CN113824575A - Method and device for identifying fault node, computing equipment and computer storage medium - Google Patents

Method and device for identifying fault node, computing equipment and computer storage medium Download PDF

Info

Publication number
CN113824575A
CN113824575A CN202010559829.0A CN202010559829A CN113824575A CN 113824575 A CN113824575 A CN 113824575A CN 202010559829 A CN202010559829 A CN 202010559829A CN 113824575 A CN113824575 A CN 113824575A
Authority
CN
China
Prior art keywords
network
target
matrix
node
training data
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.)
Granted
Application number
CN202010559829.0A
Other languages
Chinese (zh)
Other versions
CN113824575B (en
Inventor
邢彪
张卷卷
陈维新
章淑敏
林乐轩
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China Mobile Communications Group Co Ltd
China Mobile Group Zhejiang Co Ltd
Original Assignee
China Mobile Communications Group Co Ltd
China Mobile Group Zhejiang Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by China Mobile Communications Group Co Ltd, China Mobile Group Zhejiang Co Ltd filed Critical China Mobile Communications Group Co Ltd
Priority to CN202010559829.0A priority Critical patent/CN113824575B/en
Publication of CN113824575A publication Critical patent/CN113824575A/en
Application granted granted Critical
Publication of CN113824575B publication Critical patent/CN113824575B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/06—Management of faults, events, alarms or notifications
    • H04L41/0631—Management of faults, events, alarms or notifications using root cause analysis; using analysis of correlation between notifications, alarms or events based on decision criteria, e.g. hierarchy, tree or time analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/06—Management of faults, events, alarms or notifications
    • H04L41/0677—Localisation of faults
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/12—Discovery or management of network topologies

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • General Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The embodiment of the invention relates to the technical field of communication, and discloses a method, a device, a computing device and a computer storage medium for identifying a fault node, wherein the method comprises the following steps: acquiring topological information of a target network with a fault, wherein the topological information comprises alarm characteristics of each network node in the target network and connection relations among the network nodes; obtaining a target characteristic matrix of the target network according to the alarm characteristics of the network nodes, and obtaining a target adjacent matrix of the target network according to the connection relation between the network nodes; and inputting the target characteristic matrix and the target adjacency matrix into a pre-trained node identification model to obtain a fault identification result of each network node, wherein the fault identification result is used for representing whether each network node has a fault or not. Through the mode, the embodiment of the invention realizes the fault node identification.

Description

Method and device for identifying fault node, computing equipment and computer storage medium
Technical Field
The embodiment of the invention relates to the technical field of communication, in particular to a method and a device for identifying a fault node, computing equipment and a computer storage medium.
Background
The slicing network has many nodes and complex relations, and compared with the traditional network, when the slicing network fails, the nodes influenced by the failure are difficult to define.
At present, fault nodes caused by slice faults are mainly judged by manual experience, so that the efficiency is low, the human resource cost is wasted, and errors are easy to occur.
Disclosure of Invention
In view of the foregoing problems, embodiments of the present invention provide a method, an apparatus, a computing device, and a computer storage medium for identifying a failed node, which are used to solve the problems in the prior art that efficiency is low, cost of human resources is wasted, and an error is prone to occur when a failed node is manually determined.
According to an aspect of an embodiment of the present invention, there is provided a method for identifying a failed node, the method including:
acquiring topological information of a target network with a fault, wherein the topological information comprises alarm characteristics of each network node in the target network and connection relations among the network nodes;
obtaining a target characteristic matrix of the target network according to the alarm characteristics of the network nodes, and obtaining a target adjacent matrix of the target network according to the connection relation between the network nodes;
inputting the target characteristic matrix and the target adjacency matrix into a node identification model trained in advance to obtain a fault identification result of each network node, wherein the fault identification result is used for representing whether each network node is influenced by a fault; the pre-trained node identification model is obtained by training according to a plurality of groups of training data and corresponding label vectors, and each group of training data comprises a feature matrix and an adjacent matrix corresponding to the topology information of a network with a fault.
Optionally, the obtaining a target feature matrix of the target network according to the alarm feature of each network node includes:
coding each alarm characteristic of each network node respectively to obtain a multidimensional characteristic of each network node;
and taking each dimension characteristic in the multi-dimension characteristics as a column of a matrix, and taking the multi-dimension characteristics of each network node as a row of the matrix to obtain a target characteristic matrix.
Optionally, the obtaining the target feature matrix of the target network according to the alarm feature of each network node includes:
coding each alarm characteristic of each network node respectively to obtain a multidimensional alarm characteristic of each network node;
obtaining the multidimensional attribute characteristics of each network node according to the type of the slice subnet to which each network node belongs and the isolation of each network node;
and taking each dimension in the multidimensional alarm characteristic and the multidimensional attribute characteristic as a column of a matrix, and taking the multidimensional alarm characteristic and the multidimensional attribute characteristic of each network node as a row of the matrix to obtain a target characteristic matrix.
Optionally, the obtaining a target adjacency matrix of the target network according to the connection relationship between the network nodes includes:
calculating the total number of network nodes in the target alarm topological graph, and taking the total number of the network nodes as the row number and the column number of the target adjacent matrix respectively;
if the network node ViAnd said network node VjAnd if the connection relationship exists, the element of the ith row and the jth column in the target adjacent matrix is a first numerical value, otherwise, the element of the ith row and the jth column in the target adjacent matrix is a second numerical value, wherein i and j are positive integers.
Optionally, before the obtaining of the topology information of the target network with the fault, the method further includes:
acquiring a plurality of groups of training data, wherein each group of training data comprises a characteristic matrix and an adjacent matrix corresponding to the topology information of a network with a fault;
labeling each group of training data respectively to obtain a label vector corresponding to each group of training data, wherein the number of elements in the label vector is the same as the number of network nodes contained in the topology information, and the elements in the label vector are used for indicating whether the corresponding network nodes are affected by the fault when the network fails; the labels corresponding to the network nodes affected by the fault are all first labels, the labels corresponding to the network nodes not affected by the fault are all second labels, and the first labels are different from the second labels;
and obtaining the node identification model according to the multiple groups of training data and the corresponding label vector training graph convolutional neural network model.
Optionally, the obtaining the node identification model according to the multiple sets of training data and the corresponding label vector training graph convolutional neural network model includes:
inputting the multiple groups of training data into the graph convolution neural network model to obtain a first recognition result corresponding to each group of training data in the multiple groups of training data;
calculating a loss value of a preset loss function according to the first recognition result and the label vectors corresponding to the multiple groups of training data;
adjusting the weight of the graph convolution neural network model according to the loss value, and inputting the multiple groups of training data into the graph convolution neural network model for continuous training until a preset iteration number is reached;
and when the preset iteration times are reached, taking the weight which enables the loss value of the loss function to be minimum as the weight of the graph convolution neural network model to obtain the node identification model.
Optionally, the inputting the multiple sets of training data into the graph convolution neural network model to obtain a first recognition result corresponding to each set of training data in the multiple sets of training data includes:
inputting the multiple sets of training data into the graph convolution neural network model according to a formula
Figure BDA0002545880650000031
Obtaining graph convolution results corresponding to the multiple groups of training data; wherein H1 (l)And H1 (l+1)Respectively, the graph convolution results of two adjacent graph convolution layers in the graph convolution neural network, wherein the value of L is 1-L, L is the number of the graph convolution layers in the graph convolution neural network model, H is H1 (1)＝X1，H1 (L)A graph convolution result, X, corresponding to any one of the training data groups1For the set of training data;
Figure BDA0002545880650000032
i is an identity matrix, A1A adjacency matrix in the set of training data;
Figure BDA0002545880650000033
is that
Figure BDA0002545880650000034
Degree matrix of (W)1 (1)Is the weight of the first layer map convolutional layer, σ is the nonlinear activation function of each layer map convolutional layer;
sigmoid (H) according to the formula f(L)WO+ b) obtaining a first recognition result corresponding to the multiple groups of training data, wherein WOB is the weight of the fully connected layer and b is the offset value.
According to another aspect of the embodiments of the present invention, there is provided an apparatus for identifying a failed node, the apparatus including:
the system comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring the topology information of a target network with a fault, and the topology information comprises the alarm characteristics of each network node in the target network and the connection relation between each network node;
the first determining module is used for obtaining a target characteristic matrix of the target network according to the alarm characteristics of the network nodes and obtaining a target adjacent matrix of the target network according to the connection relation between the network nodes;
the input module is used for inputting the target characteristic matrix and the target adjacency matrix into a node identification model which is trained in advance to obtain a fault identification result of each network node, and the fault identification result is used for representing whether each network node is influenced by a fault; the pre-trained node identification model is obtained by training according to a plurality of groups of training data and corresponding label vectors, and each group of training data comprises a feature matrix and an adjacent matrix corresponding to the topology information of a network with a fault.
According to still another aspect of an embodiment of the present invention, there is provided a computing device including: the system comprises a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface complete mutual communication through the communication bus;
the memory is configured to store at least one executable instruction that causes the processor to perform the operations of one of the above-described methods of failed node identification.
According to a further aspect of the embodiments of the present invention, there is provided a computer-readable storage medium having at least one executable instruction stored therein, which when executed on a computing device/apparatus, causes the computing device/apparatus to perform the operations of one of the above-mentioned methods for fault node identification.
According to the embodiment of the invention, the target characteristic matrix and the target adjacent matrix are obtained according to the obtained topological information of the target network with the fault, and the target characteristic matrix and the target adjacent matrix are input into the node identification model which is trained in advance to obtain the fault identification result of each network node. The pre-trained node identification model is obtained by training according to a large amount of training data and contains the relations between the topology information of various networks with faults and the fault influence nodes, so that the obtained node identification model can accurately determine the influenced nodes when the target network has faults. Because the trained node identification model can be regarded as a black box, after the target characteristic matrix and the target adjacency matrix are input, the node identification model can output the fault identification result of each corresponding network node under the topology information, and therefore compared with the prior art that whether each network node is influenced or not needs to be manually determined, the method and the device for identifying the network nodes can greatly improve the identification efficiency of the network nodes, and meanwhile, the human resource cost is obviously saved.
The foregoing description is only an overview of the technical solutions of the embodiments of the present invention, and the embodiments of the present invention can be implemented according to the content of the description in order to make the technical means of the embodiments of the present invention more clearly understood, and the detailed description of the present invention is provided below in order to make the foregoing and other objects, features, and advantages of the embodiments of the present invention more clearly understandable.
Drawings
The drawings are only for purposes of illustrating embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to refer to like parts throughout the drawings. In the drawings:
fig. 1 is a flowchart illustrating a method for identifying a failed node according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating a training process of a node identification model in a method for identifying a fault node according to another embodiment of the present invention;
FIG. 3 is a schematic structural diagram of a convolutional neural network model according to an embodiment of the present invention;
fig. 4 shows a functional block diagram of an apparatus for identifying a failed node according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of a computing device according to an embodiment of the present invention.
Detailed Description
Exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be embodied in various forms and should not be limited to the embodiments set forth herein.
The application scenario of the embodiment of the invention is the determination of the network node which has a fault in the fault influence range when the network has a fault. The network nodes in the embodiment of the invention are network elements in a 4G network, a 5G slice network and other communication networks which are generated newly and continuously iterated with the network in the follow-up process. The embodiment of the invention trains the convolutional neural network according to the network topology information when a fault occurs, and obtains a trained node identification model. When a network fault occurs, the input data of the node identification model can be obtained according to the topology information of the target network with the fault. Inputting the obtained input data into the node identification model to obtain the fault identification result of each network node in the target network, wherein the identification result is used for representing whether each network node is influenced by the fault. By the method, when the network fails, whether each network node is affected by the failure can be determined only by acquiring the topology information of the failed network, so that the node identification efficiency is improved, and the human resource cost is saved; in addition, the node identification model is obtained by training according to a large amount of training data, and the identification result depends on the relationship between the node characteristics learned from the large amount of training data and the connection relationship between the nodes and whether the nodes have faults or not, so that the accuracy of identifying the nodes affected by the faults is higher. In an actual service scene, the embodiment of the invention can efficiently identify whether each network node is influenced by the fault, so that precautionary measures can be taken for the nodes influenced by the fault in advance to improve the operation and maintenance efficiency of the communication network. The following describes embodiments of the present invention.
Fig. 1 shows a flowchart of a method for identifying a failed node according to an embodiment of the present invention. As shown in fig. 1, the method comprises the steps of:
step 110: and acquiring the topology information of the target network with the fault.
In this step, the target network refers to any one of the communication networks in which a failure has occurred. When the target network fails, the topology information of the target network is acquired from the network management function unit. For example, for a 5G slice network, topology information of a target network is acquired from a network slice management function (nsmf). The topology information of the target network with the fault comprises the alarm characteristics of each network node in the target network and the connection relation between each network node. The alarm characteristics include, but are not limited to, whether the network node issues an alarm.
Step 120: and obtaining a target characteristic matrix of the target network according to the alarm characteristics of the network nodes, and obtaining a target adjacent matrix of the target network according to the connection relation between the network nodes.
In this step, each alarm characteristic of each network node is encoded to obtain the multidimensional characteristic of each network node. And taking each dimension characteristic in the multi-dimension characteristics as a column of the matrix, and taking the multi-dimension characteristics of each network node as a row of the matrix to obtain a target characteristic matrix. The alarm characteristics comprise whether the network node sends an alarm or not, and if the network node sends the alarm, the alarm characteristics corresponding to the node are coded by using a numerical value a; and if the network node does not send an alarm, encoding the alarm characteristic corresponding to the node by using the value b. Preferably, the value a is 1 and the value b is 0, thereby simplifying the calculation.
The number of rows and columns of the target adjacency matrix is the same, and is the total number of network nodes in the target network. If any two network nodes ViAnd VjIf there is a connection relation, the element in the ith row and the jth column in the target adjacency matrix is the first value. If network node ViAnd VjIf there is no connection relation, the element in the ith row and the jth column in the target adjacency matrix is the second value. Wherein i and j are both positive integers. The first numerical value and the second numerical value are different values, and the specific values are not limited. Preferably, the first value is 1 and the second value is 0, so as to facilitate subsequent calculation. By V1～VNRepresenting N network nodes in the communication network, the target adjacency matrix may be represented as:
Figure BDA0002545880650000071
wherein e isN1～eNNRepresenting a network node VNAnd a network node V1～VNIn a connection relationship between, VNThe nodes themselves are necessarily not connected, and therefore, the elements on the diagonal of the target adjacency matrix are all 0.
Step 130: and inputting the target characteristic matrix and the target adjacency matrix into a pre-trained node identification model to obtain a fault identification result of each network node.
In this step, the fault identification result of each network node is used to characterize whether each network node is affected by the network fault. Inputting the target characteristic matrix and the target adjacency matrix into a node recognition model which is trained in advance, wherein the node recognition model which is trained in advance is firstly based on a formula
Figure BDA0002545880650000072
Calculating the graph convolution result of the node identification model and then according to a formula
Figure BDA0002545880650000073
And calculating to obtain the fault identification result of each network node. The identification result of each network node is a vector containing two elements which respectively represent the probability that the network node is a node affected by the fault and a node not affected by the fault. The value of L is 1-L. Wherein L is the number of layers of the node identification model, H(1)＝X，H(L)The results are identified for each network node's failure.
Figure BDA0002545880650000074
I is an identity matrix, and A is a target adjacency matrix;
Figure BDA0002545880650000075
is that
Figure BDA0002545880650000076
The degree matrix of (c) is,
Figure BDA0002545880650000077
for the weight of the convolution layer of the layer I map, σ isA nonlinear activation function for each layer;
Figure BDA0002545880650000078
is the weight of the fully-connected layer,
Figure BDA0002545880650000079
is an offset value. The pre-trained node identification model is obtained by training according to a plurality of groups of training data and corresponding label vectors, wherein each group of training data comprises a feature matrix and an adjacent matrix corresponding to the network topology information with a fault. The training of the node recognition model is explained in the following embodiments, please refer to the detailed description of the next embodiment, which is not described herein.
According to the embodiment of the invention, the target characteristic matrix and the target adjacent matrix are obtained according to the obtained topological information of the target network with the fault, and the target characteristic matrix and the target adjacent matrix are input into the node identification model which is trained in advance to obtain the fault identification result of each network node. The pre-trained node identification model is obtained by training according to a large amount of training data and contains the relations between the topology information of various networks with faults and the fault influence nodes, so that the obtained node identification model can accurately determine the influenced nodes when the target network has faults. Because the trained node identification model can be regarded as a black box, after the target characteristic matrix and the target adjacency matrix are input, the node identification model can output the fault identification result of each corresponding network node under the topology information, and therefore compared with the prior art that whether each network node is influenced or not needs to be manually determined, the method and the device for identifying the network nodes can greatly improve the identification efficiency of the network nodes, and meanwhile, the human resource cost is obviously saved.
In some embodiments, the alert feature further includes alert content. The nodes sending the alarm all correspond to alarm information, namely specific alarm content, wherein the alarm content comprises alarm categories. The codes corresponding to the alarm information can be matched in the preset alarm set. The preset alarm set is a preset database, and the corresponding relation between the alarm information and the alarm codes is stored in the database. The alarm code is a string of Arabic numerals, and each digit corresponds to a field in the alarm information. Taking the warning information as an english phrase as an example, the warning codes corresponding to the english phrases may be stored in advance in a preset database, or the warning codes corresponding to the english words may be stored in advance in a preset database. When the warning codes corresponding to the English phrases are stored in the preset database, the corresponding warning codes can be directly matched according to the target warning information after the target warning information is obtained. When alarm codes corresponding to all English words are stored in a preset database, the obtained English short sentence is segmented, each English word is a segmentation word, each segmentation word is matched with the corresponding alarm code according to the sequence, and the alarm codes corresponding to all the segmentation words are combined according to the segmentation sequence to obtain the alarm code corresponding to the alarm information.
Each encoded as a dimension of the target feature matrix. For example, the alarm code corresponding to a certain network node is 123456, a "1" indicates that the node is a node that issues an alarm, and "23456" indicates a code corresponding to specific alarm content, so that the target alarm code is six-dimensional feature data corresponding to the network node in the target feature matrix. It should be understood that in the same service scenario, the formats of the alarm messages sent by the network nodes are consistent. Therefore, the dimension of the codes of the alarm contents corresponding to the network nodes in the target feature matrix is the same. For the network nodes which do not send out the alarm, the codes of the corresponding dimensions in the corresponding alarm content are 0. By the mode, the alarm characteristic is added, so that the identification result is more reliable.
In some embodiments, the target feature matrix includes attribute features of a predetermined dimension in addition to the alarm feature. And combining the codes of the alarm characteristics with the attribute characteristics of the preset dimensionality of each network node to obtain a target characteristic matrix. The preset dimensions of each network node have the same attribute, and the types of the attribute features can be set by those skilled in the art according to the characteristics of the communication network, which is not limited in the embodiments of the present invention. For example, in a 5G slice network, the preset dimension is two dimensions, each dimension represents an attribute feature, wherein one-dimensional attribute features are sub-slice networks to which network nodes belong, and the other dimension represents the isolation of the network nodes. For example, there are three subslice networks, which are respectively a wireless network subslice, a transmission network subslice, and a core network subslice, represented by codes 1, 2, and 3, and corresponding codes can be matched according to the subslice network to which each network node belongs, where the codes are values of the network node in corresponding dimensions. The isolation of a network node is related to the sub-slice network to which it belongs. When the subsliced network is a core network subslice, isolation is used to indicate whether a network node is a shared network element or an exclusive network element. When the subsliced network is a transmission network subslice, isolation is used to indicate whether a network node is in a hard pipe or a flexible pipe. When the sub-slice network is a wireless network sub-slice, isolation is used to indicate whether a node is a dedicated cell or a shared cell, whether a wireless time-frequency resource is exclusive (hard-cut) or shared (soft-cut), and the like. The coding is performed for each isolation, and the coding mode is the same as that of the sub-slice network to which the coding mode belongs, which is not described herein again. By the method, the feature types in the target feature matrix are increased, so that the identification accuracy is improved.
By V1～VNRepresenting N network nodes in a communication network, representing the dimension of an alarm code by F, and representing the dimension of a target feature matrix by M, wherein the obtained target feature matrix can be represented as follows:
Figure BDA0002545880650000091
each row of the target feature matrix represents M-dimensional codes corresponding to one network node.
In some embodiments, before the step shown in fig. 1 is executed, a plurality of sets of training data are obtained, and each set of training data is labeled to obtain a label vector corresponding to each set of training data, and a node identification model is obtained according to the plurality of sets of training data and the label vector training graph convolutional neural network. Each set of training data comprises a feature matrix and an adjacency matrix corresponding to the topology information of the failed network. The elements in the label vector represent the fault identification results of the respective network nodes. The labels of the network nodes of the same fault identification result are the same, and the labels of the network nodes of different fault identification results are different. In one embodiment, the label of each network node may be represented by an arabic number, for example, when the fault identification result indicates that the network node is a node affected by the fault, the label of the network node is 1; when the fault identification result indicates that the network node is a node which is not affected by the fault, the label of the network node is 0.
Fig. 2 shows a training flowchart of a node identification model in a method for identifying a fault node according to another embodiment of the present invention. As shown in fig. 2, the method comprises the steps of:
step 210: and inputting the multiple groups of training data into a graph convolution neural network to obtain a first recognition result corresponding to each group of training data in the multiple groups of training data.
In this step, fig. 3 shows the structure of the atlas neural network model in the embodiment of the present invention. As shown in fig. 3, the graph convolution neural network model includes four graph convolution layers, a full connection layer and an output layer, which are sequentially connected. The output result of the graph convolution layer is the input of the fully connected layer, and the output result of the fully connected layer is the input of the output layer. The graph convolution layer is used for extracting features from the training data, and the full-connection layer is used for carrying out weighting calculation on the features extracted by each neuron in the graph convolution layer to obtain a weighting result. And the output layer is used for outputting the fault identification result of each network node through the softmax function.
Inputting multiple groups of training inputs into graph convolution neural network model, and obtaining graph convolution layer according to formula
Figure BDA0002545880650000101
Obtaining a graph convolution result corresponding to the first group of training data; wherein H1 (l)And H1 (l +1)Respectively the graph convolution results of two adjacent graph convolution layers in the graph convolution neural network,l is 1-L, wherein L is the number of graph convolution layers in the graph convolution neural network model, and H1 (1)＝X1，H1 (L)The graph convolution result corresponding to any one group of training data in the multiple groups of training data, wherein X is1For the set of training data. For example, if the graph convolution neural network model shown in fig. 3 includes four graph convolution layers, L is 4, H1 (4)Is the output result of the graph convolution layer in the graph convolution neural network model.
Figure BDA0002545880650000102
I is an identity matrix, A1For the adjacency matrix in the set of training data,
Figure BDA0002545880650000103
is that
Figure BDA0002545880650000104
Degree matrix of (W)1 (1)σ is the nonlinear activation function of each layer map convolutional layer, which is the weight of the l-th layer map convolutional layer. In an embodiment of the present invention, the number of neurons in each map convolutional layer and the corresponding activation function are not limited, in one embodiment, the number of neurons in the first map convolutional layer is 256, the number of neurons in the second map convolutional layer is 256, the number of neurons in the third map convolutional layer is 128, the number of neurons in the fourth map convolutional layer is 128, and the nonlinear activation functions in each map convolutional layer are "relu" functions.
Output result H of graph convolution layer1 (L)Inputting full connection layer, outputting result H1 (L)Weighted and then signed by the output layer according to the formula f(L)WO+ b) obtaining fault identification results corresponding to the multiple groups of training data. Wherein, WOB is the weight of the fully connected layer and b is the offset value. The number of the neurons of the output layer is the same as that of the network nodes, and the output result of each neuron is a fault identification result corresponding to one network node.
Step 220: and calculating the loss value of the preset loss function according to the first recognition result and the label vectors corresponding to the multiple groups of training data.
In this step, the predetermined loss function may be any one of the loss functions. In one embodiment, the loss function is a two-class logarithmic loss function, namely a "binary _ cross _ loss" loss function. And inputting the first recognition result of each group of training data and the corresponding label vector into the loss function to obtain a loss function value.
Step 230: and adjusting the weight of the graph convolution neural network model according to the loss value, and inputting a plurality of groups of training data into the graph convolution neural network model to continue training until the preset iteration times are reached.
In this step, the weights of the convolutional neural network model are adjusted by a gradient descent method to obtain new weights, the weights of the convolutional neural network model in step 210 are updated to the new weights, and a plurality of sets of training data are continuously input for training. And repeating the steps 210 to 220 until a preset iteration number is reached, and obtaining a group of weights which enable the loss function value to be minimum.
Step 240: and when the preset iteration times are reached, the weight which enables the loss function value to be minimum is used as the weight of the graph convolution neural network model, and the node identification model is obtained.
In this step, the weight that minimizes the loss function value is used as the weight of the final graph convolution neural network model, that is, the node identification model.
According to the embodiment of the invention, the node identification model is obtained by training the convolutional neural network model of the graph through a plurality of groups of training data, the convolutional neural network model of the graph can train the topological information of different fault networks, and the network node characteristics are different when the topological information is different. Namely, the node identification model obtained through the graph convolution neural network training can obtain the fault identification result of each corresponding network node according to different topological information. In addition, the training data in the embodiment of the invention comprises historical alarm information and historical alarm network topology, so that the node identification model obtained through mass data training comprises various historical topology information, the trained node identification model calculates the input topology information, and the obtained fault identification result of each network node is more accurate.
Fig. 4 shows a schematic structural diagram of a fault identification apparatus according to an embodiment of the present invention. As shown in fig. 4, the apparatus includes: an acquisition module 410, a first determination module 420, and an input module 430. The obtaining module 410 is configured to obtain topology information of a target network with a fault, where the topology information includes alarm characteristics of each network node in the target network and a connection relationship between each network node. The first determining module 420 is configured to obtain a target feature matrix of the target network according to the alarm feature of each network node, and obtain a target adjacent matrix of the target network according to the connection relationship between the network nodes. The input module 430 is configured to input the target feature matrix and the target adjacency matrix into a node identification model trained in advance, so as to obtain a fault identification result of each network node, where the fault identification result is used to characterize whether each network node is affected by a fault; the pre-trained node identification model is obtained by training according to a plurality of groups of training data and corresponding label vectors, and each group of training data comprises a feature matrix and an adjacent matrix corresponding to the topology information of a network with a fault.
In an optional manner, the first determining module 420 is further configured to:
coding each alarm characteristic of each network node respectively to obtain a multidimensional characteristic of each network node;
and taking each dimension characteristic in the multi-dimension characteristics as a column of a matrix, and taking the multi-dimension characteristics of each network node as a row of the matrix to obtain a target characteristic matrix.
In an optional manner, the target network is a sliced subnet, and the first determining module 420 is further configured to:
coding each alarm characteristic of each network node respectively to obtain a multidimensional alarm characteristic of each network node;
obtaining the multidimensional attribute characteristics of each network node according to the type of the slice subnet to which each network node belongs and the isolation of each network node;
and taking each dimension in the multidimensional alarm characteristic and the multidimensional attribute characteristic as a column of a matrix, and taking the multidimensional alarm characteristic and the multidimensional attribute characteristic of each network node as a row of the matrix to obtain a target characteristic matrix.
In an optional manner, the first determining module 420 is further configured to:
calculating the total number of network nodes in the target alarm topological graph, and taking the total number of the network nodes as the row number and the column number of the target adjacent matrix respectively;
if the network node ViAnd said network node VjAnd if the connection relationship exists, the element of the ith row and the jth column in the target adjacent matrix is a first numerical value, otherwise, the element of the ith row and the jth column in the target adjacent matrix is a second numerical value, wherein i and j are positive integers.
In an optional manner, the apparatus further comprises:
the first obtaining module 440 is configured to obtain a plurality of sets of training data, where each set of training data includes a feature matrix and an adjacency matrix corresponding to topology information of a failed network.
A labeling module 450, configured to label each set of training data to obtain a label vector corresponding to each set of training data, where the number of elements in the label vector is the same as the number of network nodes included in the topology information, and the elements in the label vector are used to indicate whether a corresponding network node is affected by a fault when the network fails; the labels corresponding to the network nodes affected by the fault are all first labels, the labels corresponding to the network nodes not affected by the fault are all second labels, and the first labels are different from the second labels;
and the training module 460 is configured to obtain the node identification model according to the multiple sets of training data and the corresponding label vector training graph convolutional neural network model.
In an optional manner, the training module 460 is further configured to:
inputting the multiple groups of training data into the graph convolution neural network model to obtain a first recognition result corresponding to each group of training data in the multiple groups of training data;
calculating a loss value of a preset loss function according to the first recognition result and the label vectors corresponding to the multiple groups of training data;
adjusting the weight of the graph convolution neural network model according to the loss value, and inputting the multiple groups of training data into the graph convolution neural network model for continuous training until a preset iteration number is reached;
and when the preset iteration times are reached, taking the weight which enables the loss value of the loss function to be minimum as the weight of the graph convolution neural network model to obtain the node identification model.
In an optional manner, the training module 460 is further configured to:
inputting the multiple sets of training data into the graph convolution neural network model according to a formula
Figure BDA0002545880650000131
Obtaining graph convolution results corresponding to the multiple groups of training data; wherein H1 (l)And H1 (l+1)Respectively, the graph convolution results of two adjacent graph convolution layers in the graph convolution neural network, wherein the value of L is 1-L, L is the number of the graph convolution layers in the graph convolution neural network model, H is H1 (1)＝X1，H1 (L)A graph convolution result, X, corresponding to any one of the training data groups1For the set of training data;
Figure BDA0002545880650000132
i is an identity matrix, A1A adjacency matrix in the set of training data;
Figure BDA0002545880650000133
is that
Figure BDA0002545880650000134
Degree matrix of (W)1 (1)Is the weight of the first layer map convolutional layer, σ is the nonlinear activation function of each layer map convolutional layer;
sigmoid (H) according to the formula f(L)WO+ b) obtaining a first recognition result corresponding to the multiple groups of training data, wherein WOB is the weight of the fully connected layer and b is the offset value.
According to the embodiment of the invention, the target characteristic matrix and the target adjacent matrix are obtained according to the obtained topological information of the target network with the fault, and the target characteristic matrix and the target adjacent matrix are input into the node identification model which is trained in advance to obtain the fault identification result of each network node. The pre-trained node identification model is obtained by training according to a large amount of training data and contains the relations between the topology information of various networks with faults and the fault influence nodes, so that the obtained node identification model can accurately determine the influenced nodes when the target network has faults. Because the trained node identification model can be regarded as a black box, after the target characteristic matrix and the target adjacency matrix are input, the node identification model can output the fault identification result of each corresponding network node under the topology information, and therefore compared with the prior art that whether each network node is influenced or not needs to be manually determined, the method and the device for identifying the network nodes can greatly improve the identification efficiency of the network nodes, and meanwhile, the human resource cost is obviously saved.
Fig. 5 is a schematic structural diagram of a computing device according to an embodiment of the present invention, and the specific embodiment of the present invention does not limit the specific implementation of the computing device.
As shown in fig. 5, the computing device may include: a processor (processor)502, a Communications Interface 504, a memory 506, and a communication bus 508.
Wherein: the processor 502, communication interface 504, and memory 506 communicate with one another via a communication bus 508. A communication interface 504 for communicating with network elements of other devices, such as clients or other servers. The processor 502, configured to execute the program 510, may specifically perform relevant steps in the above-described method embodiment for network node identification.
In particular, program 510 may include program code comprising computer-executable instructions.
The processor 502 may be a central processing unit CPU, or an Application Specific Integrated Circuit (ASIC), or one or more Integrated circuits configured to implement an embodiment of the present invention. The computing device includes one or more processors, which may be the same type of processor, such as one or more CPUs; or may be different types of processors such as one or more CPUs and one or more ASICs.
And a memory 506 for storing a program 510. The memory 506 may comprise high-speed RAM memory, and may also include non-volatile memory (non-volatile memory), such as at least one disk memory.
Specifically, program 510 may be invoked by processor 502 to cause a computing device to perform steps 110-130 in fig. 1, steps 210-240 in fig. 2, or to implement the functions of modules 410-460 in fig. 4.
Embodiments of the present invention provide a computer-readable storage medium, where the storage medium stores at least one executable instruction, and when the executable instruction is executed on a computing device/apparatus, the computing device/apparatus is caused to perform a method for identifying a faulty node in any of the above method embodiments.
Embodiments of the present invention provide a computer program that can be invoked by a processor to enable a computing device to perform a method for fault node identification in any of the above-described method embodiments.
Embodiments of the present invention provide a computer program product comprising a computer program stored on a computer-readable storage medium, the computer program comprising program instructions that, when run on a computer, cause the computer to perform a method of fault node identification in any of the above-mentioned method embodiments.
The algorithms or displays presented herein are not inherently related to any particular computer, virtual system, or other apparatus. Various general purpose systems may also be used with the teachings herein. The required structure for constructing such a system will be apparent from the description above. In addition, embodiments of the present invention are not directed to any particular programming language. It is appreciated that a variety of programming languages may be used to implement the teachings of the present invention as described herein, and any descriptions of specific languages are provided above to disclose the best mode of the invention.
In the description provided herein, numerous specific details are set forth. It is understood, however, that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.
Similarly, it should be appreciated that in the foregoing description of exemplary embodiments of the invention, various features of the embodiments of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the invention and aiding in the understanding of one or more of the various inventive aspects. However, the disclosed method should not be interpreted as reflecting an intention that: that the invention as claimed requires more features than are expressly recited in each claim.
Those skilled in the art will appreciate that the modules in the device in an embodiment may be adaptively changed and disposed in one or more devices different from the embodiment. The modules or units or components of the embodiments may be combined into one module or unit or component, and may be divided into a plurality of sub-modules or sub-units or sub-components. All of the features disclosed in this specification (including any accompanying claims, abstract and drawings), and all of the processes or elements of any method or apparatus so disclosed, may be combined in any combination, except combinations where at least some of such features and/or processes or elements are mutually exclusive. Each feature disclosed in this specification (including any accompanying claims, abstract and drawings) may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be able to design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention may be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by one and the same item of hardware. The usage of the words first, second and third, etcetera do not indicate any ordering. These words may be interpreted as names. The steps in the above embodiments should not be construed as limiting the order of execution unless specified otherwise.

Claims (10)

1. A method of failed node identification, the method comprising:
acquiring topological information of a target network with a fault, wherein the topological information comprises alarm characteristics of each network node in the target network and connection relations among the network nodes;
obtaining a target characteristic matrix of the target network according to the alarm characteristics of the network nodes, and obtaining a target adjacent matrix of the target network according to the connection relation between the network nodes;
inputting the target characteristic matrix and the target adjacency matrix into a node identification model trained in advance to obtain a fault identification result of each network node, wherein the fault identification result is used for representing whether each network node is influenced by a fault; the pre-trained node identification model is obtained by training according to a plurality of groups of training data and corresponding label vectors, and each group of training data comprises a feature matrix and an adjacent matrix corresponding to the topology information of a network with a fault.
2. The method according to claim 1, wherein the obtaining the target feature matrix of the target network according to the alarm features of the network nodes comprises:
coding each alarm characteristic of each network node respectively to obtain a multidimensional characteristic of each network node;
and taking each dimension characteristic in the multi-dimension characteristics as a column of a matrix, and taking the multi-dimension characteristics of each network node as a row of the matrix to obtain a target characteristic matrix.
3. The method of claim 1, wherein the target network is a sliced subnet, and the obtaining the target feature matrix of the target network according to the alarm features of the network nodes comprises:
coding each alarm characteristic of each network node respectively to obtain a multidimensional alarm characteristic of each network node;
obtaining the multidimensional attribute characteristics of each network node according to the type of the slice subnet to which each network node belongs and the isolation of each network node;
and taking each dimension in the multidimensional alarm characteristic and the multidimensional attribute characteristic as a column of a matrix, and taking the multidimensional alarm characteristic and the multidimensional attribute characteristic of each network node as a row of the matrix to obtain a target characteristic matrix.
4. The method according to claim 1, wherein the obtaining the target adjacency matrix of the target network according to the connection relationship between the network nodes comprises:
calculating the total number of network nodes in the target alarm topological graph, and taking the total number of the network nodes as the row number and the column number of the target adjacent matrix respectively;
if the network node ViAnd said network node VjIf there is a connection relation, in the target adjacency matrixAnd the element of the ith row and the jth column is a first numerical value, otherwise, the element of the ith row and the jth column in the target adjacent matrix is a second numerical value, wherein i and j are positive integers.
5. The method of claim 1, wherein prior to obtaining topology information of the failed target network, the method further comprises:
acquiring a plurality of groups of training data, wherein each group of training data comprises a characteristic matrix and an adjacent matrix corresponding to the topology information of a network with a fault;
labeling each group of training data respectively to obtain a label vector corresponding to each group of training data, wherein the number of elements in the label vector is the same as the number of network nodes contained in the topology information, and the elements in the label vector are used for indicating whether the corresponding network nodes are affected by the fault when the network fails; the labels corresponding to the network nodes affected by the fault are all first labels, the labels corresponding to the network nodes not affected by the fault are all second labels, and the first labels are different from the second labels;
and obtaining the node identification model according to the multiple groups of training data and the corresponding label vector training graph convolutional neural network model.
6. The method of claim 5, wherein obtaining the node identification model from the sets of training data and the corresponding label vector training graph convolutional neural network model comprises:
inputting the multiple groups of training data into the graph convolution neural network model to obtain a first recognition result corresponding to each group of training data in the multiple groups of training data;
calculating a loss value of a preset loss function according to the first recognition result and the label vectors corresponding to the multiple groups of training data;
adjusting the weight of the graph convolution neural network model according to the loss value, and inputting the multiple groups of training data into the graph convolution neural network model for continuous training until a preset iteration number is reached;
and when the preset iteration times are reached, taking the weight which enables the loss value of the loss function to be minimum as the weight of the graph convolution neural network model to obtain the node identification model.
7. The method of claim 5, wherein inputting the plurality of sets of training data into the convolutional neural network model to obtain a first recognition result corresponding to each set of training data in the plurality of sets of training data comprises:
inputting the multiple sets of training data into the graph convolution neural network model according to a formula
Figure FDA0002545880640000031
Obtaining graph convolution results corresponding to the multiple groups of training data; wherein H1 (l)And H1 (l+1)Respectively, the graph convolution results of two adjacent graph convolution layers in the graph convolution neural network, wherein the value of L is 1-L, L is the number of the graph convolution layers in the graph convolution neural network model, H is H1 (1)＝X1，H1 (L)A graph convolution result, X, corresponding to any one of the training data groups1For the set of training data;
Figure FDA0002545880640000032
i is an identity matrix, A1A adjacency matrix in the set of training data;
Figure FDA0002545880640000033
is that
Figure FDA0002545880640000034
Degree matrix of (W)1 (1)Is the weight of the first layer map convolutional layer, σ is the nonlinear activation function of each layer map convolutional layer;
sigmoid (H) according to the formula f(L)WO+ b) obtainingFirst recognition results corresponding to the multiple sets of training data, wherein WOB is the weight of the fully connected layer and b is the offset value.
8. An apparatus for failed node identification, the apparatus comprising:
the system comprises an acquisition module, a processing module and a processing module, wherein the acquisition module is used for acquiring the topology information of a target network with a fault, and the topology information comprises the alarm characteristics of each network node in the target network and the connection relation between each network node;
the first determining module is used for obtaining a target characteristic matrix of the target network according to the alarm characteristics of the network nodes and obtaining a target adjacent matrix of the target network according to the connection relation between the network nodes;
the input module is used for inputting the target characteristic matrix and the target adjacency matrix into a node identification model which is trained in advance to obtain a fault identification result of each network node, and the fault identification result is used for representing whether each network node is influenced by a fault; the pre-trained node identification model is obtained by training according to a plurality of groups of training data and corresponding label vectors, and each group of training data comprises a feature matrix and an adjacent matrix corresponding to the topology information of a network with a fault.
9. A computing device, comprising: the system comprises a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface complete mutual communication through the communication bus;
the memory is configured to store at least one executable instruction that causes the processor to perform the operations of a method of fault node identification as claimed in any one of claims 1 to 7.
10. A computer-readable storage medium having stored therein at least one executable instruction which, when run on a computing device/apparatus, causes the computing device/apparatus to perform operations of a method of faulty node identification as claimed in any one of claims 1-7.
CN202010559829.0A 2020-06-18 2020-06-18 Method and device for identifying fault node, computing equipment and computer storage medium Active CN113824575B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010559829.0A CN113824575B (en) 2020-06-18 2020-06-18 Method and device for identifying fault node, computing equipment and computer storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010559829.0A CN113824575B (en) 2020-06-18 2020-06-18 Method and device for identifying fault node, computing equipment and computer storage medium

Publications (2)

Publication Number Publication Date
CN113824575A true CN113824575A (en) 2021-12-21
CN113824575B CN113824575B (en) 2023-04-18

Family

ID=78911821

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010559829.0A Active CN113824575B (en) 2020-06-18 2020-06-18 Method and device for identifying fault node, computing equipment and computer storage medium

Country Status (1)

Country Link
CN (1) CN113824575B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115242467A (en) * 2022-07-05 2022-10-25 北京华顺信安科技有限公司 Network data identification method and system
CN116305461A (en) * 2023-03-13 2023-06-23 清华大学 Structure response calculation method, device, electronic equipment and storage medium

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110321268A (en) * 2019-06-12 2019-10-11 平安科技（深圳）有限公司 A kind of alarm information processing method and device
CN110943857A (en) * 2019-11-20 2020-03-31 国网湖北省电力有限公司信息通信公司 Power communication network fault analysis and positioning method based on convolutional neural network
CN110995475A (en) * 2019-11-20 2020-04-10 国网湖北省电力有限公司信息通信公司 Power communication network fault detection method based on transfer learning
CN111092752A (en) * 2019-11-27 2020-05-01 中盈优创资讯科技有限公司 Fault positioning method and device spanning multiple network slices

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110321268A (en) * 2019-06-12 2019-10-11 平安科技（深圳）有限公司 A kind of alarm information processing method and device
CN110943857A (en) * 2019-11-20 2020-03-31 国网湖北省电力有限公司信息通信公司 Power communication network fault analysis and positioning method based on convolutional neural network
CN110995475A (en) * 2019-11-20 2020-04-10 国网湖北省电力有限公司信息通信公司 Power communication network fault detection method based on transfer learning
CN111092752A (en) * 2019-11-27 2020-05-01 中盈优创资讯科技有限公司 Fault positioning method and device spanning multiple network slices

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115242467A (en) * 2022-07-05 2022-10-25 北京华顺信安科技有限公司 Network data identification method and system
CN115242467B (en) * 2022-07-05 2024-02-06 北京华顺信安科技有限公司 Network data identification method and system
CN116305461A (en) * 2023-03-13 2023-06-23 清华大学 Structure response calculation method, device, electronic equipment and storage medium
CN116305461B (en) * 2023-03-13 2023-10-13 清华大学 Structure response calculation method, device, electronic equipment and storage medium

Also Published As

Publication number Publication date
CN113824575B (en) 2023-04-18

Similar Documents

Publication Publication Date Title
CN110943857B (en) Power communication network fault analysis and positioning method based on convolutional neural network
WO2019238109A1 (en) Fault root cause analysis method and apparatus
CN112199536A (en) Cross-modality-based rapid multi-label image classification method and system
CN113824575B (en) Method and device for identifying fault node, computing equipment and computer storage medium
CN115269357A (en) Micro-service abnormity detection method based on call chain
CN113778894B (en) Method, device, equipment and storage medium for constructing test cases
CN110851654A (en) Industrial equipment fault detection and classification method based on tensor data dimension reduction
CN114048468A (en) Intrusion detection method, intrusion detection model training method, device and medium
CN113706151A (en) Data processing method and device, computer equipment and storage medium
CN114355240A (en) Power distribution network ground fault diagnosis method and device
CN115905959B (en) Defect factor-based power circuit breaker relevance fault analysis method and device
CN112560997A (en) Fault recognition model training method, fault recognition method and related device
CN113541985A (en) Internet of things fault diagnosis method, training method of model and related device
CN114553671A (en) Diagnosis method for power communication network fault alarm
CN113825165B (en) 5G slice network congestion early warning method and device based on time diagram network
CN115705279A (en) Intelligent fault early warning method and device based on index data
CN113825148B (en) Method and device for determining alarm grade of network node and computing equipment
CN116032726A (en) Fault root cause positioning model training method, device, equipment and readable storage medium
CN116306951A (en) Quantum computing method and device, electronic equipment and storage medium
CN116383398A (en) Professional field term entity word vector self-correction method, system and device
CN115952438A (en) Social platform user attribute prediction method and system, mobile device and storage medium
CN114124654B (en) Alarm merging method, device, computing equipment and computer storage medium
CN115734274A (en) Cellular network fault diagnosis method based on deep learning and knowledge graph
CN116860529A (en) Fault positioning method and device
CN114913008A (en) Decision tree-based bond value analysis method, device, equipment and storage medium

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
GR01 Patent grant
GR01 Patent grant