Fault identification method, device and program product

Abstract

The invention provides a fault identification method and device and a program product, and relates to the technical field of data processing, the method comprises the following steps: obtaining at least one first node state parameter of a network node, the network node being a network node having a fault; calculating posterior probabilities of a plurality of fault types based on a conditional probability table and the at least one first node state parameter through a preset network model, the conditional probability table comprising prior probabilities and conditional probabilities between each fault type in the plurality of fault types and different node state parameters, the preset network model is used for calculating a posterior probability of a fault type according to a prior probability, and the plurality of fault types are a plurality of fault types when the network node fails; and determining a first fault type based on the posterior probability, wherein the first fault type is the fault type with the maximum posterior probability in the plurality of fault types. According to the invention, the fault identification accuracy can be improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a fault identification method, device and program product. Background Art

[0002] Network fault identification is a key technology in communications system operation and maintenance. Related technologies typically rely on expert experience or historical data to identify network faults. However, these technologies ignore the application of prior knowledge in network fault identification. As a result, the identification results are often affected by data noise or incompleteness, resulting in low accuracy.

[0003] It can be seen that the related art has the problem of low accuracy in identifying network faults. Summary of the Invention

[0004] The embodiments of the present invention provide a fault identification method, device and program product to solve the problem of low accuracy in network fault identification in related technologies.

[0005] To solve the above problems, the present invention is achieved as follows:

[0006] In a first aspect, an embodiment of the present invention provides a fault identification method, including:

[0007] Acquiring at least one first node status parameter of a network node, where the network node is a faulty network node;

[0008] Calculating, by a preset network model, a posterior probability of multiple fault types based on a conditional probability table and the at least one first node state parameter, the conditional probability table including a priori probability and conditional probability between each of the multiple fault types and different node state parameters, the preset network model being used to calculate the posterior probability of the fault type based on the prior probability, the multiple fault types being multiple fault types of faults occurring at the network node;

[0009] A first fault type is determined based on the posterior probability, where the first fault type is the fault type with the greatest posterior probability among the multiple fault types.

[0010] In a second aspect, an embodiment of the present invention provides a model training method, comprising:

[0011] Acquire a priori knowledge set, wherein the priori knowledge set includes mapping relationships between multiple fault types and different node states, and a first initial probability of each fault type;

[0012] Based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type, the initial network model is trained to obtain the preset network model and the conditional probability table. The initial network model is a model for calculating the prior probability of each fault type based on the first initial probability.

[0013] In a third aspect, an embodiment of the present invention further provides a fault identification device, comprising:

[0014] an acquisition module, configured to acquire at least one first node status parameter of a network node, wherein the network node is a network node that has a fault;

[0015] a calculation module, configured to calculate, using a preset network model and a conditional probability table and the at least one first node state parameter, a posterior probability of a plurality of fault types, the conditional probability table including a priori probabilities and conditional probabilities between each of the plurality of fault types and different node state parameters, the preset network model being configured to calculate the posterior probability of the fault type based on the priori probabilities, the plurality of fault types being a plurality of fault types in which the network node fails;

[0016] A determination module is configured to determine a first fault type based on the posterior probability, where the first fault type is the fault type with the greatest posterior probability among the multiple fault types.

[0017] In a fourth aspect, an embodiment of the present invention further provides a model training device, comprising:

[0018] A first acquisition module is configured to acquire a priori knowledge set, wherein the priori knowledge set includes mapping relationships between multiple fault types and different node states, and a first initial probability of each fault type;

[0019] A training module is used to train the initial network model based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type, to obtain the preset network model and the conditional probability table, wherein the initial network model is a model for calculating the prior probability of each fault type based on the first initial probability.

[0020] In a fifth aspect, an embodiment of the present invention further provides an electronic device, comprising a transceiver and a processor,

[0021] The transceiver is configured to obtain at least one first node status parameter of a network node, wherein the network node is a network node that has failed;

[0022] The processor is configured to calculate, using a preset network model, a posterior probability of multiple fault types based on a conditional probability table and the at least one first node state parameter, the conditional probability table including a priori probabilities and conditional probabilities between each of the multiple fault types and different node state parameters, the preset network model being configured to calculate the posterior probability of the fault type based on the prior probabilities, the multiple fault types being multiple fault types of faults occurring at the network node;

[0023] The processor is further configured to determine a first fault type based on the posterior probability, where the first fault type is the fault type with the greatest posterior probability among the multiple fault types.

[0024] In a sixth aspect, an embodiment of the present invention further provides an electronic device, including a transceiver and a processor,

[0025] The transceiver is configured to obtain a priori knowledge set, wherein the priori knowledge set includes mapping relationships between multiple fault types and different node states, and a first initial probability of each fault type;

[0026] The processor is used to train the initial network model based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type, to obtain the preset network model and the conditional probability table, wherein the initial network model is a model for calculating the prior probability of each fault type based on the first initial probability.

[0027] In the seventh aspect, an embodiment of the present invention provides an electronic device, comprising: a processor, a memory, and a program stored in the memory and runnable on the processor, wherein when the program is executed by the processor, the steps of the fault identification method described in the first aspect are implemented, or the steps of the model training method described in the first aspect are implemented.

[0028] In an eighth aspect, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the fault identification method described in the first aspect are implemented, or the steps of the model training method described in the first aspect are implemented.

[0029] In a ninth aspect, the present invention further provides a computer program product comprising computer instructions, which, when executed by a processor, implement the steps of the fault identification method described in the first aspect, or implement the steps of the model training method described in the first aspect.

[0030] In an embodiment of the present invention, a fault identification method includes: obtaining at least one first node state parameter of a network node, wherein the network node is a network node in which a fault occurs; calculating, using a preset network model, a posterior probability of multiple fault types based on a conditional probability table and the at least one first node state parameter, wherein the conditional probability table includes a prior probability and a conditional probability between each of the multiple fault types and different node state parameters; the preset network model is used to calculate the posterior probability of the fault type based on the prior probability, wherein the multiple fault types are multiple fault types in which the network node has a fault; and determining a first fault type based on the posterior probability, wherein the first fault type is the fault type with the highest posterior probability among the multiple fault types. In this way, by calculating the posterior probabilities of multiple fault types based on the conditional probability table and the at least one first node state parameter using the preset network model, the posterior probability of the fault is calculated in combination with prior knowledge, effectively improving the accuracy of fault identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0032] Figure 1 This is a flow chart of a fault identification method provided by an embodiment of the present invention;

[0033] Figure 2 This is a flow chart of a model training method provided by an embodiment of the present invention.

[0034] Figure 3 is a structural diagram of a fault identification device provided by an embodiment of the present invention;

[0035] Figure 4 is a structural diagram of an electronic device provided by an embodiment of the present invention;

[0036] Figure 5 is a structural diagram of a model training device provided by an embodiment of the present invention;

[0037] Figure 6 This is a structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0039] See Figure 1 , Figure 1 This is a flow chart of a fault identification method provided by an embodiment of the present invention. Figure 1 As shown, the following steps are included:

[0040] Step 101: Acquire at least one first node status parameter of a network node, where the network node is a faulty network node.

[0041] The network node is a node in a network system, and the at least one first node status parameter is a node status parameter when a network node failure occurs. It should be noted that when different types of network node failures occur, the node status parameters of the network node will differ. Therefore, in the present invention, when a network node failure occurs, at least one first node status parameter of the network node is obtained to further determine the specific fault type of the network node failure.

[0042] Step 102: Calculate the posterior probabilities of multiple fault types based on a conditional probability table and the at least one first node state parameter through a preset network model, wherein the conditional probability table includes a priori probability and conditional probability between each fault type in the multiple fault types and different node state parameters, and the preset network model is used to calculate the posterior probability of the fault type based on the prior probability, and the multiple fault types are multiple fault types of faults occurring in the network node.

[0043] The aforementioned preset network model is trained to calculate posterior probabilities based on prior probabilities. Prior probabilities are probabilities calculated based on prior knowledge and can describe the node state parameters corresponding to previous fault types. Posterior probabilities are the probabilities of the current node state parameters corresponding to the fault type, and conditional probabilities are the probabilities of the fault type given the current node state parameters. By calculating the posterior probabilities of each fault type using the preset network model, the posterior probabilities of different fault types corresponding to the current fault can be calculated in combination with prior knowledge, thereby determining the corresponding first fault type.

[0044] The above conditional probability table includes the priori probability between each fault type and different node state parameters. The a posteriori probability corresponding to each fault type can be calculated using the priori probability included in the conditional probability table.

[0045] Step 103: Determine a first fault type based on the posterior probability, where the first fault type is the fault type with the highest posterior probability among the multiple fault types.

[0046] In an embodiment of the present invention, a fault identification method includes: obtaining at least one first node state parameter of a network node, wherein the network node is a network node in which a fault occurs; calculating, using a preset network model, a posterior probability of multiple fault types based on a conditional probability table and the at least one first node state parameter, wherein the conditional probability table includes a prior probability and a conditional probability between each of the multiple fault types and different node state parameters; the preset network model is used to calculate the posterior probability of the fault type based on the prior probability, wherein the multiple fault types are multiple fault types in which the network node has a fault; and determining a first fault type based on the posterior probability, wherein the first fault type is the fault type with the highest posterior probability among the multiple fault types. In this way, by calculating the posterior probabilities of multiple fault types based on the conditional probability table and the at least one first node state parameter using the preset network model, the posterior probability of the fault is calculated in combination with prior knowledge, effectively improving the accuracy of fault identification.

[0047] In some embodiments, the posterior probabilities of multiple fault types calculated based on the conditional probability table and the at least one first node state parameter using a preset network model can be expressed by the following formula:

[0048]

[0049] In the formula, r i represents the i-th fault type among multiple fault types, n represents the number of multiple fault types, E represents multiple first node state parameters, P(r i ) represents the prior probability of the i-th fault type obtained through the conditional probability table, and P(E) represents the marginal likelihood probability, P(E|r i ) represents the probability of the multiple first node status parameters occurring when the i-th type of fault occurs.

[0050] In one embodiment, the preset network model is obtained by:

[0051] Acquire a priori knowledge set, wherein the priori knowledge set includes mapping relationships between the multiple fault types and different node states, and a first initial probability of each fault type;

[0052] Based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type, the initial network model is trained to obtain the preset network model and the conditional probability table. The initial network model is a model for calculating the prior probability of each fault type based on the first initial probability.

[0053] The mapping relationships between multiple fault types and different node states included in the above prior knowledge set, as well as the first initial probability of each fault type, are prior data. The first initial probability is the probability of occurrence of different fault types in network nodes under natural conditions.

[0054] In an embodiment of the present invention, an initial network model is trained based on the mapping relationship between multiple fault types and different node states, and the first initial probability of each fault type, so as to obtain a preset network model and a conditional probability table, so that the posterior probabilities of multiple fault types can be calculated based on the conditional probability table and at least one first node state parameter through the preset network model.

[0055] The initial network model is a model used to calculate the prior probability of each fault type based on the first initial probability. The model can be an identification model with a time dimension. In some embodiments, the initial network model can be a dynamic Bayesian network model.

[0056] In one embodiment, before training the initial network model based on the mapping relationship between the multiple fault types and different node states and the first initial probability of each fault type to obtain the preset network model and the conditional probability table, the training process further includes:

[0057] Obtaining an order set of the initial network model, the order set including a plurality of Markov orders and an order probability of each Markov order;

[0058] Calculating a first entropy value corresponding to each Markov order based on the first initial probabilities of the multiple fault types;

[0059] Obtaining a second entropy value corresponding to each Markov order based on the order probability calculation;

[0060] Determining a first Markov order, where the first Markov order is the Markov order with the largest difference between the third entropy value and the second entropy value, and the third entropy value is the first entropy value corresponding to the smallest Markov order among the multiple Markov orders;

[0061] The training of the initial network model based on the mapping relationship between the multiple fault types and different node states and the first initial probability of each fault type to obtain the preset network model and the conditional probability table includes:

[0062] The initial network model of the first Markov order is trained based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type to obtain the preset network model and the conditional probability table.

[0063] It should be noted that the initial network model can be set with different Markov orders. The recognition effects of initial network models with different Markov orders are different. Therefore, before training the initial network model, it is necessary to screen out the Markov order that is suitable for the node state parameters of the network nodes.

[0064] The above order set is a set of preconfigured Markov orders, which can be added in sequence. For example, the Markov orders in the order set can be j = 1, 2, ..., m, where m is the total number of Markov orders. The order set also includes the order probability of each Markov order, which is used to calculate the second entropy value of the Markov order.

[0065] The first entropy value is used to describe the basic performance of the initial network model for each Markov order in valuation and identification. The second entropy value is used to describe the performance of the initial network model for fault identification after configuring multiple Markov orders based on order probability. The third entropy value is the first entropy value corresponding to the smallest Markov order (i.e., the first entropy value when the order is 1). In this way, the difference between the third entropy value and the second entropy value can be used to obtain the gain in recognition performance after configuring different Markov orders, and then determine the first Markov order with the best recognition performance.

[0066] The gain condition is characterized by a gain parameter, which can be expressed by the following formula:

[0067] IG=H(U)-H(U|D);

[0068] In the formula, IG is the gain parameter, H(U) is the third entropy value, and H(U|D) is the second entropy value.

[0069] In an embodiment of the present invention, an order set of the initial network model is obtained, the order set including multiple Markov orders and the order probability of each Markov order; a first entropy value corresponding to each Markov order is calculated based on the first initial probabilities of the multiple fault types; a second entropy value corresponding to each Markov order is calculated based on the order probabilities; and a first Markov order is determined, where the first Markov order is the Markov order with the largest difference between the third entropy value and the second entropy value, and the third entropy value is the first entropy value corresponding to the smallest Markov order among the multiple Markov orders. In this way, the difference between the second entropy value and the third entropy value can be used to obtain the gain in recognition effect after configuring different Markov orders, thereby determining the first Markov order with the best recognition effect. On this basis, the initial network model of the first Markov order is trained based on the mapping relationship between multiple fault types and different node states, as well as the first initial probability of each fault type, to obtain a preset network model and a conditional probability table, so that the preset network model can have a time dimension and a first Markov order.

[0070] In one embodiment, the calculating the first entropy value corresponding to each Markov order based on the first initial probabilities of the multiple fault types includes:

[0071] Calculating a first product of each fault type based on the first initial probability, the first product being a product of the first initial probability and a logarithm of the first initial probability;

[0072] The first entropy value is calculated by summing the first products of each fault type.

[0073] In an embodiment of the present invention, a first product of each fault type is calculated based on the first initial probability, where the first product is the product of the first initial probability and the logarithm of the first initial probability; the sum of the first products of each fault type is calculated as the first entropy value to obtain the first entropy value through the first initial probability calculation.

[0074] The first entropy value can be expressed by the following formula:

[0075]

[0076] In the formula, H(U) represents the first entropy value, p(u i ) represents the first initial probability of the i-th fault type, i = 1, 2, ..., n, and n is the number of multiple fault types.

[0077] In one embodiment, the calculating of the second entropy value corresponding to each Markov order based on the order probability includes:

[0078] Calculating the conditional entropy of each Markov order based on the first entropy value;

[0079] Calculating a second product corresponding to each Markov order based on the order probability, where the second product is the product of the order probability and the conditional entropy corresponding to each Markov order;

[0080] The sum of the plurality of second products is set as a second entropy value corresponding to a third Markov order, wherein the Markov order of each product in the plurality of second products is less than or equal to the third Markov order.

[0081] The above conditional entropy is the entropy value at different Markov orders. In some embodiments, the first entropy value corresponding to each Markov order can be calculated as the product of the order, and then the quotient of the product and the third entropy value is set as the conditional entropy corresponding to each Markov order.

[0082] In an embodiment of the present invention, the conditional entropy of each Markov order is calculated based on the first entropy value; the second product corresponding to each Markov order is calculated based on the order probability, and the second product is the product of the order probability and the conditional entropy corresponding to each Markov order; the sum of multiple second products is set to the second entropy value corresponding to the third Markov order, and the Markov order of each product in the multiple second products is less than or equal to the third Markov order, so as to achieve the second entropy value corresponding to each Markov order calculated according to the order probability.

[0083] Specifically, the second entropy value corresponding to each Markov order obtained by calculating the order probability can be expressed by the following formula:

[0084]

[0085] In the formula, H(U|D) represents the second entropy value, p(d j ) represents the order probability, H(U|d j ) represents the conditional entropy, and j represents the Markov order in the order set D.

[0086] In one embodiment, the mapping relationship includes an independent mapping relationship and / or an associated mapping relationship, wherein the independent mapping relationship is a mapping relationship between a fault type and a node status parameter, and the associated mapping relationship is a mapping relationship between a fault type and multiple node status parameters;

[0087] The obtaining of a priori knowledge set includes:

[0088] Acquire an initial knowledge set, the initial knowledge set including a first occurrence number of each fault type, a second occurrence number of each node state parameter, and a third occurrence number of different fault types and different node state parameters occurring simultaneously;

[0089] Constructing the independent mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously;

[0090] And / or, based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously, the association mapping relationship corresponding to each fault type is constructed.

[0091] The above-mentioned independent mapping relationship and associated mapping relationship are two mapping relationships between fault types and node status parameters. For each fault type, there may be only an independent mapping relationship or an associated mapping relationship, or both.

[0092] Among them, the independent mapping relationship and the associated mapping relationship can represent the relationship and probability between the fault type and the status parameters of different nodes. The initial network model can be trained through the independent mapping relationship and the associated mapping relationship to obtain the preset network model and conditional probability table.

[0093] In one embodiment, constructing the independent mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously includes:

[0094] calculating a first initial probability of a second fault type based on the first number of occurrences, calculating a second initial probability of a first node state parameter based on the second number of occurrences, and calculating a third initial probability of simultaneous occurrence of the second fault type and the first node state parameter based on the third number of occurrences, wherein the second fault type is one of the multiple fault types and the first node state parameter is one of the multiple node state parameters;

[0095] Calculating a first lift, where the first lift is a quotient of the third initial probability and a third product, where the third product is a sum of the first initial probability and the second initial probability;

[0096] In a case where the first promotion degree is greater than a preset promotion degree threshold, an independent mapping relationship between the second fault type and the first node state parameter is constructed.

[0097] The first initial probability is the quotient of the first number of occurrences and the total number of occurrences, the second initial probability is the quotient of the second number of occurrences and the total number of occurrences, and the third initial probability is the quotient of the third number of occurrences and the total number of occurrences.

[0098] The calculation of the first lift can be expressed by the following formula:

[0099]

[0100] In the formula, L(A→B) represents the first lift, P(A) represents the first initial probability, P(B) represents the second initial probability, and P(A∩B) represents the third initial probability.

[0101] When the first improvement degree is greater than the preset improvement degree threshold, it is considered that there is an association between the second fault type and the single first node state parameter, so an independent mapping relationship between the second fault type and the first node state parameter is constructed.

[0102] In one embodiment, constructing the association mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously includes:

[0103] Calculating a conditional probability of each node state parameter and a second fault type, the conditional probability being a quotient of a third number of occurrences of each node state parameter and the second fault type occurring simultaneously and the first number of occurrences of the second fault type;

[0104] Determine a frequent item set corresponding to the second fault type, where the frequent item set includes a plurality of second node state parameters;

[0105] calculating a vector field of the second fault type and the plurality of second node state parameters;

[0106] Calculating response values ​​corresponding to the plurality of second node state parameters based on the vector field;

[0107] Determine a plurality of third node status parameters based on the response value, wherein the plurality of third node status parameters are node status parameters whose response values ​​are greater than a set response threshold among the plurality of second node status parameters;

[0108] An association mapping relationship between the second fault type and the plurality of third node status parameters is constructed based on the conditional probability.

[0109] The above conditional probability can be calculated by the following formula:

[0110]

[0111] In the formula, N(X i ,Y j ) represents the third occurrence number of the i-th node state parameter and the j-th fault type occurring simultaneously, N(Y j ) represents the second occurrence number of the jth fault type.

[0112] The frequent item sets are used to characterize the node state parameters with a high frequency of occurrence with the second fault type. The node state parameters associated with different fault types can be determined through the frequent item sets.

[0113] The response value above is used to represent the response to the fault type when the node status parameter appears. It should be noted that the larger the response value, the higher the probability of the fault type occurring when the node status parameter appears. By setting the response threshold and response value, it is possible to determine the node status parameter that is associated with the fault type.

[0114] In an embodiment of the present invention, a conditional probability of each node state parameter and the second fault type is calculated, where the conditional probability is the quotient of the third number of occurrences of each node state parameter and the second fault type occurring simultaneously and the first number of occurrences of the second fault type; a frequent item set corresponding to the second fault type is determined, where the frequent item set includes multiple second node state parameters; a vector field of the second fault type and the multiple second node state parameters is calculated; response values ​​corresponding to the multiple second node state parameters are calculated based on the vector field; multiple third node state parameters are determined based on the response values, where the multiple third node state parameters are node state parameters among the multiple second node state parameters whose response values ​​are greater than a set response threshold; and an association mapping relationship between the second fault type and the multiple third node state parameters is established based on the conditional probability. In this way, by first determining the frequent item set to obtain multiple second node intermediate parameters, and then calculating the response value to obtain and filter multiple third node state parameters from the frequent item set, an association mapping relationship between the second fault type and the multiple third node state parameters can be established based on the conditional probability.

[0115] In some embodiments, the response values ​​corresponding to the plurality of second node state parameters are calculated based on the vector field, and may be calculated according to the following path formula:

[0116]

[0117] Where Y(t) represents the response value corresponding to the fault type at time t, α0 represents the fault response value without node state activity, X(x,t) represents the state vector including the state parameter of the second node at time t, ∫ψ(X(x,t))dx represents the radial basis function ψ applied to the state vector X(x,t) and the integration of the second node state parameter x, ε(t) represents the error term at time t simulating random disturbance, X(t) represents the parameter value of the second node state parameter at time t, and F(X(t),t) represents the vector field of the fault type flow caused by X(t) at time t. It represents the divergence of the vector field F.

[0118] In one embodiment, determining the frequent itemsets corresponding to the second fault type includes:

[0119] Calculating the mutual information (MI) and support corresponding to each node state parameter;

[0120] Calculate the score corresponding to each node state parameter based on the MI and the support;

[0121] A frequent item set corresponding to the second fault type is determined based on the score, where the frequent item set includes a plurality of second node state parameters, and the scores of the plurality of second node state parameters are greater than a set score threshold.

[0122] It should be noted that MI and support can reflect the degree of co-occurrence between each node state parameter and the fault type. By combining MI and support to calculate the score, the frequent item set corresponding to each fault type can be determined through the score.

[0123] Among them, MI and support can be calculated by the following formula:

[0124]

[0125] In the formula, I(X i ; Y j ) represents the value of the MI of the node state parameter x and the fault type y, P(x,y) represents the probability of the node state parameter x and the fault type y occurring at the same time, P(x) and P(y) represent the probability of the node state parameter x occurring alone and the fault type y occurring alone, respectively, log represents the logarithmic function, N Y Indicates the total number of fault types, S(X i ,Y j ) represents the node state parameter X i For fault type Y j support.

[0126] In some implementations, after the MI and support are calculated, the MI and support may be normalized, and then the score may be calculated based on the normalized MI and support.

[0127] The score corresponding to each node state parameter is calculated based on the MI and the support, which can be specifically expressed by the following formula:

[0128]

[0129] In the formula, W(X i ,Y j ) indicates the node status X i Relative to fault type Y j The numerical value of the rating, and They represent the normalized mutual information and normalized support respectively.

[0130] In an embodiment of the present invention, the mutual information and support corresponding to each node state parameter are calculated; the score corresponding to each node state parameter is calculated based on the MI and the support; and the frequent item set corresponding to the second fault type is determined based on the score, where the frequent item set includes multiple second node state parameters, and the scores of the multiple second node state parameters are greater than a set score threshold, thereby obtaining the frequent item set corresponding to each fault type.

[0131] See Figure 2 , Figure 2 This is a flow chart of a model training method provided by an embodiment of the present invention. Figure 2 As shown, the following steps are included:

[0132] Step 201: Acquire a priori knowledge set, wherein the priori knowledge set includes mapping relationships between multiple fault types and different node states, and a first initial probability of each fault type;

[0133] Step 202: Based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type, the initial network model is trained to obtain the preset network model and the conditional probability table. The initial network model is a model for calculating the prior probability of each fault type based on the first initial probability.

[0134] In an embodiment of the present invention, the initial network model is trained based on the mapping relationship between multiple fault types and different node states, and the first initial probability of each fault type, so as to obtain a preset network model and a conditional probability table through training, so that the posterior probabilities of multiple fault types can be calculated based on the conditional probability table and at least one first node state parameter through the preset network model.

[0135] In one embodiment, before training the initial network model based on the mapping relationship between the multiple fault types and different node states and the first initial probability of each fault type to obtain the preset network model and the conditional probability table, the method includes:

[0136] Obtaining an order set of the initial network model, the order set including a plurality of Markov orders and an order probability of each Markov order;

[0137] Calculating a first entropy value corresponding to each Markov order based on the first initial probabilities of the multiple fault types;

[0138] Obtaining a second entropy value corresponding to each Markov order based on the order probability calculation;

[0139] Determining a first Markov order, where the first Markov order is the Markov order with the largest difference between the third entropy value and the second entropy value, and the third entropy value is the first entropy value corresponding to the smallest Markov order among the multiple Markov orders;

[0140] The training of the initial network model based on the mapping relationship between the multiple fault types and different node states and the first initial probability of each fault type to obtain the preset network model and the conditional probability table includes:

[0141] The initial network model of the first Markov order is trained based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type to obtain the preset network model and the conditional probability table.

[0142] In one embodiment, the calculating the first entropy value corresponding to each Markov order based on the first initial probabilities of the multiple fault types includes:

[0143] Calculating a first product of each fault type based on the first initial probability, the first product being a product of the first initial probability and a logarithm of the first initial probability;

[0144] The first entropy value is calculated by summing the first products of each fault type.

[0145] In one embodiment, the calculating of the second entropy value corresponding to each Markov order based on the order probability includes:

[0146] Calculating the conditional entropy of each Markov order based on the first entropy value;

[0147] Calculating a second product corresponding to each Markov order based on the order probability, where the second product is the product of the order probability and the conditional entropy corresponding to each Markov order;

[0148] The sum of the plurality of second products is set as a second entropy value corresponding to a third Markov order, wherein the Markov order of each product in the plurality of second products is less than or equal to the third Markov order.

[0149] In one embodiment, the mapping relationship includes an independent mapping relationship and / or an associated mapping relationship, wherein the independent mapping relationship is a mapping relationship between a fault type and a node status parameter, and the associated mapping relationship is a mapping relationship between a fault type and multiple node status parameters;

[0150] The obtaining of a priori knowledge set includes:

[0151] Acquire an initial knowledge set, the initial knowledge set including a first occurrence number of each fault type, a second occurrence number of each node state parameter, and a third occurrence number of different fault types and different node state parameters occurring simultaneously;

[0152] Constructing the independent mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously;

[0153] And / or, based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously, the association mapping relationship corresponding to each fault type is constructed.

[0154] In one embodiment, constructing the independent mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously includes:

[0155] calculating a first initial probability of a second fault type based on the first number of occurrences, calculating a second initial probability of a first node state parameter based on the second number of occurrences, and calculating a third initial probability of simultaneous occurrence of the second fault type and the first node state parameter based on the third number of occurrences, wherein the second fault type is one of the multiple fault types and the first node state parameter is one of the multiple node state parameters;

[0156] Calculating a first lift, where the first lift is a quotient of the third initial probability and a third product, where the third product is a sum of the first initial probability and the second initial probability;

[0157] In a case where the first promotion degree is greater than a preset promotion degree threshold, an independent mapping relationship between the second fault type and the first node state parameter is constructed.

[0158] In one embodiment, constructing the association mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously includes:

[0159] Calculating a conditional probability of each node state parameter and a second fault type, the conditional probability being a quotient of a third number of occurrences of each node state parameter and the second fault type occurring simultaneously and the first number of occurrences of the second fault type;

[0160] Determine a frequent item set corresponding to the second fault type, where the frequent item set includes a plurality of second node state parameters;

[0161] calculating a vector field of the second fault type and the plurality of second node state parameters;

[0162] Calculating response values ​​corresponding to the plurality of second node state parameters based on the vector field;

[0163] Determine a plurality of third node status parameters based on the response value, wherein the plurality of third node status parameters are node status parameters whose response values ​​are greater than a set response threshold among the plurality of second node status parameters;

[0164] An association mapping relationship between the second fault type and the plurality of third node status parameters is constructed based on the conditional probability.

[0165] In one embodiment, determining the frequent itemsets corresponding to the second fault type includes:

[0166] Calculate the mutual information MI and support corresponding to each node state parameter;

[0167] Calculate the score corresponding to each node state parameter based on the MI and the support;

[0168] A frequent item set corresponding to the second fault type is determined based on the score, where the frequent item set includes a plurality of second node state parameters, and the scores of the plurality of second node state parameters are greater than a set score threshold.

[0169] See Figure 3 , Figure 3 This is a structural diagram of a fault identification device provided by an embodiment of the present invention. Figure 3 As shown, the fault identification device 300 includes:

[0170] An acquisition module 301 is configured to acquire at least one first node status parameter of a network node, where the network node is a faulty network node;

[0171] a calculation module 302, configured to calculate, using a preset network model, a posterior probability of multiple fault types based on a conditional probability table and the at least one first node state parameter, the conditional probability table including a priori probabilities and conditional probabilities between each of the multiple fault types and different node state parameters, the preset network model being configured to calculate the posterior probability of the fault type based on the prior probabilities, the multiple fault types being multiple fault types that may occur at the network node;

[0172] The determination module 303 is configured to determine a first fault type based on the posterior probability, where the first fault type is the fault type with the highest posterior probability among the multiple fault types.

[0173] In one embodiment, the preset network model is obtained by:

[0174] Acquire a priori knowledge set, wherein the priori knowledge set includes mapping relationships between the multiple fault types and different node states, and a first initial probability of each fault type;

[0175] Based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type, the initial network model is trained to obtain the preset network model and the conditional probability table. The initial network model is a model for calculating the prior probability of each fault type based on the first initial probability.

[0176] In one embodiment, before training the initial network model based on the mapping relationship between the multiple fault types and different node states and the first initial probability of each fault type to obtain the preset network model and the conditional probability table, the training process further includes:

[0177] Obtaining an order set of the initial network model, the order set including a plurality of Markov orders and an order probability of each Markov order;

[0178] Calculating a first entropy value corresponding to each Markov order based on the first initial probabilities of the multiple fault types;

[0179] Obtaining a second entropy value corresponding to each Markov order based on the order probability calculation;

[0180] Determining a first Markov order, where the first Markov order is the Markov order with the largest difference between the third entropy value and the second entropy value, and the third entropy value is the first entropy value corresponding to the smallest Markov order among the multiple Markov orders;

[0181] The training of the initial network model based on the mapping relationship between the multiple fault types and different node states and the first initial probability of each fault type to obtain the preset network model and the conditional probability table includes:

[0182] The initial network model of the first Markov order is trained based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type to obtain the preset network model and the conditional probability table.

[0183] In one embodiment, the calculating the first entropy value corresponding to each Markov order based on the first initial probabilities of the multiple fault types includes:

[0184] Calculating a first product of each fault type based on the first initial probability, the first product being a product of the first initial probability and a logarithm of the first initial probability;

[0185] The first entropy value is calculated by summing the first products of each fault type.

[0186] In one embodiment, the calculating of the second entropy value corresponding to each Markov order based on the order probability includes:

[0187] Calculating the conditional entropy of each Markov order based on the first entropy value;

[0188] Calculating a second product corresponding to each Markov order based on the order probability, where the second product is the product of the order probability and the conditional entropy corresponding to each Markov order;

[0189] The sum of the plurality of second products is set as a second entropy value corresponding to a third Markov order, wherein the Markov order of each product in the plurality of second products is less than or equal to the third Markov order.

[0190] In one embodiment, the mapping relationship includes an independent mapping relationship and / or an associated mapping relationship, wherein the independent mapping relationship is a mapping relationship between a fault type and a node status parameter, and the associated mapping relationship is a mapping relationship between a fault type and multiple node status parameters;

[0191] The obtaining of a priori knowledge set includes:

[0192] Acquire an initial knowledge set, the initial knowledge set including a first occurrence number of each fault type, a second occurrence number of each node state parameter, and a third occurrence number of different fault types and different node state parameters occurring simultaneously;

[0193] Constructing the independent mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously;

[0194] And / or, based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously, the association mapping relationship corresponding to each fault type is constructed.

[0195] In one embodiment, constructing the independent mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously includes:

[0196] calculating a first initial probability of a second fault type based on the first number of occurrences, calculating a second initial probability of a first node state parameter based on the second number of occurrences, and calculating a third initial probability of simultaneous occurrence of the second fault type and the first node state parameter based on the third number of occurrences, wherein the second fault type is one of the multiple fault types and the first node state parameter is one of the multiple node state parameters;

[0197] Calculating a first lift, where the first lift is a quotient of the third initial probability and a third product, where the third product is a sum of the first initial probability and the second initial probability;

[0198] In a case where the first promotion degree is greater than a preset promotion degree threshold, an independent mapping relationship between the second fault type and the first node state parameter is constructed.

[0199] In one embodiment, constructing the association mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously includes:

[0200] Calculating a conditional probability of each node state parameter and a second fault type, the conditional probability being a quotient of a third number of occurrences of each node state parameter and the second fault type occurring simultaneously and the first number of occurrences of the second fault type;

[0201] Determine a frequent item set corresponding to the second fault type, where the frequent item set includes a plurality of second node state parameters;

[0202] calculating a vector field of the second fault type and the plurality of second node state parameters;

[0203] Calculating response values ​​corresponding to the plurality of second node state parameters based on the vector field;

[0204] Determine a plurality of third node status parameters based on the response value, wherein the plurality of third node status parameters are node status parameters whose response values ​​are greater than a set response threshold among the plurality of second node status parameters;

[0205] An association mapping relationship between the second fault type and the plurality of third node status parameters is constructed based on the conditional probability.

[0206] In one embodiment, determining the frequent itemsets corresponding to the second fault type includes:

[0207] Calculate the mutual information MI and support corresponding to each node state parameter;

[0208] Calculate the score corresponding to each node state parameter based on the MI and the support;

[0209] A frequent item set corresponding to the second fault type is determined based on the score, where the frequent item set includes a plurality of second node state parameters, and the scores of the plurality of second node state parameters are greater than a set score threshold.

[0210] The fault identification device provided by the embodiment of the present invention is capable of achieving the above Figure 1 The various processes and technical features of the various embodiments of the fault identification method shown correspond to each other and can achieve the same technical effects. To avoid repetition, they will not be described here.

[0211] It should be noted that the fault identification device in the embodiment of the present invention may be a device, or a component, integrated circuit, or chip in an electronic device.

[0212] An embodiment of the present invention further provides an electronic device, comprising: a processor, a memory, and a program stored in the memory and operable on the processor, wherein the program is executed by the processor to implement the above-mentioned Figure 1The various processes of the fault identification method embodiment shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0213] For details, see Figure 4 As shown, an embodiment of the present invention further provides an electronic device, including a bus 401 , a transceiver 402 , an antenna 403 , a bus interface 404 , a processor 405 and a memory 406 .

[0214] The transceiver 402 is configured to obtain at least one first node status parameter of a network node, wherein the network node is a faulty network node;

[0215] The processor 405 is configured to calculate, using a preset network model, a posterior probability of multiple fault types based on a conditional probability table and the at least one first node state parameter, the conditional probability table including a priori probabilities and conditional probabilities between each of the multiple fault types and different node state parameters, the preset network model being configured to calculate the posterior probability of the fault type based on the prior probabilities, the multiple fault types being multiple fault types of faults occurring at the network node;

[0216] The processor 405 is further configured to determine a first fault type based on the posterior probability, where the first fault type is the fault type with the highest posterior probability among the multiple fault types.

[0217] In one embodiment, the preset network model is obtained by:

[0218] Acquire a priori knowledge set, wherein the priori knowledge set includes mapping relationships between the multiple fault types and different node states, and a first initial probability of each fault type;

[0219] Based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type, the initial network model is trained to obtain the preset network model and the conditional probability table. The initial network model is a model for calculating the prior probability of each fault type based on the first initial probability.

[0220] In one embodiment, before training the initial network model based on the mapping relationship between the multiple fault types and different node states and the first initial probability of each fault type to obtain the preset network model and the conditional probability table, the training process further includes:

[0221] Obtaining an order set of the initial network model, the order set including a plurality of Markov orders and an order probability of each Markov order;

[0222] Calculating a first entropy value corresponding to each Markov order based on the first initial probabilities of the multiple fault types;

[0223] Obtaining a second entropy value corresponding to each Markov order based on the order probability calculation;

[0224] Determining a first Markov order, where the first Markov order is the Markov order with the largest difference between the third entropy value and the second entropy value, and the third entropy value is the first entropy value corresponding to the smallest Markov order among the multiple Markov orders;

[0225] The training of the initial network model based on the mapping relationship between the multiple fault types and different node states and the first initial probability of each fault type to obtain the preset network model and the conditional probability table includes:

[0226] The initial network model of the first Markov order is trained based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type to obtain the preset network model and the conditional probability table.

[0227] In one embodiment, the calculating the first entropy value corresponding to each Markov order based on the first initial probabilities of the multiple fault types includes:

[0228] Calculating a first product of each fault type based on the first initial probability, the first product being a product of the first initial probability and a logarithm of the first initial probability;

[0229] The first entropy value is calculated by summing the first products of each fault type.

[0230] In one embodiment, the calculating of the second entropy value corresponding to each Markov order based on the order probability includes:

[0231] Calculating the conditional entropy of each Markov order based on the first entropy value;

[0232] Calculating a second product corresponding to each Markov order based on the order probability, where the second product is the product of the order probability and the conditional entropy corresponding to each Markov order;

[0233] The sum of the plurality of second products is set as a second entropy value corresponding to a third Markov order, wherein the Markov order of each product in the plurality of second products is less than or equal to the third Markov order.

[0234] In one embodiment, the mapping relationship includes an independent mapping relationship and / or an associated mapping relationship, wherein the independent mapping relationship is a mapping relationship between a fault type and a node status parameter, and the associated mapping relationship is a mapping relationship between a fault type and multiple node status parameters;

[0235] The obtaining of a priori knowledge set includes:

[0236] Acquire an initial knowledge set, the initial knowledge set including a first occurrence number of each fault type, a second occurrence number of each node state parameter, and a third occurrence number of different fault types and different node state parameters occurring simultaneously;

[0237] Constructing the independent mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously;

[0238] And / or, based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously, the association mapping relationship corresponding to each fault type is constructed.

[0239] In one embodiment, constructing the independent mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously includes:

[0240] calculating a first initial probability of a second fault type based on the first number of occurrences, calculating a second initial probability of a first node state parameter based on the second number of occurrences, and calculating a third initial probability of simultaneous occurrence of the second fault type and the first node state parameter based on the third number of occurrences, wherein the second fault type is one of the multiple fault types and the first node state parameter is one of the multiple node state parameters;

[0241] Calculating a first lift, where the first lift is a quotient of the third initial probability and a third product, where the third product is a sum of the first initial probability and the second initial probability;

[0242] In a case where the first promotion degree is greater than a preset promotion degree threshold, an independent mapping relationship between the second fault type and the first node state parameter is constructed.

[0243] In one embodiment, constructing the association mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously includes:

[0244] Calculating a conditional probability of each node state parameter and a second fault type, the conditional probability being a quotient of a third number of occurrences of each node state parameter and the second fault type occurring simultaneously and the first number of occurrences of the second fault type;

[0245] Determine a frequent item set corresponding to the second fault type, where the frequent item set includes a plurality of second node state parameters;

[0246] calculating a vector field of the second fault type and the plurality of second node state parameters;

[0247] Calculating response values ​​corresponding to the plurality of second node state parameters based on the vector field;

[0248] Determine a plurality of third node status parameters based on the response value, wherein the plurality of third node status parameters are node status parameters whose response values ​​are greater than a set response threshold among the plurality of second node status parameters;

[0249] An association mapping relationship between the second fault type and the plurality of third node status parameters is constructed based on the conditional probability.

[0250] In one embodiment, determining the frequent itemsets corresponding to the second fault type includes:

[0251] Calculate the mutual information MI and support corresponding to each node state parameter;

[0252] Calculate the score corresponding to each node state parameter based on the MI and the support;

[0253] A frequent item set corresponding to the second fault type is determined based on the score, where the frequent item set includes a plurality of second node state parameters, and the scores of the plurality of second node state parameters are greater than a set score threshold.

[0254] exist Figure 4In the embodiment, the bus architecture (represented by bus 401) is shown. Bus 401 may include any number of interconnected buses and bridges. Bus 401 links together various circuits including one or more processors represented by processor 405 and memory represented by memory 406. Bus 401 may also link together various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and are therefore not described further herein. Bus interface 404 provides an interface between bus 401 and transceiver 402. Transceiver 402 may be one element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices on a transmission medium. Data processed by processor 405 is transmitted on a wireless medium via antenna 403. Furthermore, antenna 403 receives data and transmits the data to processor 405.

[0255] Processor 405 is responsible for managing bus 401 and general processing, and may also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory 406 may be used to store data used by processor 405 when performing operations.

[0256] Optionally, the processor 405 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or a graphics processing unit (GPU).

[0257] See Figure 5 , Figure 5 This is a structural diagram of a model training device provided by an embodiment of the present invention. Figure 5 As shown, the fault identification device 500 includes:

[0258] A first acquisition module 501 is configured to acquire a priori knowledge set, wherein the priori knowledge set includes mapping relationships between multiple fault types and different node states, and a first initial probability of each fault type;

[0259] The training module 502 is used to train the initial network model based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type, to obtain the preset network model and the conditional probability table. The initial network model is a model that calculates the prior probability of each fault type based on the first initial probability.

[0260] In one embodiment, the fault identification device 500 further includes:

[0261] A second acquisition module is used to obtain an order set of the initial network model, where the order set includes multiple Markov orders and an order probability of each Markov order;

[0262] A first calculation module, configured to calculate a first entropy value corresponding to each Markov order based on the first initial probabilities of the multiple fault types;

[0263] A second calculation module is used to calculate and obtain a second entropy value corresponding to each Markov order based on the order probability;

[0264] a determination module, configured to determine a first Markov order, where the first Markov order is the Markov order having the largest difference between the third entropy value and the second entropy value, and the third entropy value is the first entropy value corresponding to the smallest Markov order among the multiple Markov orders;

[0265] The training module 502 includes:

[0266] A training unit is used to train the initial network model of the first Markov order based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type, to obtain the preset network model and the conditional probability table.

[0267] In one embodiment, the first calculation module includes:

[0268] a first calculation submodule, configured to calculate a first product of each fault type based on the first initial probability, where the first product is a product of the first initial probability and a logarithm of the first initial probability;

[0269] The second calculation submodule is configured to convert the sum of the first products of each fault type into the first entropy value.

[0270] In one embodiment, the second calculation module includes:

[0271] A third calculation submodule, configured to calculate the conditional entropy of each Markov order based on the first entropy value;

[0272] a fourth calculation submodule, configured to calculate a second product corresponding to each Markov order based on the order probability, where the second product is a product of the order probability and the conditional entropy corresponding to each Markov order;

[0273] The fifth calculation submodule is configured to set the sum of a plurality of second products as a second entropy value corresponding to a third Markov order, wherein the Markov order of each of the plurality of second products is less than or equal to the third Markov order.

[0274] In one embodiment, the mapping relationship includes an independent mapping relationship and / or an associated mapping relationship, wherein the independent mapping relationship is a mapping relationship between a fault type and a node status parameter, and the associated mapping relationship is a mapping relationship between a fault type and multiple node status parameters;

[0275] The first acquisition module 501 includes:

[0276] an acquisition submodule, configured to acquire an initial knowledge set, the initial knowledge set including a first occurrence number of each fault type, a second occurrence number of each node state parameter, and a third occurrence number of simultaneous occurrences of different fault types and different node state parameters;

[0277] A first construction submodule is configured to construct the independent mapping relationship corresponding to each fault type based on a first occurrence number of each fault type, a second occurrence number of each node status parameter, and a third occurrence number of different fault types and different node status parameters occurring simultaneously;

[0278] The second construction submodule is used to construct the association mapping relationship corresponding to each fault type based on the first number of occurrences of each fault type, the second number of occurrences of each node status parameter, and the third number of occurrences of different fault types and different node status parameters occurring simultaneously.

[0279] In one embodiment, the first building block includes:

[0280] a first calculation unit, configured to calculate a first initial probability of a second fault type based on the first number of occurrences, calculate a second initial probability of a first node state parameter based on the second number of occurrences, and calculate a third initial probability of simultaneous occurrence of the second fault type and the first node state parameter based on the third number of occurrences, wherein the second fault type is one of the multiple fault types and the first node state parameter is one of the multiple node state parameters;

[0281] a second calculation unit, configured to calculate a first lift, where the first lift is a quotient of the third initial probability and a third product, where the third product is a sum of the first initial probability and the second initial probability;

[0282] The first constructing unit is configured to construct an independent mapping relationship between the second fault type and the first node state parameter when the first improvement degree is greater than a preset improvement degree threshold.

[0283] In one embodiment, the second building submodule includes:

[0284] a third calculating unit, configured to calculate a conditional probability of each node state parameter and a second fault type, the conditional probability being a quotient of a third number of occurrences of each node state parameter and the second fault type occurring simultaneously and the first number of occurrences of the second fault type;

[0285] A first determining unit is configured to determine a frequent item set corresponding to the second fault type, wherein the frequent item set includes a plurality of second node state parameters;

[0286] a fourth calculation unit, configured to calculate a vector field of the second fault type and the plurality of second node state parameters;

[0287] a fifth calculation unit, configured to calculate response values ​​corresponding to the plurality of second node state parameters based on the vector field;

[0288] a second determining unit, configured to determine a plurality of third node status parameters based on the response value, wherein the plurality of third node status parameters are node status parameters having response values ​​greater than a set response threshold among the plurality of second node status parameters;

[0289] The second construction unit is configured to construct an association mapping relationship between the second fault type and the plurality of third node status parameters based on the conditional probability.

[0290] In one embodiment, the first determining unit includes:

[0291] A first calculation subunit is used to calculate the mutual information MI and support corresponding to each node state parameter;

[0292] A second calculation subunit, configured to calculate a score corresponding to each node state parameter based on the MI and the support;

[0293] A determining subunit is configured to determine a frequent item set corresponding to the second fault type based on the score, wherein the frequent item set includes a plurality of second node state parameters, and the scores of the plurality of second node state parameters are greater than a set score threshold.

[0294] The model training device provided by the embodiment of the present invention is capable of achieving the above Figure 2 The various processes and technical features of each embodiment of the model training method shown correspond one to one and can achieve the same technical effect. To avoid repetition, they will not be described here.

[0295] It should be noted that the model training device in the embodiment of the present invention can be a device, or a component, integrated circuit, or chip in an electronic device.

[0296] An embodiment of the present invention further provides an electronic device, comprising: a processor, a memory, and a program stored in the memory and operable on the processor, wherein the program is executed by the processor to implement the above-mentioned Figure 2 The various processes of the model training method embodiment shown can achieve the same technical effect. To avoid repetition, they will not be described here.

[0297] For details, see Figure 6 As shown, an embodiment of the present invention further provides an electronic device, including a bus 601 , a transceiver 602 , an antenna 603 , a bus interface 604 , a processor 605 and a memory 606 .

[0298] The transceiver 602 is configured to obtain a priori knowledge set, wherein the priori knowledge set includes mapping relationships between multiple fault types and different node states, and a first initial probability of each fault type;

[0299] The processor 605 is used to train the initial network model based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type, to obtain the preset network model and the conditional probability table, where the initial network model is a model that calculates the prior probability of each fault type based on the first initial probability.

[0300] In one embodiment, the transceiver 602 is further configured to obtain an order set of the initial network model, wherein the order set includes a plurality of Markov orders and an order probability of each Markov order;

[0301] The processor 605 is configured to calculate a first entropy value corresponding to each Markov order based on the first initial probabilities of the multiple fault types;

[0302] The processor 605 is configured to calculate and obtain a second entropy value corresponding to each Markov order based on the order probability;

[0303] The processor 605 is configured to determine a first Markov order, where the first Markov order is the Markov order having the largest difference between the third entropy value and the second entropy value, and the third entropy value is the first entropy value corresponding to the smallest Markov order among the multiple Markov orders;

[0304] The training of the initial network model based on the mapping relationship between the multiple fault types and different node states and the first initial probability of each fault type to obtain the preset network model and the conditional probability table includes:

[0305] The initial network model of the first Markov order is trained based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type to obtain the preset network model and the conditional probability table.

[0306] In one embodiment, the calculating the first entropy value corresponding to each Markov order based on the first initial probabilities of the multiple fault types includes:

[0307] Calculating a first product of each fault type based on the first initial probability, the first product being a product of the first initial probability and a logarithm of the first initial probability;

[0308] The first entropy value is calculated by summing the first products of each fault type.

[0309] In one embodiment, the calculating of the second entropy value corresponding to each Markov order based on the order probability includes:

[0310] Calculating the conditional entropy of each Markov order based on the first entropy value;

[0311] Calculating a second product corresponding to each Markov order based on the order probability, where the second product is the product of the order probability and the conditional entropy corresponding to each Markov order;

[0312] The sum of the plurality of second products is set as a second entropy value corresponding to a third Markov order, wherein the Markov order of each product in the plurality of second products is less than or equal to the third Markov order.

[0313] In one embodiment, the mapping relationship includes an independent mapping relationship and / or an associated mapping relationship, wherein the independent mapping relationship is a mapping relationship between a fault type and a node status parameter, and the associated mapping relationship is a mapping relationship between a fault type and multiple node status parameters;

[0314] The obtaining of a priori knowledge set includes:

[0315] Acquire an initial knowledge set, the initial knowledge set including a first occurrence number of each fault type, a second occurrence number of each node state parameter, and a third occurrence number of different fault types and different node state parameters occurring simultaneously;

[0316] Constructing the independent mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously;

[0317] And / or, based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously, the association mapping relationship corresponding to each fault type is constructed.

[0318] In one embodiment, constructing the independent mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously includes:

[0319] calculating a first initial probability of a second fault type based on the first number of occurrences, calculating a second initial probability of a first node state parameter based on the second number of occurrences, and calculating a third initial probability of simultaneous occurrence of the second fault type and the first node state parameter based on the third number of occurrences, wherein the second fault type is one of the multiple fault types and the first node state parameter is one of the multiple node state parameters;

[0320] Calculating a first lift, where the first lift is a quotient of the third initial probability and a third product, where the third product is a sum of the first initial probability and the second initial probability;

[0321] In a case where the first promotion degree is greater than a preset promotion degree threshold, an independent mapping relationship between the second fault type and the first node state parameter is constructed.

[0322] In one embodiment, constructing the association mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously includes:

[0323] Calculating a conditional probability of each node state parameter and a second fault type, the conditional probability being a quotient of a third number of occurrences of each node state parameter and the second fault type occurring simultaneously and the first number of occurrences of the second fault type;

[0324] Determine a frequent item set corresponding to the second fault type, where the frequent item set includes a plurality of second node state parameters;

[0325] calculating a vector field of the second fault type and the plurality of second node state parameters;

[0326] Calculating response values ​​corresponding to the plurality of second node state parameters based on the vector field;

[0327] Determine a plurality of third node status parameters based on the response value, wherein the plurality of third node status parameters are node status parameters whose response values ​​are greater than a set response threshold among the plurality of second node status parameters;

[0328] An association mapping relationship between the second fault type and the plurality of third node status parameters is constructed based on the conditional probability.

[0329] In one embodiment, determining the frequent itemsets corresponding to the second fault type includes:

[0330] Calculate the mutual information MI and support corresponding to each node state parameter;

[0331] Calculate the score corresponding to each node state parameter based on the MI and the support;

[0332] A frequent item set corresponding to the second fault type is determined based on the score, where the frequent item set includes a plurality of second node state parameters, and the scores of the plurality of second node state parameters are greater than a set score threshold.

[0333] exist Figure 6 In the embodiment, the bus architecture (represented by bus 601) is shown. Bus 601 may include any number of interconnected buses and bridges. Bus 601 links together various circuits including one or more processors represented by processor 605 and memory represented by memory 606. Bus 601 may also link together various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and are therefore not described further herein. Bus interface 604 provides an interface between bus 601 and transceiver 602. Transceiver 602 may be one element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices on a transmission medium. Data processed by processor 605 is transmitted on a wireless medium via antenna 603. Furthermore, antenna 603 receives data and transmits the data to processor 605.

[0334] The processor 605 is responsible for managing the bus 601 and general processing, and may also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. The memory 606 may be used to store data used by the processor 605 when performing operations.

[0335] Optionally, the processor 605 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or a graphics processing unit (GPU).

[0336] The embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, which, when executed by a processor, implements the above Figure 1 Each process of the corresponding fault identification method embodiment, or to achieve the above Figure 2 The various processes of the corresponding model training method embodiments can achieve the same technical effects, so they are not described here to avoid repetition. The computer-readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0337] The present invention also provides a computer program product, comprising computer instructions, which, when executed by a processor, implement the above Figure 1 Each process of the corresponding fault identification method embodiment, or to achieve the above Figure 2 The various processes of the corresponding model training method embodiments can achieve the same technical effects, so they will not be repeated here to avoid repetition.

[0338] The terms "first", "second" etc. in the embodiments of the present invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. In addition, the terms "comprise" and "have" and any of their variations are intended to cover non-exclusive inclusions, for example, the process, method, system, product or equipment comprising a series of steps or units are not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or that are inherent to these processes, methods, products or equipment. In addition, "and / or" is used in this application to represent at least one of the connected objects, for example A and / or B and / or C, which means comprising eight situations including single A, single B, single C, and both A and B exist, both B and C exist, both A and C exist, and both A, B and C exist.

[0339] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0340] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for enabling a terminal (which can be a mobile phone, computer, server, air conditioner, or second terminal device, etc.) to execute the methods of each embodiment of the present application.

[0341] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of this application, ordinary technicians in this field can also make many forms without departing from the purpose of this application and the scope of protection of the claims, all of which are within the protection of this application.

Claims

1. A fault identification method, characterized in that: include: Acquiring at least one first node status parameter of a network node, where the network node is a faulty network node; Calculating, by a preset network model, a posterior probability of multiple fault types based on a conditional probability table and the at least one first node state parameter, the conditional probability table including a priori probability and conditional probability between each of the multiple fault types and different node state parameters, the preset network model being used to calculate the posterior probability of the fault type based on the prior probability, the multiple fault types being multiple fault types of faults occurring at the network node; A first fault type is determined based on the posterior probability, where the first fault type is the fault type with the greatest posterior probability among the multiple fault types.

2. The method according to claim 1, wherein The preset network model is obtained in the following way: Acquire a priori knowledge set, wherein the priori knowledge set includes mapping relationships between the multiple fault types and different node states, and a first initial probability of each fault type; Based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type, the initial network model is trained to obtain the preset network model and the conditional probability table. The initial network model is a model for calculating the prior probability of each fault type based on the first initial probability.

3. The method according to claim 2, wherein Before training the initial network model based on the mapping relationship between the multiple fault types and different node states and the first initial probability of each fault type to obtain the preset network model and the conditional probability table, the training process further includes: Obtaining an order set of the initial network model, the order set including a plurality of Markov orders and an order probability of each Markov order; Calculating a first entropy value corresponding to each Markov order based on the first initial probabilities of the multiple fault types; Obtaining a second entropy value corresponding to each Markov order based on the order probability calculation; Determining a first Markov order, where the first Markov order is the Markov order with the largest difference between the third entropy value and the second entropy value, and the third entropy value is the first entropy value corresponding to the smallest Markov order among the multiple Markov orders; The training of the initial network model based on the mapping relationship between the multiple fault types and different node states and the first initial probability of each fault type to obtain the preset network model and the conditional probability table includes: The initial network model of the first Markov order is trained based on the mapping relationship between the multiple fault types and different node states, and the first initial probability of each fault type to obtain the preset network model and the conditional probability table.

4. The method according to claim 3, wherein The step of obtaining the second entropy value corresponding to each Markov order based on the order probability calculation includes: Calculating the conditional entropy of each Markov order based on the first entropy value; Calculating a second product corresponding to each Markov order based on the order probability, where the second product is the product of the order probability and the conditional entropy corresponding to each Markov order; The sum of the plurality of second products is set as a second entropy value corresponding to a third Markov order, wherein the Markov order of each product in the plurality of second products is less than or equal to the third Markov order.

5. The method according to any one of claims 2 to 4, characterized in that The mapping relationship includes an independent mapping relationship and / or an associated mapping relationship, wherein the independent mapping relationship is a mapping relationship between a fault type and a node status parameter, and the associated mapping relationship is a mapping relationship between a fault type and multiple node status parameters; The obtaining of a priori knowledge set includes: Acquire an initial knowledge set, the initial knowledge set including a first occurrence number of each fault type, a second occurrence number of each node state parameter, and a third occurrence number of different fault types and different node state parameters occurring simultaneously; Constructing the independent mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously; And / or, based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of different fault types and different node status parameters occurring simultaneously, the association mapping relationship corresponding to each fault type is constructed.

6. The method according to claim 5, wherein The constructing the independent mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of simultaneous occurrences of different fault types and different node status parameters includes: calculating a first initial probability of a second fault type based on the first number of occurrences, calculating a second initial probability of a first node state parameter based on the second number of occurrences, and calculating a third initial probability of simultaneous occurrence of the second fault type and the first node state parameter based on the third number of occurrences, wherein the second fault type is one of the multiple fault types and the first node state parameter is one of the multiple node state parameters; Calculating a first lift, where the first lift is a quotient of the third initial probability and a third product, where the third product is a sum of the first initial probability and the second initial probability; In a case where the first promotion degree is greater than a preset promotion degree threshold, an independent mapping relationship between the second fault type and the first node state parameter is constructed.

7. The method according to claim 5, wherein The constructing the association mapping relationship corresponding to each fault type based on the first occurrence number of each fault type, the second occurrence number of each node status parameter, and the third occurrence number of simultaneous occurrences of different fault types and different node status parameters includes: Calculating a conditional probability of each node state parameter and a second fault type, the conditional probability being a quotient of a third number of occurrences of each node state parameter and the second fault type occurring simultaneously and the first number of occurrences of the second fault type; Determine a frequent item set corresponding to the second fault type, where the frequent item set includes a plurality of second node state parameters; calculating a vector field of the second fault type and the plurality of second node state parameters; Calculating response values ​​corresponding to the plurality of second node state parameters based on the vector field; Determine a plurality of third node status parameters based on the response value, wherein the plurality of third node status parameters are node status parameters whose response values ​​are greater than a set response threshold among the plurality of second node status parameters; An association mapping relationship between the second fault type and the plurality of third node status parameters is constructed based on the conditional probability.

8. The method according to claim 7, wherein The determining of the frequent item sets corresponding to the second fault type includes: Calculate the mutual information MI and support corresponding to each node state parameter; Calculate the score corresponding to each node state parameter based on the MI and the support; A frequent item set corresponding to the second fault type is determined based on the score, where the frequent item set includes a plurality of second node state parameters, and the scores of the plurality of second node state parameters are greater than a set score threshold.

9. A fault identification device, characterized in that: include: an acquisition module, configured to acquire at least one first node status parameter of a network node, wherein the network node is a network node that has a fault; a calculation module, configured to calculate, using a preset network model and a conditional probability table and the at least one first node state parameter, a posterior probability of a plurality of fault types, the conditional probability table including a priori probabilities and conditional probabilities between each of the plurality of fault types and different node state parameters, the preset network model being configured to calculate the posterior probability of the fault type based on the priori probabilities, the plurality of fault types being a plurality of fault types in which the network node fails; A determination module is configured to determine a first fault type based on the posterior probability, where the first fault type is the fault type with the greatest posterior probability among the multiple fault types.

10. A computer program product, characterized in that The method comprises computer instructions, which, when executed by a processor, implement the steps of the fault identification method according to any one of claims 1 to 8.

Images