
the first processing unit 62 is configured to mark all data in the second sample data set by using the first sample data set and combining with a plurality of preset classifiers, so as to obtain a third sample data set marked with a network fault category;
a second processing unit 63, configured to use the third sample data set as training data; inputting the training data into a prediction model for training to obtain a neural network model for predicting the network fault category;
and the third processing unit 64 is configured to predict the sample set to be tested by using the neural network model, so as to obtain a network fault category.
In an embodiment, the first processing unit 62 is specifically configured to:
inputting the first sample data set and the second sample data set into a first classifier and a second classifier respectively to obtain a first classification result and a second classification result;
determining sample data from the second sample data set that can be marked as a network failure category based on the first classification result and the second classification result;
updating the first sample data set by using the determined sample data; and excluding the determined sample data from the second sample data set;
inputting the updated first sample data set and the second sample data set excluding the determined sample data into the first classifier and the second classifier again;
and repeating the steps until all the data in the second sample data set are marked.
In an embodiment, the first processing unit 62 is specifically configured to:
performing fusion processing on the first classification result and the second classification result to obtain a classification result after the fusion processing;
and determining sample data which can be marked as a network fault category from the second sample data set based on the classification result after the fusion processing.
In an embodiment, the first processing unit 62 is specifically configured to:
calculating an evidence entropy corresponding to corresponding data aiming at each data in the classification result after the fusion processing to obtain a plurality of evidence entropies;
and marking the data corresponding to the minimum evidence entropy in the plurality of evidence entropies as sample data of the network fault category.
In an embodiment, the first processing unit 62 is specifically configured to:
for each data in the first classification result, calculating an evidence entropy corresponding to the corresponding data to obtain a plurality of evidence entropies; marking the data corresponding to the minimum evidence entropy in the multiple evidence entropies as sample data with network fault categories;
calculating an evidence entropy corresponding to corresponding data for each data in the second classification result to obtain a plurality of evidence entropies; and marking the data corresponding to the minimum evidence entropy in the plurality of evidence entropies as sample data of the network fault category.
In an embodiment, the first processing unit 62 is further configured to:
and after marking all data in the second sample data set, taking the updated first sample data set as third sample data marked with a network fault category.
In practical applications, the obtaining unit 61 may be implemented by a communication interface in the network failure prediction apparatus. The first processing unit 62, the second processing unit 63 and the third processing unit 64 may be implemented by processors in a network failure prediction apparatus.
It should be noted that: in the network failure prediction apparatus provided in the foregoing embodiment, when performing network failure prediction, the division of each program module is merely used as an example, and in practical applications, the processing allocation may be completed by different program modules as needed, that is, the internal structure of the apparatus is divided into different program modules, so as to complete all or part of the above-described processing. In addition, the network fault prediction apparatus and the network fault prediction method provided in the foregoing embodiments belong to the same concept, and specific implementation processes thereof are described in detail in the method embodiments, and are not described herein again.
An embodiment of the present invention further provides a network device, as shown in fig. 7, including:
a communication interface 71 capable of performing information interaction with other devices;
and the processor 72 is connected with the communication interface 71 and is used for executing the method provided by one or more technical schemes of the network equipment side when running the computer program. And the computer program is stored on the memory 73.
It should be noted that: the specific processing procedures of the processor 72 and the communication interface 71 are detailed in the method embodiment, and are not described herein again.
Of course, in practice, the various components of the network device 70 are coupled together by a bus system 74. It will be appreciated that the bus system 74 is used to enable connected communication between these components. The bus system 74 includes a power bus, a control bus, and a status signal bus in addition to a data bus. For clarity of illustration, however, the various buses are labeled as bus system 74 in fig. 7.
The method disclosed in the above embodiments of the present application may be applied to the processor 72, or implemented by the processor 72. The processor 72 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware or instructions in the form of software in the processor 72. The Processor 72 may be a general purpose Processor, a Digital data Processor (DSP), or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, etc. The processor 72 may implement or perform the methods, steps, and logic blocks disclosed in the embodiments of the present application. A general purpose processor may be a microprocessor or any conventional processor or the like. The steps of the method disclosed in the embodiments of the present application may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software modules may be located in a storage medium located in the memory 73, and the processor 72 reads the information in the memory 73 and performs the steps of the aforementioned method in conjunction with its hardware.
In an exemplary embodiment, the network Device 70 may be implemented by one or more Application Specific Integrated Circuits (ASICs), DSPs, programmable Logic Devices (PLDs), complex Programmable Logic Devices (CPLDs), field Programmable Gate Arrays (FPGAs), general purpose processors, controllers, micro Controllers (MCUs), microprocessors (microprocessors), or other electronic components for performing the aforementioned methods.
It will be appreciated that the memory (memory 73) of embodiments of the present application may be either volatile memory or nonvolatile memory, and may include both volatile and nonvolatile memory. Among them, the nonvolatile Memory may be a Read Only Memory (ROM), a Programmable Read Only Memory (PROM), an Erasable Programmable Read-Only Memory (EPROM), an Electrically Erasable Programmable Read-Only Memory (EEPROM), a magnetic random access Memory (FRAM), a magnetic random access Memory (Flash Memory), a magnetic surface Memory, an optical Disc, or a Compact Disc Read-Only Memory (CD-ROM); the magnetic surface storage may be disk storage or tape storage. Volatile Memory can be Random Access Memory (RAM), which acts as external cache Memory. By way of illustration and not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), synchronous Static Random Access Memory (SSRAM), dynamic Random Access Memory (DRAM), synchronous Dynamic Random Access Memory (SDRAM), double Data Rate Synchronous Dynamic Random Access Memory (DDRSDRAM), enhanced Synchronous Dynamic Random Access Memory (ESDRAM), enhanced Synchronous Dynamic Random Access Memory (Enhanced DRAM), synchronous Dynamic Random Access Memory (SLDRAM), direct Memory (DRmb Access), and Random Access Memory (DRAM). The memories described in the embodiments of the present application are intended to comprise, without being limited to, these and any other suitable types of memory.
In an exemplary embodiment, the present invention further provides a storage medium, specifically a computer storage medium, for example, a memory storing a computer program, which is executable by the processor 72 of the network device 70 to perform the steps of the network device side method. The computer readable storage medium may be Memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash Memory, magnetic surface Memory, optical disk, or CD-ROM.
It should be noted that: "," and the like are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order.
In addition, the technical solutions described in the embodiments of the present invention may be arbitrarily combined without conflict.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention.
Claims (10)
1. A method for predicting network failure, the method comprising:
acquiring a first sample data set and a second sample data set; the first sample data set is sample data marked with network fault categories; the second sample data set is sample data which is not marked with a network fault category;
marking all data in the second sample data set by using the first sample data set and combining a plurality of preset classifiers to obtain a third sample data set marked with network fault categories;
taking the third sample data set as training data; inputting the training data into a prediction model for training to obtain a neural network model for predicting the network fault category;
and predicting the sample set to be tested by using the neural network model to obtain the network fault category.
2. The method of claim 1, wherein said tagging all data in said second sample data set with said first sample data set in conjunction with a plurality of preset classifiers comprises:
inputting the first sample data set and the second sample data set into a first classifier and a second classifier respectively to obtain a first classification result and a second classification result;
determining sample data from the second sample data set that can be marked as a network failure category based on the first classification result and the second classification result;
updating the first sample data set by using the determined sample data; and excluding the determined sample data from the second sample data set;
inputting the updated first sample data set and the second sample data set excluding the determined sample data into the first classifier and the second classifier again;
and repeating the steps until all the data in the second sample data set are marked.
3. The method of claim 2, wherein said determining sample data from said second sample data set that can be marked as a network failure category based on said first and second classification results comprises:
performing fusion processing on the first classification result and the second classification result to obtain a classification result after the fusion processing;
and determining sample data which can be marked as a network fault category from the second sample data set based on the classification result after the fusion processing.
4. The method according to claim 3, wherein the determining sample data from the second sample data set that can be marked as a network failure category based on the classification result after the fusion process comprises:
calculating an evidence entropy corresponding to corresponding data aiming at each data in the classification result after the fusion processing to obtain a plurality of evidence entropies;
and marking the data corresponding to the minimum evidence entropy in the plurality of evidence entropies as sample data of the network fault category.
5. The method of claim 2, wherein said determining sample data from said second sample data set that can be marked as a network failure category based on said first and second classification results comprises:
calculating an evidence entropy corresponding to corresponding data for each data in the first classification result to obtain a plurality of evidence entropies; marking the data corresponding to the minimum evidence entropy in the multiple evidence entropies as sample data with network fault categories;
calculating an evidence entropy corresponding to corresponding data for each data in the second classification result to obtain a plurality of evidence entropies; and marking the data corresponding to the minimum evidence entropy in the plurality of evidence entropies as sample data of the network fault category.
6. The method according to any one of claims 2 to 5, further comprising:
and after marking all data in the second sample data set, taking the updated first sample data set as third sample data marked with a network fault category.
7. A network failure prediction apparatus, comprising:
an obtaining unit, configured to obtain a first sample data set and a second sample data set; the first sample data set is sample data marked with network fault categories; the second sample data set is sample data which is not marked with a network fault category;
the first processing unit is used for marking all data in the second sample data set by utilizing the first sample data set and combining a plurality of preset classifiers to obtain a third sample data set marked with network fault categories;
a second processing unit, configured to use the third sample data set as training data; inputting the training data into a prediction model for training to obtain a neural network model for predicting the network fault category;
and the third processing unit is used for predicting the sample set to be tested by utilizing the neural network model to obtain the network fault category.
8. A network device, comprising:
a communication interface for acquiring a first sample data set and a second sample data set; the first sample data set is sample data marked with network fault categories; the second sample data set is sample data which is not marked with a network fault category;
the processor is used for marking all data in the second sample data set by utilizing the first sample data set and combining a plurality of preset classifiers to obtain a third sample data set marked with network fault categories; taking the third sample data set as training data; inputting the training data into a prediction model for training to obtain a neural network model for predicting the network fault category; and predicting the sample set to be tested by using the neural network model to obtain the network fault category.
9. A network device comprising a processor and a memory for storing a computer program capable of running on the processor,
wherein the processor is adapted to perform the steps of the method of any one of claims 1 to 6 when running the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111092331.9A CN115828090A (en) | 2021-09-17 | 2021-09-17 | Network fault prediction method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111092331.9A CN115828090A (en) | 2021-09-17 | 2021-09-17 | Network fault prediction method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115828090A true CN115828090A (en) | 2023-03-21 |
Family
ID=85515217
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111092331.9A Pending CN115828090A (en) | 2021-09-17 | 2021-09-17 | Network fault prediction method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115828090A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117555279A (en) * | 2024-01-11 | 2024-02-13 | 杭州企茏电子科技有限公司 | Remote instant monitoring system and method for dangerous chemical storage warehouse |
-
2021
- 2021-09-17 CN CN202111092331.9A patent/CN115828090A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117555279A (en) * | 2024-01-11 | 2024-02-13 | 杭州企茏电子科技有限公司 | Remote instant monitoring system and method for dangerous chemical storage warehouse |
| CN117555279B (en) * | 2024-01-11 | 2024-04-05 | 杭州企茏电子科技有限公司 | Remote instant monitoring system and method for dangerous chemical storage warehouse |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20210099336A1 (en) | Fault root cause analysis method and apparatus | |
| CN110751224B (en) | Training method of video classification model, video classification method, device and equipment | |
| EP2854053A1 (en) | Defect prediction method and device | |
| CN107809766B (en) | Method and device for generating machine learning sample for network optimization | |
| CN109635292B (en) | Work order quality inspection method and device based on machine learning algorithm | |
| CN111368636A (en) | Object classification method and device, computer equipment and storage medium | |
| CN113254254A (en) | Root cause positioning method and device of system fault, storage medium and electronic device | |
| CN114169389A (en) | Class-expanded target detection model training method and storage device | |
| CN115828090A (en) | Network fault prediction method, device, equipment and storage medium | |
| CN110716857B (en) | Test case management method, device, computer equipment and storage medium | |
| CA2997888A1 (en) | Method and system for mapping attributes of entities | |
| CN114492601A (en) | Resource classification model training method and device, electronic equipment and storage medium | |
| CN111163482A (en) | Data processing method, device and storage medium | |
| CN112164223B (en) | Intelligent traffic information processing method and device based on cloud platform | |
| CN110728229B (en) | Image processing method, device, equipment and storage medium | |
| CN107517474B (en) | Network analysis optimization method and device | |
| CN112383436A (en) | Network monitoring method and device | |
| CN116821903A (en) | Detection rule determination and malicious binary file detection method, device and medium | |
| Yang et al. | Telops: Ai-driven operations and maintenance for telecommunication networks | |
| CN114119953A (en) | Method for quickly positioning and correcting license plate, storage medium and equipment | |
| CN113361811A (en) | Method, system, device and computer readable storage medium for predicting operation state | |
| CN112101412A (en) | Intelligent classification garbage information processing method and device and server | |
| CN107305540A (en) | Address cutting recognition methods | |
| CN114697208B (en) | Configuration analysis method, system, equipment and storage medium of network element equipment | |
| Dong et al. | A region selection model to identify unknown unknowns in image datasets |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |