An AI-based intelligent communication fault prediction system

Abstract

The present invention relates to the field of communication data processing technology, and in particular to an AI-based intelligent communication fault prediction system, including a sample acquisition module for acquiring communication sample information; a fault statistics module for analyzing the sample failure rate according to the communication sample information; a sample analysis module for analyzing sample association parameters and sample feature parameters; a user acquisition module for acquiring communication user information in real time; an association analysis module for analyzing user association parameters and association similarity; a user analysis module for analyzing user bandwidth usage and user feature parameters; a fault prediction module for analyzing similar feature parameters and analyzing user fault risks according to similar feature parameters and sample operation information. A prediction output module for outputting user fault risks. The present invention realizes accurate prediction of communication faults.

Description

Technical Field

[0001] The present invention relates to the field of communication data processing technology, and in particular to an AI-based intelligent communication fault prediction system. Background Art

[0002] With the development of communication networks, traditional fault prediction methods that rely on manual inspections and regular troubleshooting are no longer able to meet the efficient maintenance needs of modern networks. Artificial intelligence mines and analyzes large amounts of historical data, automatically extracts fault characteristics, and builds prediction models to accurately predict network failures.

[0003] Chinese Patent Publication No. CN102487343A discloses a method for diagnosing and predicting hidden faults in satellite communication systems. The method involves normalizing heterogeneous data in communication records, applying a SOM to cluster analysis of communication record data for hidden faults in earth stations within the satellite communication system, classifying communication states to form communication state sets, analyzing each communication state set based on fault pattern summary, comparing the relationship between different communication state parameters and fault phenomena, and summarizing decision-making rules using an incremental decision tree algorithm to form a fault rule set. The communication state set and fault rule set are then presented in an interpretable form, thereby implementing a prototype system for diagnosing and predicting hidden faults in satellite communication systems. This invention implements unified cluster analysis of communication data but fails to comprehensively analyze the characteristics of sample communication data and user communication data. This results in low communication data analysis efficiency and inaccurate communication fault prediction. Summary of the Invention

[0004] To this end, the present invention provides an AI-based intelligent communication fault prediction system to overcome the problems of low efficiency in communication data analysis and inaccurate communication fault prediction in the prior art.

[0005] To achieve the above objectives, the present invention provides an AI-based intelligent communication fault prediction system, comprising:

[0006] A sample acquisition module is used to obtain communication sample information;

[0007] Fault statistics module, used to analyze sample failure rate based on communication sample information;

[0008] A sample analysis module, configured to analyze sample-related parameters based on communication sample information, and further configured to analyze sample characteristic parameters based on sample-related parameters, communication sample information, and sample failure rate;

[0009] User acquisition module, used to obtain communication user information in real time;

[0010] An association analysis module is used to analyze user association parameters based on communication user information, and also to analyze association similarity based on user association parameters and sample association parameters;

[0011] A user analysis module is used to analyze user bandwidth usage based on communication user information, and is also used to analyze user characteristic parameters based on user bandwidth usage, communication user information and correlation similarity;

[0012] A fault prediction module is used to analyze similar characteristic parameters based on sample characteristic parameters and user characteristic parameters, and also to analyze user fault risks based on similar characteristic parameters and communication sample information;

[0013] The prediction output module is used to output user failure risks.

[0014] Furthermore, the fault statistics module calculates the sample failure rate according to the first fault number, the second fault number, and the third fault number through a fault analysis formula. The fault analysis module is provided with a fault analysis formula as follows:

[0015] A1=(N1 / NA) Na / NA ;

[0016] A2=(N2 / NA) Na / NA ;

[0017] Na=N1+N2+N3-NA;

[0018] Among them, A1 represents the first sample failure rate, N1 represents the first failure number, NA represents the number of communication sample information, Na represents the sample failure parameter, A2 represents the second sample failure rate, N2 represents the second failure number, and N3 represents the third failure number.

[0019] Furthermore, the sample analysis module is provided with a sample correlation analysis unit, which is used to calculate the sample correlation parameter according to the sample signal strength, the sample transmission rate, the first fault number and the second fault number through a sample correlation analysis formula. The sample correlation analysis unit is provided with a sample correlation analysis formula as follows:

[0020] ,

[0021] ,

[0022] ,

[0023] ,

[0024] ,

[0025] ,

[0026] ,

[0027] Where R1 represents the first sample association parameter, R2 represents the second sample association parameter, R3 represents the third sample association parameter, X1(i) represents the sample signal strength parameter, x1 represents the sample signal strength mean, X2(i) represents the sample transmission rate parameter, x2 represents the sample transmission rate mean, S(i,j) represents the sample signal strength, K1 represents the set of the first fault number, K2 represents the set of the second fault number, i represents the sample number, i∈N + And i≤NA, j represents the sample time number, j∈N + .

[0028] Furthermore, the sample analysis module is further provided with a sample feature analysis unit, which is used to calculate the sample feature parameters according to the sample signal strength, sample bandwidth information and sample bandwidth usage rate through a sample feature analysis formula. The sample feature analysis unit is provided with a sample feature analysis formula as follows:

[0029] ,

[0030] Among them, Q(i) represents the sample characteristic parameter, j max It represents the maximum value of the sample time number, D(i,j) represents the sample bandwidth usage, and d1(i,j) represents the sample transmission rate.

[0031] Furthermore, the sample analysis module is further provided with a sample feature adjustment unit, which is used to adjust the analysis process of the sample feature parameters according to the sample association parameters and the sample operation information, wherein:

[0032] When the sample operation information is a hardware failure and R2 / R1>r, the sample feature adjustment unit determines that the sample feature parameter does not meet the threshold, adjusts the analysis process of the sample feature parameter, and the adjusted sample feature parameter is Q1(i), setting Q1(i)=Q(i)×(1-R2+R1);

[0033] When the sample operation information is a network failure and R3 / R1>r, the sample feature adjustment unit determines that the sample feature parameter does not meet the threshold, adjusts the analysis process of the sample feature parameter, and the adjusted sample feature parameter is Q1(i), setting Q1(i)=Q(i)×(1-R3+R1);

[0034] When R2 / R1≤r or R3 / R1≤r, the sample feature adjustment unit determines that the sample feature parameter meets the threshold value and does not adjust the analysis process of the sample feature parameter;

[0035] Where r represents the correlation threshold;

[0036] The sample analysis module is further provided with a sample feature optimization unit, which is used to optimize the sample associated parameters according to the sample failure rate, so as to optimize the adjustment process of the sample feature parameters, wherein:

[0037] When A1+A2≤a, the sample feature optimization unit determines that the sample failure rate meets the threshold value and does not optimize the adjustment process of the sample feature parameters;

[0038] When the sample operation information is a hardware failure and A1+A2>a, the sample feature optimization unit determines that the sample failure rate does not meet the threshold, optimizes the sample association parameter, and sets the optimized sample association parameter R1'=R1×e A1 / 2 ;

[0039] When the sample operation information is a network failure and A1+A2>a, the sample feature optimization unit determines that the sample failure rate does not meet the threshold, optimizes the sample association parameter, and sets the optimized sample association parameter R1'=R1×e A2 / 2 ;

[0040] Where a represents the failure rate threshold.

[0041] Furthermore, the association analysis module is provided with a user association analysis unit, which is used to calculate the user association parameter according to the user signal strength and the user transmission rate through the user association analysis formula. The user association analysis unit is provided with the user association analysis formula as follows:

[0042] ,

[0043] ,

[0044] ,

[0045] Where Z represents the user association parameter, G(u) represents the user signal strength, u represents the user time number, u∈N + , g represents the mean user signal strength, H(u) represents the user transmission rate, h represents the mean user transmission rate, u max Indicates the maximum value of the user time number.

[0046] Furthermore, the association analysis module is further provided with a user similarity analysis unit, which is used to calculate the association similarity according to the user association parameter and the sample association parameter through a similarity analysis formula. The user similarity analysis unit is provided with a similarity analysis formula as follows:

[0047] P=(Z / R1) Z-(R2+R3) / 2 ;

[0048] Where P represents the association similarity.

[0049] Furthermore, the user analysis module is provided with a user feature analysis unit, which is used to calculate user feature parameters according to user signal strength, user bandwidth usage and user bandwidth information through a user feature analysis formula. The user feature analysis unit is provided with a user feature analysis formula as follows:

[0050] ,

[0051] Where F represents the user characteristic parameter, and M(u) represents the user bandwidth usage rate.

[0052] Furthermore, the user analysis module is further provided with a user feature adjustment unit, which is used to adjust the analysis process of the user feature parameters according to the correlation similarity, wherein:

[0053] When P≥p, the user feature adjustment unit determines that the correlation similarity meets the threshold value and does not adjust the analysis process of the user feature parameters;

[0054] When P < p, the user feature adjustment unit determines that the correlation similarity does not meet the threshold, and adjusts the analysis process of the user feature parameter. The adjusted user feature parameter is F1, and F1 = F / e (1-P) ;

[0055] Where p represents the similarity threshold.

[0056] Furthermore, the fault prediction module analyzes the user fault risk based on the first feature quantity and the second feature quantity, wherein:

[0057] When B1 / NB>A1, the fault prediction module sets the user fault risk to high hardware fault risk;

[0058] When B2 / NB>A2, the fault prediction module sets the user fault risk to high network fault risk;

[0059] When B1 / NB≤A1 and B2 / NB≤A2, the fault prediction module does not analyze the user fault risk;

[0060] Wherein, B1 represents the number of first features, B2 represents the number of second features, and NB represents the number of similar feature parameters.

[0061] Compared with the prior art, the beneficial effects of the present invention are that, by acquiring communication sample information through the sample acquisition module, and analyzing the communication sample information through the fault statistics module and the sample analysis module, sample communication characteristics are analyzed, thereby realizing feature analysis of sample communication data; by acquiring communication user information through the user acquisition module, and analyzing the communication user information through the association analysis module and the user analysis module, user communication characteristics are analyzed, thereby realizing feature analysis of user communication data; and by analyzing the sample communication characteristics and user communication characteristics through the fault prediction module, user communication faults are predicted, thereby improving the system's analysis efficiency of user communication data and improving the accuracy of user communication fault prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 This is a schematic diagram of the structure of the AI-based intelligent communication fault prediction system in this embodiment.

[0063] Figure 2 FIG. 4 is a structural diagram of the sample analysis module of this embodiment.

[0064] Figure 3 Schematic diagram of the structure of the association analysis module in this embodiment.

[0065] Figure 4 This is a structural diagram of the user analysis module of this embodiment. DETAILED DESCRIPTION

[0066] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are merely used to explain the present invention and are not intended to limit the present invention.

[0067] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0068] Furthermore, it should be noted that, in the description of the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," and "connected" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; mechanical connections or electrical connections; direct connections or indirect connections through an intermediate medium; and internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0069] See also Figure 1 As shown in FIG, an AI-based intelligent communication fault prediction system of this embodiment includes:

[0070] A sample acquisition module is used to acquire communication sample information, wherein the communication sample information includes sample signal strength, sample bandwidth information, and sample operation information. The sample bandwidth information includes sample transmission rate and sample maximum bandwidth. The unit of the sample transmission rate is kb / s, and the unit of the sample maximum bandwidth is Mb. The sample operation information includes hardware failure, network failure, and normal operation;

[0071] A fault statistics module, used for analyzing the sample failure rate according to the communication sample information, the fault statistics module is connected to the sample acquisition module;

[0072] A sample analysis module, configured to analyze sample-related parameters based on communication sample information, and further configured to analyze sample characteristic parameters based on the sample-related parameters, communication sample information, and sample failure rate, the sample analysis module being connected to the failure statistics module;

[0073] A user acquisition module is used to acquire communication user information in real time. The communication user information includes user signal strength and user bandwidth information. The user bandwidth information includes user transmission rate and user maximum bandwidth. The unit of the user transmission rate is kb / s, and the unit of the user maximum bandwidth is Mb;

[0074] An association analysis module, configured to analyze user association parameters based on communication user information, and to analyze association similarity based on user association parameters and sample association parameters, the association analysis module being connected to the user acquisition module and the sample analysis module;

[0075] A user analysis module, configured to analyze user bandwidth usage based on communication user information, and to analyze user characteristic parameters based on user bandwidth usage, communication user information, and correlation similarity, the user analysis module being connected to the correlation analysis module;

[0076] A fault prediction module, configured to analyze similar characteristic parameters based on sample characteristic parameters and user characteristic parameters, and further configured to analyze user fault risks based on similar characteristic parameters and communication sample information, the fault prediction module being connected to the user analysis module;

[0077] The prediction output module is used to output the user's failure risk, and the prediction output module is connected to the failure prediction module.

[0078] Specifically, this embodiment is applied to a server cloud where users communicate through a base station server. By analyzing the characteristics of user data and sample data in the server, prediction of user communication failures is achieved. The sample acquisition module acquires communication sample information, and the fault statistics module and the sample analysis module analyze the communication sample information to analyze the sample communication characteristics, thereby achieving feature analysis of the sample communication data. The user acquisition module acquires communication user information, and the association analysis module and the user analysis module analyze the communication user information to analyze the user communication characteristics, thereby achieving feature analysis of the user communication data. The fault prediction module analyzes the sample communication characteristics and the user communication characteristics to achieve prediction of user communication failures, thereby improving the system's analysis efficiency of user communication data and improving the accuracy of user communication failure prediction.

[0079] Specifically, the fault statistics module in this embodiment counts the number of communication sample information whose sample operation information is hardware failure as the first fault number, counts the number of communication sample information whose sample operation information is network failure as the second fault number, and counts the number of communication sample information whose sample operation information is normal operation as the third fault number, and calculates the sample failure rate through the fault analysis formula based on the first fault number, the second fault number, and the third fault number. The fault analysis module is provided with a fault analysis formula as follows:

[0080] A1=(N1 / NA) Na / NA ;

[0081] A2=(N2 / NA) Na / NA ;

[0082] Na=N1+N2+N3-NA;

[0083] Among them, A1 represents the first sample failure rate, N1 represents the first failure number, NA represents the number of communication sample information, Na represents the sample failure parameter, A2 represents the second sample failure rate, N2 represents the second failure number, and N3 represents the third failure number.

[0084] Specifically, in this embodiment, the fault statistics module analyzes the sample operation information to count the first fault number, the second fault number and the third fault number, and realizes the statistics of the communication sample information with different faults in the sample data, thereby analyzing the sample failure rate, and using the sample failure rate to represent the proportion of data with faults in the communication sample information, thereby improving the system's analysis efficiency of user communication data and improving the accuracy of user communication failure prediction.

[0085] See also Figure 2 As shown, the sample analysis module includes:

[0086] A sample bandwidth analysis unit, configured to analyze the sample bandwidth usage rate based on the sample bandwidth information;

[0087] a sample correlation analysis unit, configured to analyze sample correlation parameters according to sample signal strength, sample transmission rate, and sample operation information, the sample correlation analysis unit being connected to the sample bandwidth analysis unit;

[0088] A sample feature analysis unit, configured to analyze sample feature parameters according to sample signal strength, sample bandwidth information, and sample bandwidth usage, the sample feature analysis unit being connected to the sample association analysis unit;

[0089] A sample feature adjustment unit, configured to adjust the analysis process of the sample feature parameters according to the sample association parameters, the sample feature adjustment unit being connected to the sample feature analysis unit;

[0090] The sample feature optimization unit is used to optimize the adjustment process of the sample feature parameters according to the sample failure rate. The sample feature optimization unit is connected to the sample feature adjustment unit.

[0091] Specifically, the sample bandwidth analysis unit in this embodiment analyzes the sample bandwidth utilization rate according to the sample bandwidth information, sets the sample bandwidth utilization rate to D(i,j), and sets D(i,j)=d1(i,j) / [d2(i)×128], where D(i,j) represents the sample bandwidth utilization rate, i represents the sample number, and i∈N + And i≤NA, j represents the sample time number, j∈N + , d1(i,j) represents the sample transmission rate, and d2(i) represents the maximum bandwidth of the sample. The sample number is used to distinguish different communication sample information, and the sample time number is used to distinguish the user signal strength and user bandwidth information of the communication sample information of the same sample user at different times during communication.

[0092] Specifically, in this embodiment, the sample bandwidth analysis unit analyzes the sample bandwidth information to analyze the bandwidth utilization rate, and uses the bandwidth utilization rate to represent the occupancy of the user network, thereby improving the system's analysis efficiency of user communication data and improving the accuracy of user communication failure prediction.

[0093] Specifically, the sample association analysis unit in this embodiment extracts the number corresponding to the communication sample information whose sample operation information is a hardware fault as the first fault number, extracts the number corresponding to the communication sample information whose sample operation information is a network fault as the second fault number, and calculates the sample association parameter through the sample association analysis formula according to the sample signal strength, the sample transmission rate, the first fault number and the second fault number. The sample association analysis unit is provided with a sample association analysis formula as follows:

[0094] ,

[0095] ,

[0096] ,

[0097] ,

[0098] ,

[0099] ,

[0100] ,

[0101] Among them, R1 represents the first sample association parameter, R2 represents the second sample association parameter, R3 represents the third sample association parameter, X1(i) represents the sample signal strength parameter, x1 represents the sample signal strength mean, X2(i) represents the sample transmission rate parameter, x2 represents the sample transmission rate mean, S(i,j) represents the sample signal strength, K1 represents the set of first fault numbers, and K2 represents the set of second fault numbers.

[0102] Specifically, in this embodiment, the sample correlation analysis unit analyzes the sample signal strength, sample transmission rate and sample operation information to analyze the sample correlation parameters, and uses the sample correlation parameters to express the correlation between the signal strength and transmission rate in the communication sample information, thereby realizing a comprehensive analysis of the data correlation between the data with faults and the data running normally, thereby improving the system's analysis efficiency of user communication data and improving the accuracy of user communication fault prediction.

[0103] Specifically, the sample feature analysis unit in this embodiment calculates the sample feature parameter according to the sample signal strength, sample bandwidth information and sample bandwidth usage rate through a sample feature analysis formula. The sample feature analysis unit is provided with a sample feature analysis formula as follows:

[0104] ,

[0105] Among them, Q(i) represents the sample characteristic parameter, j max Indicates the maximum value of the sample time number.

[0106] Specifically, in this embodiment, the sample feature analysis unit analyzes the sample signal strength, sample bandwidth information and sample bandwidth usage rate to analyze the sample feature parameters, and uses the sample feature parameters to represent the characteristics of the communication sample data to achieve feature analysis of the sample data, thereby improving the system's analysis efficiency of user communication data and improving the accuracy of user communication failure prediction.

[0107] Specifically, the sample feature adjustment unit in this embodiment adjusts the analysis process of the sample feature parameters according to the sample association parameters and the sample operation information, wherein:

[0108] When the sample operation information is a hardware failure and R2 / R1>r, the sample feature adjustment unit determines that the sample feature parameter does not meet the threshold, adjusts the analysis process of the sample feature parameter, and the adjusted sample feature parameter is Q1(i), setting Q1(i)=Q(i)×(1-R2+R1);

[0109] When the sample operation information is a network failure and R3 / R1>r, the sample feature adjustment unit determines that the sample feature parameter does not meet the threshold, adjusts the analysis process of the sample feature parameter, and the adjusted sample feature parameter is Q1(i), setting Q1(i)=Q(i)×(1-R3+R1);

[0110] When R2 / R1≤r or R3 / R1≤r, the sample feature adjustment unit determines that the sample feature parameter meets the threshold value and does not adjust the analysis process of the sample feature parameter;

[0111] Where r represents the correlation threshold, 1≤r≤1.1. It is understood that the value of the correlation threshold is not specifically limited in this embodiment, and can be freely set by those skilled in the art, as long as it satisfies the adjustment of the analysis process of the sample characteristic parameters. The optimal value of the correlation threshold is: r = 1.05.

[0112] Specifically, in this embodiment, the sample feature adjustment unit analyzes the sample association parameters and the sample operation information to adjust the analysis process of the sample feature parameters, so that the adjusted sample feature parameters are correlated with the correlation of each data in the sample data. When the correlation of the data with the fault is large, the sample feature parameters are adjusted, thereby improving the system's analysis efficiency of user communication data and improving the accuracy of user communication fault prediction.

[0113] Specifically, the sample feature optimization unit in this embodiment optimizes the sample feature according to the sample failure rate.

[0114] The sample association parameters are optimized to optimize the adjustment process of the sample characteristic parameters, where:

[0115] When A1+A2≤a, the sample feature optimization unit determines that the sample failure rate meets the threshold value and does not optimize the adjustment process of the sample feature parameters;

[0116] When the sample operation information is a hardware failure and A1+A2>a, the sample feature optimization unit determines that the sample failure rate does not meet the threshold, optimizes the sample association parameter, and sets the optimized sample association parameter R1'=R1×e A1 / 2 ;

[0117] When the sample operation information is a network failure and A1+A2>a, the sample feature optimization unit determines that the sample failure rate does not meet the threshold, optimizes the sample association parameter, and sets the optimized sample association parameter R1'=R1×e A2 / 2 ;

[0118] Where a represents the failure rate threshold, and 0.05≤a≤0.15. It is understood that the present embodiment does not impose any specific restrictions on the value of the failure rate threshold, and those skilled in the art may freely set it, as long as it satisfies the optimization of the sample characteristic parameter adjustment process. The optimal value of the failure rate threshold is: a = 0.1.

[0119] Specifically, in this embodiment, the sample failure rate is analyzed by the sample feature optimization unit to optimize the sample association parameters and optimize the adjustment process of the sample feature parameters so that the optimized sample association parameters are correlated with the sample failure rate. When the sample failure rate is large, the sample association parameters are optimized to reduce the impact of a large amount of fault data in the sample data on the sample feature parameters, thereby improving the system's analysis efficiency of user communication data and improving the accuracy of user communication failure prediction.

[0120] See also Figure 3 As shown, the association analysis module includes:

[0121] A user association analysis unit, configured to analyze user association parameters based on communication user information;

[0122] The user similarity analysis unit is used to analyze the association similarity according to the user association parameter and the sample association parameter, and the user similarity analysis unit is connected to the user association analysis unit.

[0123] Specifically, in this embodiment, the user association analysis unit calculates the user association parameter according to the user signal strength and the user transmission rate through the user association analysis formula. The user association analysis unit is provided with the following user association analysis formula:

[0124] ,

[0125] ,

[0126] ,

[0127] Where Z represents the user association parameter, G(u) represents the user signal strength, u represents the user time number, u∈N + , g represents the mean user signal strength, H(u) represents the user transmission rate, h represents the mean user transmission rate, u max Indicates the maximum value of the user time number. The user time number is a number that distinguishes the communication user information of the user at different times during communication.

[0128] Specifically, in this embodiment, the user association analysis unit analyzes the communication user information to analyze the user association parameters, and uses the user association parameters to represent the correlation between the signal strength and the transmission rate in the user's communication data, thereby realizing a unified analysis of the correlation between user communication data, thereby improving the system's analysis efficiency of user communication data and improving the accuracy of user communication failure prediction.

[0129] Specifically, in this embodiment, the user similarity analysis unit calculates the association similarity according to the user association parameter and the sample association parameter using a similarity analysis formula. The user similarity analysis unit is provided with a similarity analysis formula as follows:

[0130] P=(Z / R1) Z-(R2+R3) / 2 ;

[0131] Where P represents the association similarity.

[0132] Specifically, in this embodiment, the user similarity analysis unit analyzes the user association parameters and the sample association parameters to analyze the association similarity, and uses the association similarity to represent the similarity between the user association parameters and the sample association parameters, thereby realizing the similarity comparison between the user data and the sample data, thereby improving the system's analysis efficiency of the user communication data and improving the accuracy of the prediction of user communication failures.

[0133] See also Figure 4 As shown, the user analysis module includes:

[0134] A user bandwidth analysis unit is used to analyze user bandwidth usage based on user bandwidth information;

[0135] User feature analysis unit for analyzing user characteristic parameters based on user signal strength, user bandwidth utilization and user bandwidth information, the user feature analysis unit is connected to the user bandwidth analysis unit;

[0136] The user feature adjustment unit is used to adjust the analysis process of the user feature parameters according to the correlation similarity, and the user feature adjustment unit is connected to the user feature analysis unit.

[0137] Specifically, the user bandwidth analysis unit in this embodiment analyzes the user bandwidth utilization rate according to the user bandwidth information, sets the user bandwidth utilization rate to M(u), and sets M(u)=H(u) / [128×MH], where M(u) represents the user bandwidth utilization rate and MH represents the user's maximum bandwidth.

[0138] Specifically, the user feature analysis unit in this embodiment calculates the user feature parameters according to the user signal strength, user bandwidth usage rate and user bandwidth information through the user feature analysis formula. The user feature analysis unit is provided with the following user feature analysis formula:

[0139] ,

[0140] Where F represents the user feature parameter.

[0141] Specifically, in this embodiment, the user feature analysis unit analyzes the user signal strength, user bandwidth usage and user bandwidth information to analyze user feature parameters, and uses the user feature parameters to represent the characteristics of the user communication data, so that the feature analysis of the user communication data is consistent with the feature analysis of the sample communication data, thereby improving the system's analysis efficiency of the user communication data and improving the accuracy of the prediction of user communication failures.

[0142] Specifically, the user feature adjustment unit in this embodiment adjusts the analysis process of the user feature parameters according to the correlation similarity, wherein:

[0143] When P≥p, the user feature adjustment unit determines that the correlation similarity meets the threshold value and does not adjust the analysis process of the user feature parameters;

[0144] When P < p, the user feature adjustment unit determines that the correlation similarity does not meet the threshold, and adjusts the analysis process of the user feature parameter. The adjusted user feature parameter is F1, and F1 = F / e (1-P) ;

[0145] Where p represents the similarity threshold, 1≤p≤1.2. It is understood that the value of the similarity threshold is not specifically limited in this embodiment, and those skilled in the art can freely set it as long as it satisfies the adjustment of user feature parameters. The optimal value of the similarity threshold is: p=1.1.

[0146] Specifically, in this embodiment, the user feature adjustment unit analyzes the association similarity to adjust the analysis process of the user feature parameters, so that the adjusted user feature parameters are correlated with the correlation between the user data and the sample data, thereby improving the system's analysis efficiency of user communication data and improving the accuracy of user communication failure prediction.

[0147] Specifically, the fault prediction module in this embodiment analyzes similar feature parameters based on sample feature parameters and user feature parameters, wherein:

[0148] When α1≤F / Q(i)≤α2, the fault prediction module uses the characteristic parameters of the current analysis sample as similar characteristic parameters;

[0149] When F / Q(i)<α1 or F / Q(i)>α2, the fault prediction module does not analyze similar characteristic parameters;

[0150] Wherein, α1 represents the first comparison threshold, 0.95≤α≤1.05, and α2 represents the second comparison threshold, 1.1≤α2≤1.15. It is understood that the values ​​of the comparison thresholds are not specifically limited in this embodiment, and those skilled in the art may freely set them, as long as they satisfy the analysis of similar feature parameters. The optimal values ​​of the comparison thresholds are α1=1 and α2=1.1.

[0151] Specifically, the fault prediction module in this embodiment counts the number of similar feature parameters of the sample operation information for hardware failure as the first feature number, counts the number of similar feature parameters of the sample operation information for network failure as the second feature number, and analyzes the user fault risk based on the first feature number and the second feature number, wherein:

[0152] When B1 / NB>A1, the fault prediction module sets the user fault risk to high hardware fault risk;

[0153] When B2 / NB>A2, the fault prediction module sets the user fault risk to high network fault risk;

[0154] When B1 / NB≤A1 and B2 / NB≤A2, the fault prediction module does not analyze the user fault risk;

[0155] Wherein, B1 represents the number of first features, B2 represents the number of second features, and NB represents the number of similar feature parameters.

[0156] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

Claims

1. An AI-based intelligent communication fault prediction system, characterized in that: include: A sample acquisition module, used to acquire communication sample information; A fault statistics module is used to analyze the sample failure rate based on the communication sample information; A sample analysis module, used to analyze sample-related parameters according to communication sample information, and also used to analyze sample characteristic parameters according to sample-related parameters, communication sample information and sample failure rate; User acquisition module, used to obtain communication user information in real time; An association analysis module, used to analyze user association parameters according to communication user information, and also used to analyze association similarity according to user association parameters and sample association parameters; A user analysis module is used to analyze the user bandwidth usage rate according to the communication user information, and is also used to analyze the user characteristic parameters according to the user bandwidth usage rate, the communication user information and the correlation similarity; A fault prediction module is used to analyze similar characteristic parameters according to sample characteristic parameters and user characteristic parameters, and also to analyze user fault risks according to similar characteristic parameters and communication sample information; Prediction output module, used to output user failure risk; The sample analysis module is provided with a sample correlation analysis unit, which is used to calculate the sample correlation parameter according to the sample signal strength, the sample transmission rate, the first fault number and the second fault number through the sample correlation analysis formula. The sample correlation analysis unit is provided with a sample correlation analysis formula as follows: ， ， ， ， ， ， ， Wherein, R1 represents the first sample association parameter, R2 represents the second sample association parameter, R3 represents the third sample association parameter, X1(i) represents the sample signal strength parameter, x1 represents the sample signal strength mean, X2(i) represents the sample transmission rate parameter, x2 represents the sample transmission rate mean, S(i,j) represents the sample signal strength, K1 represents the set of the first fault number, K2 represents the set of the second fault number, i represents the sample number, i∈N + And i≤NA, j represents the sample time number, j∈N + , NA represents the number of communication sample information; The sample analysis module is further provided with a sample feature analysis unit, which is used to calculate the sample feature parameters through a sample feature analysis formula according to the sample signal strength, sample bandwidth information and sample bandwidth usage rate. The sample feature analysis unit is provided with a sample feature analysis formula as follows: ， Among them, Q(i) represents the sample characteristic parameter, j max represents the maximum value of the sample time number, D(i,j) represents the sample bandwidth utilization, and d1(i,j) represents the sample transmission rate.

2. The AI-based intelligent communication fault prediction system according to claim 1, characterized in that: The fault statistics module calculates the sample failure rate according to the first fault number, the second fault number and the third fault number through a fault analysis formula. The fault statistics module is provided with a fault analysis formula as follows: A1=(N1 / NA) Na / NA ; <h2 style=";text-align:left;direction:ltr">A2=(N2 / NA)<h2 style=";text-align:left;direction:ltr"> Na / NA <h2 style=";text-align:left;direction:ltr"> ; Na=N1+N2+N3-NA; Among them, A1 represents the first sample failure rate, N1 represents the first failure number, NA represents the number of communication sample information, Na represents the sample failure parameter, A2 represents the second sample failure rate, N2 represents the second failure number, and N3 represents the third failure number.

3. The AI-based intelligent communication fault prediction system according to claim 1, characterized in that: The sample analysis module is also provided with a sample characteristic adjustment unit, which is used to adjust the analysis process of the sample characteristic parameters according to the sample association parameters and the sample operation information, wherein: When the sample operation information is a hardware failure and R2 / R1>r, the sample feature adjustment unit determines that the sample feature parameter does not meet the threshold, adjusts the analysis process of the sample feature parameter, and the adjusted sample feature parameter is Q1(i), setting Q1(i)=Q(i)×(1-R2+R1); When the sample operation information is a network failure and R3 / R1>r, the sample feature adjustment unit determines that the sample feature parameter does not meet the threshold, adjusts the analysis process of the sample feature parameter, and the adjusted sample feature parameter is Q1(i), setting Q1(i)=Q(i)×(1-R3+R1); When R2 / R1≤r or R3 / R1≤r, the sample feature adjustment unit determines that the sample feature parameter meets the threshold value and does not adjust the analysis process of the sample feature parameter; Among them, r represents the correlation threshold; The sample analysis module is further provided with a sample feature optimization unit, which is used to optimize the sample associated parameters according to the sample failure rate, so as to optimize the adjustment process of the sample feature parameters, wherein: When A1+A2≤a, the sample feature optimization unit determines that the sample failure rate meets the threshold value and does not optimize the adjustment process of the sample feature parameters; When the sample operation information is a hardware failure and A1+A2>a, the sample feature optimization unit determines that the sample failure rate does not meet the threshold, optimizes the sample association parameter, and the optimized sample association parameter R1' is set to R1'=R1×e A1 / 2 ; When the sample operation information is a network failure and A1+A2>a, the sample feature optimization unit determines that the sample failure rate does not meet the threshold, optimizes the sample association parameter, and sets R1'=R1×e A2 / 2 ; Wherein, a represents the failure rate threshold.

4. The AI-based intelligent communication fault prediction system according to claim 3 is characterized in that: The association analysis module is provided with a user association analysis unit, which is used to calculate the user association parameter according to the user signal strength and the user transmission rate through the user association analysis formula. The user association analysis unit is provided with the user association analysis formula as follows: ， ， ， Where Z represents the user association parameter, G(u) represents the user signal strength, u represents the user time number, and u∈N + , g represents the mean value of user signal strength, H(u) represents the user transmission rate, h represents the mean value of user transmission rate, u max Indicates the maximum value of the user time number.

5. The AI-based intelligent communication fault prediction system according to claim 4 is characterized in that: The association analysis module is further provided with a user similarity analysis unit, which is used to calculate the association similarity according to the user association parameter and the sample association parameter through a similarity analysis formula. The user similarity analysis unit is provided with a similarity analysis formula as follows: P=(Z / R1) Z-(R2+R3) / 2 ; Among them, P represents the association similarity.

6. The AI-based intelligent communication fault prediction system according to claim 5, characterized in that: The user analysis module is provided with a user feature analysis unit, which is used to calculate user feature parameters through a user feature analysis formula according to user signal strength, user bandwidth usage rate and user bandwidth information. The user feature analysis unit is provided with a user feature analysis formula as follows: ， Wherein, F represents the user characteristic parameter, and M(u) represents the user bandwidth utilization rate.

7. The AI-based intelligent communication fault prediction system according to claim 6, characterized in that: The user analysis module is also provided with a user feature adjustment unit, which is used to adjust the analysis process of the user feature parameters according to the association similarity, wherein: When P≥p, the user feature adjustment unit determines that the correlation similarity meets the threshold value and does not adjust the analysis process of the user feature parameters; When P<p, the user feature adjustment unit determines that the correlation similarity does not meet the threshold value, and adjusts the analysis process of the user feature parameter. The adjusted user feature parameter is F1, and F1=F / e is set. (1-P) ; Where p represents the similarity threshold.

8. The AI-based intelligent communication fault prediction system according to claim 7, characterized in that: The fault prediction module analyzes the user fault risk according to the first feature quantity and the second feature quantity, wherein: When B1 / NB>A1, the fault prediction module sets the user fault risk to high hardware fault risk; When B2 / NB>A2, the fault prediction module sets the user fault risk to high network fault risk; When B1 / NB≤A1 and B2 / NB≤A2, the fault prediction module does not analyze the user fault risk; Among them, B1 represents the number of first features, B2 represents the number of second features, and NB represents the number of similar feature parameters.

Images