A fault location method and system for an information communication system

Abstract

The embodiment of the application provides a fault positioning method and system for an information communication system, and belongs to the technical field of diagnosis of information communication systems.The fault positioning method comprises the following steps: obtaining historical state data of an information communication system; encoding the historical state data to obtain a state matrix sequence; obtaining historical fault point data of the information communication system; associating each historical fault point with a state matrix from the time of the last system parameter adjustment closest to the historical fault point to the time of the historical fault point to form a data set; training a preset fault positioning network using the data set; determining whether the information communication system at the current time has a fault; in the case that the information communication system at the current time is determined to have a fault, obtaining a state matrix from the time of the last system parameter adjustment to the time of the current fault; and inputting the obtained state matrix into the fault positioning network to obtain a fault point.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of diagnostic techniques for information communication systems, and in particular to a fault locating method and system for information communication systems. BACKGROUND

[0002] In the field of information communication systems, with the continuous progress of technology, we are facing the challenges of increasing system complexity and data volume. This change not only promotes the rapid development of the industry, but also puts higher requirements on the stability and reliability of the system. In this context, the demand for fault prediction and diagnosis technology is becoming more and more urgent. System failure, whether it is hardware failure or software problem, can cause a chain reaction, leading to service interruption or even the collapse of the entire system. More seriously, these failures can also cause the loss or damage of important data, posing a huge risk to the normal operation of the enterprise and directly affecting customer satisfaction and trust.

[0003] In order to effectively respond to these challenges, it has become an urgent need of the industry to develop a technology that can accurately predict the location of fault points. Such a solution not only needs to have a high degree of accuracy, but also needs to be able to quickly locate and warn potential fault points so that maintenance personnel can intervene in time to prevent the occurrence or expansion of faults.

[0004] Traditional fault prediction methods, such as manual inspection and regular maintenance, while to some extent can find problems, but the efficiency is relatively low, and it is difficult to achieve timely discovery and handling of potential fault hazards. SUMMARY

[0005] The purpose of the embodiments of the present application is to provide a fault locating method and system for information communication systems, which can accurately and quickly find the fault locating point of the system.

[0006] In order to achieve the above-mentioned purpose, the embodiments of the present application provide a fault locating method for information communication systems:

[0007] Obtain historical state data of the information communication system;

[0008] Encode the historical state data to obtain a sequence of state matrices;

[0009] Obtain historical fault point data of the information communication system;

[0010] Correlate each historical fault point with the state matrix from the time of the last system parameter adjustment closest to the historical fault point to the time of the historical fault point to form a data set;

[0011] Train a pre-set fault locating network using the data set;

[0012] determining whether the information communication system at the current time instant has a fault;

[0013] in the case where it is determined that the information communication system at the current time instant has a fault, obtaining a state matrix from a time instant of a previous system parameter adjustment to a time instant of the current fault;

[0014] inputting the obtained state matrix into the fault locating network to obtain a fault point.

[0015] Optionally, the historical state data is encoded to obtain a state matrix sequence, comprising:

[0016] the state matrix sequence is constructed according to formula (1),

[0017]

[0018] wherein S is the state matrix, s ij is a state value of an i-th type state at a j-th sampling point, 1≤i,j≤n,m.

[0019] Optionally, each historical fault point is associated with a state matrix from a time instant of a previous system parameter adjustment closest to the historical fault point to a time instant of the historical fault point to form a data set, comprising:

[0020] the state matrix is intercepted in the state matrix sequence according to a time instant of a previous system parameter adjustment to a time instant of the historical fault point.

[0021] Optionally, the state matrix is intercepted in the state matrix sequence according to a time instant of a previous system parameter adjustment to a time instant of the historical fault point, comprising:

[0022] determining a single type of sampling point to be intercepted according to the time instant;

[0023] determining whether a number of sampling points of the state matrix is less than or equal to a preset minimum sampling point threshold;

[0024] in the case where it is determined that the number of sampling points of the state matrix is less than or equal to the minimum sampling point threshold, deleting the state matrix and the historical fault point.

[0025] Optionally, the state matrix is intercepted in the state matrix sequence according to a time instant of a previous system parameter adjustment to a time instant of the historical fault point, comprising:

[0026] determining whether a number of sampling points of the state matrix is greater than or equal to a preset maximum sampling point threshold;

[0027] In a case where the number of sampling points is greater than or equal to the maximum sampling point threshold, the state matrix is marked as an abnormal state matrix;

[0028] determining a standard sampling point number;

[0029] interpolating or sampling all state matrices except the deleted state matrix to standardize each state matrix to obtain the truncated state matrix.

[0030] Optionally, determining the standard sampling point number comprises:

[0031] determining the standard sampling point number according to formula (2),

[0032]

[0033] wherein c is the standard sampling point number, m i is the number of sampling points of a single class of the i-th state matrix, and p is the number of all state matrices except the deleted state matrix and the abnormal state matrix.

[0034] Optionally, training the preset fault positioning network using the data set comprises:

[0035] generating a decision tree according to the state matrix;

[0036] training the decision tree using the data set;

[0037] In a case where the decision tree does not meet the prediction accuracy requirement, performing a pruning operation on the decision tree to update the decision tree, and returning to the step of training the decision tree using the data set;

[0038] In a case where the decision tree meets the prediction accuracy requirement, determining that the training of the fault positioning network is completed.

[0039] Optionally, training the decision tree using the data set comprises

[0040] using formula (3) as a loss function,

[0041] W = - [2w k 2 +∑(w max -w n-1 ) 2 ], (3)

[0042] wherein W is the loss function, w k is the probability that the sample belongs to the k-th class, w max is the maximum probability of the classification result, and w n-1 is the second maximum probability of the classification result.

[0043] In another aspect, the present application also provides a fault locating system for an information communication system, the fault locating system comprising a processor configured to perform the fault locating method according to any one of the above.

[0044] In another aspect, the present application also provides a computer readable storage medium storing instructions for being read by a machine to cause the machine to perform the fault locating method according to any one of the above.

[0045] By the above technical solution, the present application provides a fault locating method and system for an information communication system, which constructs a state matrix by intercepting historical state data in the information communication system, and then performs fault locating in combination with a pre-set fault locating network, so as to realize accurate and efficient fault locating of the information communication system. Compared with the prior art, the fault locating method and system provided by the present application can realize fault locating by only using the state data of the system after the previous parameter adjustment, without the need of combining additional system diagnosis operations, thereby simplifying the fault locating process and improving the fault locating efficiency.

[0046] Other features and advantages of the present application will be described in detail in the following specific implementation part. BRIEF DESCRIPTION OF DRAWINGS

[0047] The accompanying drawings are included to provide a further understanding of the present application, and constitute a part of the specification, and are used together with the following specific implementation part to explain the present application, but do not constitute a limitation on the present application. In the drawings:

[0048] Figure 1 is a flowchart of a fault locating method for an information communication system according to an embodiment of the present application;

[0049] Figure 2 is a flowchart of an association manner according to an embodiment of the present application. DETAILED DESCRIPTION

[0050] The specific implementation part of the present application will be described in detail below in combination with the drawings. It should be understood that the specific implementation part described herein is only used to illustrate and explain the present application, and does not limit the present application.

[0051] As shown in Figure 1 is a flowchart of a fault locating method for an information communication system according to an embodiment of the present application. In this Figure 1 , the fault locating method can include the following steps:

[0052] In step S10, historical state data of the information communication system is acquired;

[0053] In step S11, the historical state data is encoded to obtain a state matrix sequence;

[0054] In step S12, historical fault point data of the information communication system is acquired;

[0055] In step S13, each historical fault point is associated with a state matrix from a time of a previous system parameter adjustment closest to the historical fault point to a time of the historical fault point to form a data set;

[0056] In step S14, the data set is used to train a preset fault locating network;

[0057] In step S15, it is determined whether the information communication system at a current time has a fault;

[0058] In step S16, in a case where it is determined that the information communication system at the current time has a fault, a state matrix from a time of a previous system parameter adjustment to a time of the current fault is acquired;

[0059] In step S17, the acquired state matrix is input into the fault locating network to obtain a fault point.

[0060] In the method as shown in the figure, Figure 1 In the method as shown in the figure,

[0061] Step S11 can be used to encode the historical state data to obtain a state matrix sequence. The state matrix can be used to represent all the sampling data corresponding to the acquired historical state data. In an example of the present application, the state matrix sequence can be represented by the following formula (1):

[0062]

[0063] wherein S is a state matrix, s ij is a state value of an i-th type state at a j-th sampling point, 1≤i,j≤n,m.

[0064] Step S12 can be used to associate each historical failure point with a state matrix from the time of the previous system tuning closest to the historical failure point to the time of the historical failure point to form a data set. Specifically, the way of association can be to intercept the state matrix in the state matrix sequence according to the time of the previous system tuning to the time of the historical failure point. Further, step S12 can further include the method shown in Figure 2 In the method shown in Figure 2 Step S12 can include the following steps:

[0065] In step S20, the sampling points of a single category to be intercepted are determined according to the time;

[0066] In step S21, it is judged whether the number of sampling points of the state matrix is less than or equal to the preset minimum sampling point threshold;

[0067] In step S22, in the case where the number of sampling points of the state matrix is less than or equal to the minimum sampling point threshold, the state matrix and the historical failure point are deleted.

[0068] In step S23, it is judged whether the number of sampling points of the state matrix is greater than or equal to the preset maximum sampling point threshold;

[0069] In step S24, in the case where the number of sampling points is greater than or equal to the maximum sampling point threshold, the state matrix is marked as an abnormal state matrix;

[0070] In step S25, the standard sampling point number is determined;

[0071] In step S26, interpolation or sampling operation is performed on all state matrices except the deleted state matrices to standardize each state matrix to obtain the intercepted state matrix.

[0072] In the method shown in Figure 2 Step S20 can be used to determine the sampling points of a single category to be intercepted according to the time of the historical failure point. Wherein, the interception method of the sampling points can be a time interval from the time of the previous system tuning to the time of the historical failure point.

[0073] Steps S21 to S24 can be used to determine whether the number of sampling points meets the requirement of determining the standard sampling point number. Specifically, steps S21 and S22 are mainly used to delete the state matrix with too few sampling points, because too few sampling points will lead to lack of sufficient vector parameters in subsequent network iteration, thereby reducing the prediction accuracy. Steps S23 and S24 are mainly used to mark the state matrix with too many sampling points, so as to avoid the interference caused by the calculation of the standard sampling points of such state matrix.

[0074] Step S25 can be used to determine the standard sampling point number. For the specific method of determining the standard sampling point number, it can be various forms known to those skilled in the art. In an example of the present application, the method of determining the standard sampling point number can be calculated using the following formula (2):

[0075]

[0076] Wherein c is the standard sampling point number, m i is the number of sampling points of a single class of the i-th state matrix, and p is the number of all state matrices except the deleted state matrix and the abnormal state matrix.

[0077] Step S26 can be used to interpolate or sample all state matrices except the deleted state matrix to standardize each state matrix to obtain the truncated state matrix. In this step S26, the reason why the deleted state matrix is not interpolated is that the number of sampling points of the deleted state matrix is too small, and if the interpolation operation is performed to augment the matrix, it will cause the data characteristics of the matrix to be distorted, which not only cannot reflect the data characteristics, but also will interfere with the subsequent network training. And the sampling operation is performed on the abnormal state matrix, because in most cases, although the sampling operation reduces the sampling points, it can still maintain the general data characteristics. In addition, the specific method of interpolation operation can be, for example, linear interpolation operation.

[0078] Step S14 can be used to train the preset fault locating network using the data set. For the specific method of training the fault locating network, it can be various forms known to those skilled in the art. In an example of the present application, the method of training the fault locating network can be to first generate a decision tree according to the dimension of the state matrix, and then divide the data set into a training set and a test set to divide the decision tree. In the case where the decision tree does not meet the prediction accuracy requirement, the decision tree is pruned to update the decision tree and continue to train the decision tree; in the case where it is judged that the decision tree meets the prediction accuracy requirement, the fault locating network training is determined to be completed. For the loss function in the process of training the decision tree, considering that the conventional loss function is only through the maximum prediction probability, that is, the final probability to perform loss iteration. Although this way can also iterate and update a better network model, this way does not consider the influence of the second prediction result on the first prediction result, so that the robustness of the prediction result is low, therefore in an example of the present application, the following formula (3) can be used as the loss function:

[0079] W = -[∑w k 2 +∑(w max -w n-1 )2 ], (3)

[0080] wherein, W is a loss function, w k is a probability that a sample belongs to the kth class, w max is a maximum probability of a classification result, w n-1 is a second maximum probability of a classification result.

[0081] In another aspect, the present application also provides a fault locating system for an information communication system, the fault locating system comprising a processor configured to perform the fault locating method according to any one of the above. Specifically, the fault locating method can comprise the following steps:

[0082] In step S10, historical state data of the information communication system is obtained;

[0083] In step S11, the historical state data is encoded to obtain a sequence of state matrices;

[0084] In step S12, historical fault point data of the information communication system is obtained;

[0085] In step S13, each historical fault point is associated with a state matrix from a time of a previous system parameter adjustment closest to the historical fault point to a time of the historical fault point to form a data set;

[0086] In step S14, the data set is used to train a pre-set fault locating network;

[0087] In step S15, it is determined whether the information communication system at a current time has a fault;

[0088] In step S16, in a case where it is determined that the information communication system at the current time has a fault, a state matrix from a time of a previous system parameter adjustment to a time of the current fault is obtained;

[0089] In step S17, the obtained state matrix is input into the fault locating network to obtain a fault point.

[0090] In the method shown in the above, Figure 1 Step S10 can be used to obtain historical state data of the information communication system. The historical state data can include state sampling information of the information communication system. The state sampling information can include, but is not limited to, device load state, network response rate (packet loss rate), network delay value, interface state, etc.

[0091] Step S11 can be used to encode historical state data to obtain a state matrix sequence. The state matrix can represent all sampled data corresponding to all acquired historical state data. In one example of the present invention, the state matrix sequence can be represented by the following formula (1):

[0092]

[0093] Where S is the state matrix, s ij Let be the state value of the j-th sampling point of the i-th state, where 1≤i,j≤n,m.

[0094] Step S12 may involve associating each historical fault point with the state matrix from the time of the most recent system parameter tuning to the time of the historical fault point, thereby forming a dataset. Specifically, the association can be performed by truncating the state matrix from the sequence of state matrices based on the time from the previous system parameter tuning to the time of the historical fault point. Further, step S12 may include further steps such as... Figure 2 The method shown in the diagram. Figure 2 In step S12, the steps may include the following:

[0095] In step S20, the sampling points for a single category to be extracted are determined based on the time.

[0096] In step S21, it is determined whether the number of sampling points in the state matrix is ​​less than or equal to the preset minimum sampling point threshold.

[0097] In step S22, if the number of sampling points in the state matrix is ​​less than or equal to the minimum sampling point threshold, the state matrix and historical fault points are deleted.

[0098] In step S23, it is determined whether the number of sampling points in the state matrix is ​​greater than or equal to the preset maximum sampling point threshold.

[0099] In step S24, if the number of sampling points is greater than or equal to the maximum sampling point threshold, the state matrix is ​​marked as an abnormal state matrix.

[0100] In step S25, the number of standard sampling points is determined;

[0101] In step S26, interpolation or sampling operations are performed on all state matrices except for the deleted state matrix to standardize each state matrix and obtain the truncated state matrix.

[0102] In such Figure 2In the illustrated method, step S20 can be used to determine the sampling points of the single category to be intercepted according to the time of the historical fault point. In this case, the interception manner of the sampling points can be a time interval from the time of the previous system parameter adjustment to the time of the historical fault point.

[0103] Steps S21 to S24 can be used to determine whether the number of sampling points meets the requirement of the determined standard sampling point number. Specifically, steps S21 and S22 are mainly used to delete the state matrices with too few sampling points, because too few sampling points will lead to a lack of sufficient vector parameters in subsequent network iterations, thereby reducing the prediction accuracy. Steps S23 and S24 are mainly used to mark the state matrices with too many sampling points, so as to avoid the interference caused by the calculation of the standard sampling points of such state matrices.

[0104] Step S25 can be used to determine the standard sampling point number. For the specific method of determining the standard sampling point number, it can be various forms known to those skilled in the art. In one example of the present application, the method of determining the standard sampling point number can be calculated by using the following formula (2):

[0105]

[0106] where c is the standard sampling point number, m i is the number of sampling points of the single category of the i-th state matrix, and p is the number of all state matrices except the deleted state matrices and the abnormal state matrices.

[0107] Step S26 can be used to interpolate or sample all state matrices except the deleted state matrices to standardize each state matrix to obtain the intercepted state matrix. In this step S26, the deleted state matrices are not interpolated because the number of sampling points of the deleted state matrices is too small. If interpolation is performed to augment the matrix, the data characteristics of the matrix will be distorted, which not only cannot reflect the data characteristics, but also will interfere with the subsequent network training. The abnormal state matrices are sampled because in most cases, although the sampling operation reduces the sampling points, it can still maintain the general data characteristics. In addition, the specific method of interpolation can be, for example, a linear interpolation operation.

[0108] Step S14 can be used to train the preset fault locating network with the data set. For the specific method of training the fault locating network, it can be various forms known to those skilled in the art, and in an example of the present application, the method of training the fault locating network can be to first generate a decision tree according to the dimension of the state matrix, and then divide the data set into a training set and a test set to divide the decision tree. In the case where the decision tree does not meet the prediction accuracy requirement, the decision tree is pruned to update the decision tree, and the training of the decision tree is continued; in the case where it is judged that the decision tree meets the prediction accuracy requirement, the fault locating network training is determined to be completed at this time. For the loss function in the process of training the decision tree, considering that the conventional loss function is only through the maximum prediction probability, that is, the final probability to perform loss iteration. Although this way can also update a better network model, but this way does not consider the influence of the second prediction result on the first prediction result, so that the robustness of the prediction result is low, therefore, in an example of the present application, the following formula (3) can be used as the loss function:

[0109] W=-[∑w k 2 +∑(w max -w n-1 ) 2 ], (3)

[0110] wherein, W is the loss function, w k is the probability of the sample belonging to the kth class, w max is the maximum probability of the classification result, and w n-1 is the second maximum probability of the classification result.

[0111] In still another aspect, the present application also provides a computer readable storage medium, which stores instructions for being read by a machine to make the machine execute the fault locating method as described in any of the above. Specifically, the fault locating method can include the following steps:

[0112] In step S10, the historical state data of the information communication system is obtained;

[0113] In step S11, the historical state data is encoded to obtain a state matrix sequence;

[0114] In step S12, the historical fault point data of the information communication system is obtained;

[0115] In step S13, each historical fault point is associated with the state matrix from the time of the nearest previous system parameter adjustment to the time of the historical fault point to form a data set;

[0116] In step S14, the preset fault locating network is trained by using the data set;

[0117] In step S15, it is judged whether the information communication system at the current time has a fault;

[0118] In step S16, in the case that it is judged that the information communication system at the current time has a fault, the state matrix from the time of the previous system tuning to the time of the current fault is obtained;

[0119] In step S17, the obtained state matrix is input into the fault locating network to obtain the fault point.

[0120] In the method shown in the Figure 1 , step S10 can be used to obtain the historical state data of the information communication system. The historical state data can include state sampling information of the information communication system. The state sampling information can include, but is not limited to, device load state, network response rate (packet loss rate), network delay value, interface state, etc.

[0121] Step S11 can be used to encode the historical state data to obtain a state matrix sequence. The state matrix can be used to represent all the sampling data corresponding to all the obtained historical state data. In an example of the present application, the state matrix sequence can be represented by the following formula (1):

[0122]

[0123] Wherein, S is the state matrix, s ij is the state value of the jth sampling point of the ith state, 1≤i,j≤n,m.

[0124] Step S12 can be used to associate each historical fault point with the state matrix from the time of the previous system tuning closest to the historical fault point to the time of the historical fault point to form a data set. Specifically, the association can be performed by intercepting the state matrix in the state matrix sequence according to the time of the previous system tuning to the time of the historical fault point. Further, step S12 can further include the method shown in Figure 2 . In the Figure 2 , step S12 can include the following steps:

[0125] In step S20, the sampling points of a single category to be intercepted are determined according to the time;

[0126] In step S21, it is judged whether the number of sampling points of the state matrix is less than or equal to the preset minimum sampling point threshold;

[0127] In step S22, in the case that the number of sampling points of the state matrix is less than or equal to the minimum sampling point threshold, the state matrix and the historical fault point are deleted.

[0128] In step S23, it is judged whether the number of sampling points of the state matrix is greater than or equal to the preset maximum sampling point threshold.

[0129] In step S24, in the case that the number of sampling points is greater than or equal to the maximum sampling point threshold, the state matrix is marked as an abnormal state matrix.

[0130] In step S25, the standard sampling point number is determined.

[0131] In step S26, interpolation or sampling operation is performed on all state matrices except the deleted state matrix to standardize each state matrix to obtain a truncated state matrix.

[0132] In the method shown in the figure, Figure 2 Step S20 can be used to determine the sampling points of a single category to be truncated according to the time of the historical fault point. The truncation of the sampling points can be a time interval from the time of the last system parameter adjustment to the time of the historical fault point.

[0133] Steps S21 to S24 can be used to determine whether the number of sampling points meets the requirement of determining the standard sampling point number. Specifically, steps S21 and S22 are mainly used to delete state matrices with too few sampling points, because too few sampling points will lead to a lack of sufficient vector parameters in subsequent network iteration, thereby reducing the prediction accuracy. Steps S23 and S24 are mainly used to mark state matrices with too many sampling points, so as to avoid the interference caused by the calculation of the standard sampling points of such state matrices.

[0134] Step S25 can be used to determine the standard sampling point number. The specific method for determining the standard sampling point number can be various forms known to those skilled in the art. In an example of the present application, the method for determining the standard sampling point number can be calculated using the following formula (2):

[0135]

[0136] Where c is the standard sampling point number, m i is the number of sampling points of a single category of the i th state matrix, and p is the number of all state matrices except the deleted state matrix and the abnormal state matrix.

[0137] Step S26 can be used to interpolate or sample all state matrices except the deleted state matrix, to standardize each state matrix, to obtain the truncated state matrix. In this step S26, the reason why the deleted state matrix is not interpolated is that the number of sampling points of the deleted state matrix is too small, and if the interpolation operation is performed and the matrix is augmented, the data characteristics of the matrix will be distorted, which not only cannot reflect the data characteristics, but also will interfere with the subsequent network training. The sampling operation on the abnormal state matrix is because in most cases, although the sampling operation reduces the sampling points, it can still maintain the general data characteristics. In addition, the specific method of the interpolation operation can be, for example, a linear interpolation operation.

[0138] Step S14 can be used to train the preset fault positioning network with the data set. The specific method for training the fault positioning network can be various forms known to those skilled in the art, and in an example of the present application, the method for training the fault positioning network can be to first generate a decision tree according to the dimension of the state matrix, and then divide the data set into a training set and a test set to divide the decision tree. In the case where the decision tree does not meet the prediction accuracy requirement, the decision tree is pruned to update the decision tree, and the decision tree is returned to continue training. In the case where it is judged that the decision tree meets the prediction accuracy requirement, the fault positioning network training is determined to be completed. For the loss function in the process of training the decision tree, considering that the conventional loss function is only through the maximum prediction probability, i.e., the final probability to perform loss iteration. Although this way can also iterate and update a better network model, this way does not consider the influence of the second prediction result on the first prediction result, so that the robustness of the prediction result is low, and therefore in an example of the present application, the following formula (3) can be used as the loss function:

[0139] W = -[∑w k 2 +∑(w max -w n-1 ) 2 ], (3)

[0140] wherein W is the loss function, w k is the probability that the sample belongs to the kth class, w max is the maximum probability of the classification result, and w n-1 is the second maximum probability of the classification result.

[0141] By means of the technical scheme, the embodiment of the present application provides a fault positioning method and system for an information communication system, which constructs a state matrix by intercepting historical state data in the information communication system, and then combines a preset fault positioning network to perform fault positioning, so that accurate and efficient fault positioning of the information communication system is realized. Compared with the prior art, the fault positioning method and system provided by the present application can realize fault positioning by using state data of the system after the previous parameter adjustment, without combining additional system diagnosis operations, thereby simplifying the fault positioning process and improving the fault positioning efficiency.

[0142] Those skilled in the art will understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0143] The present application is described with reference to flowcharts and / or block diagrams according to the methods, devices (systems), and computer program products of the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus generate a means for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus that implements the functions specified in the flowcharts and / or block diagrams.

[0144] These computer program instructions can also be stored in a computer-readable memory that can direct the computer or other programmable data processing apparatus to work in a specific manner, so that the instructions stored in the computer-readable memory produce a product including instruction apparatus, which implements the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus that implements the functions specified in the flowcharts and / or block diagrams.

[0145] These computer program instructions can also be loaded into a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable data processing apparatus to produce a computer-implemented process, so that the instructions executed on the computer or other programmable data processing apparatus provide a means for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocksFigure 1 the function specified in the block or blocks.

[0146] In one typical arrangement, the computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0147] The memory can include non-persistent memory and / or volatile memory, such as random access memory (RAM) about which the processor can execute instructions. The memory can also include non-volatile memory, such as read only memory (ROM), electrically programmable read only memory (EPROM), electrically erasable programmable read only memory (EEPROM), flash memory, or other memory technologies, about which the processor can execute instructions. The memory is an example of computer readable storage media.

[0148] Computer readable media includes permanent and non-permanent, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), electrically programmable read only memory (EEPROM), flash memory or other memory technologies, compact disc read only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer readable media does not include transitory media, such as modulated data signals and carrier waves.

[0149] It should also be noted that the terms "comprising", "containing", or any other variant thereof, are intended to encompass a non-exclusive inclusion, such that a process, method, article or apparatus that comprises a list of elements does not include only those elements recited, but can also include other elements not expressly listed or inherent to such process, method, article or apparatus. Without further limitation, an element preceded by "comprises a" does not, without more constraints, foreclose the existence of additional identical elements in the process, method, article or apparatus that includes the element.

[0150] The above merely provides an embodiment of the present application and is not intended to limit the present application. The present application can have various modifications and changes for those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of claims of the present application.

Claims

1. A fault locating method for an information communication system, characterized by, The fault locating method comprises: acquiring historical state data of an information communication system; encoding the historical state data to obtain a state matrix sequence; acquiring historical fault point data of the information communication system; associating each historical fault point with a state matrix from a time of a previous system tuning closest to a time of the historical fault point to the time of the historical fault point to form a data set; training a preset fault locating network using the data set; judging whether the information communication system at a current time is faulty; in a case where it is judged that the information communication system at the current time is faulty, acquiring a state matrix from the time of the previous system tuning to the time of the current fault; inputting the acquired state matrix into the fault locating network to obtain a fault point; associating each historical fault point with a state matrix from a time of a previous system tuning closest to a time of the historical fault point to the time of the historical fault point to form a data set, comprising: intercepting the state matrix in the state matrix sequence according to the time of the previous system tuning to the time of the historical fault point.

2. The fault locating method of claim 1, wherein, encoding the historical state data to obtain a state matrix sequence, comprising: constructing the state matrix sequence according to formula (1), ，（1） wherein is the state matrix, is the state matrix, is the state matrix, is the state value of the th sample point of the 3. The fault locating method of claim 1, wherein, intercepting the state matrix in the state matrix sequence according to the time of the previous system tuning to the time of the historical fault point, comprising: determining a sampling point of a single category to be intercepted according to the time; judging whether the number of sampling points of the state matrix is less than or equal to a preset minimum sampling point threshold; in a case where it is judged that the number of sampling points of the state matrix is less than or equal to the minimum sampling point threshold, deleting the state matrix and the historical fault point.

4. The fault locating method of claim 1, wherein, intercepting the state matrix in the state matrix sequence according to the time of the previous system tuning to the time of the historical fault point, comprising: judging whether the number of sampling points of the state matrix is greater than or equal to a preset maximum sampling point threshold; in a case where it is judged that the number of sampling points is greater than or equal to the maximum sampling point threshold, marking the state matrix as an abnormal state matrix; determining a standard sampling point number; interpolating or sampling all state matrices except the deleted state matrix to standardize each state matrix to obtain the intercepted state matrix.

5. The fault locating method of claim 4, wherein, determining a standard sampling point number, comprising: determining the standard sampling point number according to formula (2), ，（2） wherein, is the number of the standard sampling points, is the number of the sampling points of a single class of the th state matrix, is the number of all state matrices except the state matrix deleted and the abnormal state matrix.

6. The fault locating method of claim 1, wherein, training a preset fault locating network using the data set, comprising: generating a decision tree according to the state matrix; training the decision tree using the data set; in a case where the decision tree does not meet a prediction accuracy requirement, performing a pruning operation on the decision tree to update the decision tree, and returning to perform the step of training the decision tree using the data set; in a case where it is judged that the decision tree meets the prediction accuracy requirement, determining that the fault locating network is trained.

7. The fault locating method of claim 6, wherein, training the decision tree using the data set, comprising using formula (3) as a loss function, ，（3） wherein, is the loss function, is the probability that the sample belongs to the class, is the maximum probability of the classification result, is the second maximum probability of the classification result.

8. A fault locating system for an information communication system, characterized by The fault locating system comprises a processor configured to perform the fault locating method of any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores instructions for being read by a machine to cause the machine to perform the fault locating method of any one of claims 1 to 7.

Images