IT operation and maintenance management software operation monitoring method based on artificial intelligence

Abstract

The invention discloses an IT operation and maintenance management software operation monitoring method based on artificial intelligence, and relates to the technical field of intelligent operation and maintenance. The method comprises the following steps: constructing an IT operation and maintenance management overall framework; the overall architecture comprises a data acquisition layer, a communication layer, a data storage layer, a logic calculation layer and a user service layer. IT system data are collected in real time and preprocessed; classifying the data into basic data, operation data, security data, performance data and business data; establishing a BP neural network model to carry out operation and maintenance fault diagnosis, and obtaining a fault category, a fault position and a fault evaluation coefficient; fault reason judgment is carried out by utilizing association rule mining, and fault repair work is guided; comprehensive monitoring and intelligent operation and maintenance of the IT system are achieved, system faults can be diagnosed, performance bottleneck can be predicted, operation and maintenance efficiency and system stability are effectively improved, fault response time is shortened, operation cost is reduced, and meanwhile long-term efficient operation of the system is guaranteed through continuous performance optimization suggestions.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent operation and maintenance technology, and specifically relates to an operation monitoring method of IT operation and maintenance management software based on artificial intelligence. Background Art

[0002] With the rapid development of information technology, IT systems have become increasingly complex, covering multiple components such as servers, network equipment, databases, and applications. The interactions between these components are also becoming increasingly complex.

[0003] Traditional IT operation and maintenance methods often rely on manual monitoring and manual intervention, which makes it difficult to cope with large-scale, highly complex IT system operation and maintenance needs. Traditional IT operation and maintenance management often has delays in fault discovery and location, and is easily affected by human factors, resulting in low operation and maintenance efficiency and insufficient accuracy. In order to improve operation and maintenance efficiency and accuracy, it is necessary to introduce intelligent monitoring and management methods that can detect and handle faults and problems in IT systems in real time and accurately.

[0004] With the rapid development of big data and artificial intelligence technologies, these technologies have been widely used in various fields, including IT operation and maintenance management. By collecting and analyzing large amounts of data in IT systems and using artificial intelligence technologies such as machine learning and deep learning, it is possible to achieve intelligent monitoring and management of IT systems and improve the automation and intelligence level of operation and maintenance.

[0005] With the intensification of market competition and the continuous changes in business needs, enterprises have higher and higher requirements for the stability and reliability of IT systems. In order to meet these requirements, enterprises need to introduce more efficient and intelligent IT operation and maintenance management methods to improve the stability and reliability of the system and ensure the normal operation of the business. Summary of the invention

[0006] In order to overcome the shortcomings and deficiencies of the above-mentioned prior art, the present invention adopts the following technical solutions:

[0007] The operation monitoring method of IT operation and maintenance management software based on artificial intelligence, the method flow is as follows:

[0008] Step 1: Build an overall IT operation and maintenance management architecture; the overall architecture includes data collection layer, communication layer, data storage layer, logic computing layer and user service layer;

[0009] Step 2: Collect IT system data in real time and pre-process it;

[0010] Step 3: Classify the data into basic data, operational data, security data, performance data and business data;

[0011] Step 4: Establish a BP neural network model to perform IT operation and maintenance fault diagnosis and obtain the fault category, fault location and fault assessment coefficient;

[0012] Step 5: Use association rule mining to determine the cause of the fault and guide the fault repair work;

[0013] Step 6: Performance optimization and preventive measures.

[0014] Preferably, the data collection layer collects various types of data in the IT system in real time through various data sources including API interfaces, log collection tools and network listeners; the communication layer uses Ethernet, wireless networks and message queue technology to transmit the collected data to the monitoring system; the data storage layer stores the collected data to ensure the persistence and high availability of the data; the logical computing layer analyzes and processes the collected data, including data cleaning, feature extraction and model training, and uses a machine learning framework to build fault diagnosis models and performance optimization models; the user service layer provides real-time monitoring of IT system data, data visualization and user management functions.

[0015] Preferably, the preprocessing includes data cleaning, denoising and feature encoding;

[0016] Data cleaning: remove invalid records and correct erroneous data;

[0017] Denoising: eliminate noise interference and retain effective information;

[0018] Feature encoding: converting non-numeric data into numeric data to facilitate processing by machine learning algorithms;

[0019] Dataset division: The preprocessed data is divided into training set, validation set and test set, which are used for model training, parameter adjustment and final performance evaluation respectively.

[0020] As a preference, step three is specifically as follows:

[0021] According to the function, the data is classified into basic data, operation data, security data, performance data and business data. The number of data sets to be classified is determined to be 5, and 5 data points are randomly selected as the initial clustering centers;

[0022] The coordinates of each data point (x i ,y i ) and the coordinates of the initial cluster center (x j ,x j ) is substituted into the calculation formula to obtain the distance between the five initial cluster centers and assign it to the cluster center closest to it. The calculation formula is as follows:

[0023]

[0024] Where i represents the number of non-initial cluster center points, i is a natural number greater than 0; j represents the number of initial cluster center points, and the value of j is 1, 2, 3, 4, 5;

[0025] Update the initial cluster center coordinates and change the coordinates of each data point (x i ,y i ) is substituted into the calculation formula to obtain the coordinates of the new cluster center θ1, which is as follows:

[0026]

[0027] Where a represents the number of data points in the data set;

[0028] The coordinates of each data point (x i ,y i ) is substituted into the calculation formula to determine whether the optimal result is achieved. The calculation formula is as follows:

[0029]

[0030] Among them, A is the sum of squared errors of data points in the data set, which represents the sum of the distances from all data points to their cluster centers; Q is a data point, which represents a given object; e represents the mean of the cluster center points; the minimum value of e indicates that the optimal result is achieved;

[0031] Repeat the above steps until the cluster center point no longer changes or reaches the predetermined number of iterations, and the data clustering is completed.

[0032] As a preference, step four is specifically as follows:

[0033] S41, setting the initial parameters of the neural network, including the number of nodes in the input layer, hidden layer and output layer, and the initial values ​​of weights and biases; setting the BP neural network to a structure of 5 inputs and 3 outputs;

[0034] S42, using five items of data, namely, basic data, operation data, safety data, performance data and business data, as input parameters of the BP neural network, and the corresponding fault category, fault location and fault assessment coefficient as output layer parameters, and calculating the node output value of each layer;

[0035] S43, multiply the input data by the initialization weight matrix from the input layer to the hidden layer, that is

[0036] Z = Y * W;

[0037] Among them, Y represents the input matrix, Z represents the output matrix, and W represents the initial weight matrix; the initial weight matrix W is a small value that obeys the normal distribution;

[0038] After obtaining the output matrix, add the bias of the hidden layer and perform nonlinear transformation through the Sigmoid function to obtain the output value of the hidden layer. The formula of the Sigmoid function is as follows:

[0039]

[0040] Where, e is the base of natural logarithm, and x is the input value;

[0041] S44, multiplying the output value of the hidden layer by the weight matrix from the hidden layer to the output layer, adding the bias of the output layer, and obtaining the output value of the output layer through the Sigmoid function again;

[0042] S45, calculate the loss function, compare the output value of the BP neural network with the actual label value, and obtain the error value; the loss function calculation formula is as follows:

[0043]

[0044] Among them, X i ,Y i are the total measured and predicted parameters when the number of elements is i; n is the number of elements in the test sample; when E RMSE When the value is minimum, the number of neurons in the hidden layer is determined;

[0045] S46. According to the value of the loss function, the gradient of each layer is calculated by the chain rule; first, the gradient of the output layer is calculated, that is, the derivative of the output error with respect to the input of the output layer; then, the gradient of the output layer is multiplied by the transpose of the weight matrix from the output layer to the hidden layer to obtain the gradient of the hidden layer; finally, the value of the weight matrix and the bias is updated according to the gradient and the learning rate;

[0046] S47, repeating steps S42 to S46 until the change in the loss function is less than a set threshold;

[0047] S48. Use the trained BP neural network model to perform fault diagnosis, input the test samples into the neural network, obtain the output value through forward propagation, and establish a stable BP neural network model through continuous training to obtain the fault category, fault location and fault assessment coefficient; wherein the value of the fault assessment coefficient is a number greater than 0 and less than or equal to 1, which is used to quantify the severity of the fault.

[0048] As a preference, step five is specifically as follows:

[0049] S51, scanning the system operation and maintenance fault database data and fault data, and counting the support of each item; wherein the support is the frequency of occurrence of the item set;

[0050] S52, selecting a first-level frequent item set according to the support threshold;

[0051] S53, generating a candidate second-level option set based on the first-level frequent item set, and counting its support;

[0052] S54, selecting secondary frequent itemsets according to the support threshold;

[0053] S55. Generate candidate third-level option sets based on the second-level frequent item sets, and count their support;

[0054] S56, repeat the above process until no more candidate item sets can be generated;

[0055] S57. Generate association rules based on frequent item sets and calculate their confidence; wherein the confidence is the degree of association between the fault data and the fault causes in the system operation and maintenance fault database data;

[0056] S58. Select an association rule that meets the requirements according to the confidence threshold;

[0057] S59. Use the generated association rules to quickly locate the cause of the fault and guide the fault repair work; wherein the association rules are association rules that meet the minimum support and minimum confidence.

[0058] Preferably, the performance optimization in step six includes adjusting resource configuration and optimizing code logic.

[0059] Preferably, the preventive measures in step six include regularly evaluating the health of the system, predicting potential risk points, and taking measures in advance to avoid failures.

[0060] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0061] 1. The present invention can quickly and accurately diagnose the fault type, fault location and fault severity through the BP neural network model, reducing the fault response time; using association rule mining technology, quickly find the root cause of the fault, guide the operation and maintenance personnel to carry out targeted repair work, and improve the efficiency of fault handling.

[0062] 2. Based on the fault diagnosis results and system operation data, the present invention proposes optimization suggestions, such as adjusting resource configuration, optimizing code logic, etc., to improve the overall performance of the system; regularly evaluates the system health status, predicts potential risk points, and takes measures in advance to avoid faults and ensure the stability and reliability of the system.

[0063] 3. The present invention reduces manual intervention, improves operation and maintenance efficiency, reduces operation and maintenance costs through intelligent IT operation and maintenance management, provides real-time system monitoring and data visualization functions, enables users to intuitively understand the system status, and improves user experience and management efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0065] Figure 1 A flow chart of an artificial intelligence-based IT operation and maintenance management software operation monitoring method of the present invention is shown;

[0066] Figure 2 The flowchart of the present invention for training the BP neural network model is shown;

[0067] Figure 3 The flowchart of the present invention using association rule mining to determine the cause of a fault is shown. DETAILED DESCRIPTION

[0068] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0069] In addition, the described features, structures or characteristics may be combined in one or more example embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the example embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, steps, etc. may be adopted. In other cases, well-known structures, methods, implementations or operations are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0070] Embodiment 1:

[0071] See also Figure 1 As shown, the operation monitoring method of IT operation and maintenance management software based on artificial intelligence in this embodiment has the following process:

[0072] Step 1: Build an overall IT operation and maintenance management architecture.

[0073] Data collection layer: responsible for real-time collection of various data in the IT system, including but not limited to server logs, application logs, network traffic data, performance indicators, etc. The data collection layer connects to various devices and services in the IT system through API interfaces, log collection tools (such as Logstash), network monitors, etc., and converts the collected data into digital signals.

[0074] Communication layer: Utilizes technologies such as Ethernet, wireless networks, and message queues (such as Kafka) to transmit the collected data to the monitoring system through the communication network, and is responsible for efficient data transmission and communication with the upper layer.

[0075] Data storage layer: stores the collected data, including data processing chips, storage devices (such as Hadoop HDFS, Amazon S3) and storage management software (such as Elasticsearch), ensuring data persistence and high availability.

[0076] Logical computing layer: Analyze and process the collected data, including data cleaning, feature extraction, model training, etc. Use machine learning frameworks (such as TensorFlow, Scikit-learn) to build fault diagnosis models and performance optimization models.

[0077] User service layer: Provides real-time monitoring of IT system data, data visualization and user management functions. Through front-end display platforms such as web portals and mobile applications, it displays various system parameters, generates reports, and performs data analysis and fault analysis.

[0078] Step 2: Data collection and preprocessing.

[0079] Data collection: Collect IT system data in real time from various data sources (such as server logs, application logs, network traffic data, etc.).

[0080] Preprocessing includes data cleaning, denoising and feature encoding.

[0081] Data cleaning: remove invalid records and correct erroneous data.

[0082] Denoising: Eliminate noise interference and retain valid information.

[0083] Feature encoding: Convert non-numeric data into numeric data to facilitate processing by machine learning algorithms.

[0084] Dataset division: The preprocessed data is divided into training set, validation set and test set, which are used for model training, parameter adjustment and final performance evaluation respectively.

[0085] Step 3: Data clustering.

[0086] The data are classified into basic data, operation data, security data, performance data and business data according to their functions. The number of data sets to be classified is determined to be 5, and 5 data points are randomly selected as the initial clustering centers.

[0087] The coordinates of each data point (x i ,y i ) and the coordinates of the initial cluster center (x j ,x j ) is substituted into the calculation formula to obtain the distance between the five initial cluster centers and assign it to the cluster center closest to it. The calculation formula is as follows:

[0088]

[0089] Where i represents the number of non-initial cluster center points, i is a natural number greater than 0; j represents the number of initial cluster center points, and the value of j is 1, 2, 3, 4, 5;

[0090] Update the initial cluster center coordinates and change the coordinates of each data point (x i ,y i ) is substituted into the calculation formula to obtain the coordinates of the new cluster center θ1, which is as follows:

[0091]

[0092] Where a represents the number of data points in the data set;

[0093] The coordinates of each data point (x i ,y i ) is substituted into the calculation formula to determine whether the optimal result is achieved. The calculation formula is as follows:

[0094]

[0095] Among them, A is the sum of squared errors of data points in the data set, which represents the sum of the distances from all data points to their cluster centers; Q is a data point, which represents a given object; e represents the mean of the cluster center points; the minimum value of e indicates that the optimal result is achieved;

[0096] Repeat the above steps until the cluster center point no longer changes or reaches the predetermined number of iterations, and the data clustering is completed.

[0097] Step 4: Establish a BP neural network model for fault diagnosis.

[0098] See also Figure 2 As shown, the training process is as follows:

[0099] S41, setting the initial parameters of the neural network, including the number of nodes in the input layer, hidden layer and output layer, and the initial values ​​of weights and biases. Set the BP neural network to a 5-input, 3-output structure.

[0100] S42, using basic data, operation data, safety data, performance data and business data as input parameters of the BP neural network, and the corresponding fault category, fault location and fault assessment coefficient as output layer parameters, and calculating the node output value of each layer.

[0101] S43, multiply the input data by the initialization weight matrix from the input layer to the hidden layer, that is

[0102] Z = Y * W;

[0103] Among them, Y represents the input matrix, Z represents the output matrix, and W represents the initial weight matrix; the initial weight matrix W is a small value that obeys the normal distribution;

[0104] After obtaining the output matrix, add the bias of the hidden layer and perform nonlinear transformation through the Sigmoid function to obtain the output value of the hidden layer. The formula of the Sigmoid function is as follows:

[0105]

[0106] Where, e is the base of natural logarithm, and x is the input value;

[0107] S44. Multiply the output value of the hidden layer by the weight matrix from the hidden layer to the output layer, add the bias of the output layer, and obtain the output value of the output layer through the Sigmoid function again.

[0108] S45, calculate the loss function, compare the output value of the BP neural network with the actual label value, and obtain the error value. The loss function calculation formula is as follows:

[0109]

[0110] Among them, X i ,Y i are the total measured and predicted parameters when the number of elements is i; n is the number of elements in the test sample; when E RMSE When the value is minimum, the number of neurons in the hidden layer is determined.

[0111] S46. According to the value of the loss function, the gradient of each layer is calculated by the chain rule. First, the gradient of the output layer is calculated, that is, the derivative of the output error with respect to the input of the output layer; then, the gradient of the output layer is multiplied by the transpose of the weight matrix from the output layer to the hidden layer to obtain the gradient of the hidden layer; finally, the value of the weight matrix and bias is updated according to the gradient and the learning rate.

[0112] S47. Repeat steps S42 to S46 until the change in the loss function is less than a set threshold.

[0113] S48, using the trained BP neural network model to perform fault diagnosis, input the test sample into the neural network, obtain the output value through forward propagation, and establish a stable BP neural network model through continuous training to obtain the fault category, fault location and fault assessment coefficient. The value of the fault assessment coefficient is a number greater than 0 and less than or equal to 1, which is used to quantify the severity of the fault.

[0114] Step 5: Use association rule mining to determine the cause of the fault.

[0115] See also Figure 3 As shown in the figure, the use of association rule mining to determine the cause of the fault is as follows:

[0116] S51, scanning the system operation and maintenance fault database data and fault data, and counting the support of each item, wherein the support is the frequency of occurrence of the item set.

[0117] S52. Select a first-level frequent itemset according to the support threshold.

[0118] S53. Generate candidate second-level option sets based on the first-level frequent item sets, and count their support.

[0119] S54. Select the secondary frequent itemsets according to the support threshold.

[0120] S55. Generate candidate third-level option sets based on the second-level frequent item sets, and count their support.

[0121] S56. Repeat the above process until no more candidate item sets can be generated.

[0122] S57: Generate association rules based on frequent item sets and calculate their confidence, wherein the confidence is the degree of association between the fault data and the fault causes in the system operation and maintenance fault database data.

[0123] S58. According to the confidence threshold, select an association rule that meets the requirements.

[0124] S59: Use the generated association rules to quickly locate the cause of the fault and guide the fault repair work. The association rules are association rules that meet the minimum support and minimum confidence.

[0125] Step 6: Performance optimization and preventive measures.

[0126] Based on fault diagnosis results and system operation data, optimization suggestions are put forward, such as adjusting resource configuration, optimizing code logic, etc.

[0127] Regularly evaluate system health, predict potential risk points, and take measures in advance to avoid failures.

[0128] The beneficial effects of this embodiment are: by building a multi-layer architecture to achieve comprehensive monitoring and intelligent operation and maintenance of the IT system, using technical means such as data collection, preprocessing, cluster analysis, BP neural network modeling and association rule mining, it can accurately diagnose system faults and predict performance bottlenecks, effectively improving IT operation and maintenance efficiency and system stability, reducing fault response time, and reducing operating costs. At the same time, through continuous performance optimization suggestions, the long-term and efficient operation of the system is guaranteed.

[0129] The weight coefficients of the present invention are used to measure the degree of influence of different factors or variables on a certain result or decision. The definition of weight coefficient refers to the value assigned to each factor when comparing and evaluating multiple factors to reflect its importance or priority. These weight coefficients can be determined according to specific circumstances and needs, and are usually jointly formulated and confirmed by professionals or relevant stakeholders. By reasonably setting the weight coefficients, it can help the program or system to make decisions or predictions more accurately.

[0130] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

[0131] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. An artificial intelligence-based IT operation and maintenance management software operation monitoring method, characterized in that: The method flow is as follows: Step 1: Build an overall IT operation and maintenance management architecture; the overall architecture includes data collection layer, communication layer, data storage layer, logic computing layer and user service layer; Step 2: Collect IT system data in real time and pre-process it; Step 3: Classify the data into basic data, operational data, security data, performance data and business data; Step 4: Establish a BP neural network model to perform IT operation and maintenance fault diagnosis and obtain the fault category, fault location and fault assessment coefficient; Step 5: Use association rule mining to determine the cause of the fault and guide the fault repair work; Step 6: Performance optimization and preventive measures.

2. The method for monitoring the operation of IT operation and maintenance management software based on artificial intelligence according to claim 1, characterized in that: The data collection layer collects various data in the IT system in real time through various data sources including API interfaces, log collection tools and network monitors; The communication layer uses Ethernet, wireless network and message queue technology to transmit the collected data to the monitoring system; The data storage layer stores the collected data to ensure data persistence and high availability; The logic computing layer analyzes and processes the collected data, including data cleaning, feature extraction and model training, and uses a machine learning framework to build a fault diagnosis model and a performance optimization model; The user service layer provides real-time monitoring of IT system data, data visualization and user management functions.

3. The method for monitoring the operation of IT operation and maintenance management software based on artificial intelligence according to claim 1, characterized in that: The preprocessing includes data cleaning, denoising and feature encoding; Data cleaning: remove invalid records and correct erroneous data; Denoising: eliminate noise interference and retain effective information; Feature encoding: converting non-numeric data into numeric data to facilitate processing by machine learning algorithms; Dataset division: The preprocessed data is divided into training set, validation set and test set, which are used for model training, parameter adjustment and final performance evaluation respectively.

4. The method for monitoring the operation of IT operation and maintenance management software based on artificial intelligence according to claim 1, characterized in that: The step three is as follows: According to the function, the data is classified into basic data, operation data, security data, performance data and business data. The number of data sets to be classified is determined to be 5, and 5 data points are randomly selected as the initial clustering centers; The coordinates of each data point (x i ,y i ) and the coordinates of the initial cluster center (x j ,x j ) is substituted into the calculation formula to obtain the distance between the five initial cluster centers and assign it to the cluster center closest to it. The calculation formula is as follows: Where i represents the number of non-initial cluster center points, i is a natural number greater than 0; j represents the number of initial cluster center points, and the value of j is 1, 2, 3, 4, 5; Update the initial cluster center coordinates and change the coordinates of each data point (x i ,y i ) is substituted into the calculation formula to obtain the coordinates of the new cluster center θ1, which is as follows: Where a represents the number of data points in the data set; The coordinates of each data point (x i ,y i ) is substituted into the calculation formula to determine whether the optimal result is achieved. The calculation formula is as follows: Among them, A is the sum of squared errors of data points in the data set, which represents the sum of the distances from all data points to their cluster centers; Q is a data point, which represents a given object; e represents the mean of the cluster center points; the minimum value of e indicates that the optimal result is achieved; Repeat the above steps until the cluster center point no longer changes or reaches the predetermined number of iterations, and the data clustering is completed.

5. The method for monitoring the operation of IT operation and maintenance management software based on artificial intelligence according to claim 1, characterized in that: The step 4 is as follows: S41, setting the initial parameters of the neural network, including the number of nodes in the input layer, hidden layer and output layer, and the initial values ​​of weights and biases; setting the BP neural network to a structure of 5 inputs and 3 outputs; S42, using five items of data, namely, basic data, operation data, safety data, performance data and business data, as input parameters of the BP neural network, and the corresponding fault category, fault location and fault assessment coefficient as output layer parameters, and calculating the node output value of each layer; S43, multiply the input data by the initialization weight matrix from the input layer to the hidden layer, that is Z = Y * W; Among them, Y represents the input matrix, Z represents the output matrix, and W represents the initial weight matrix; the initial weight matrix W is a small value that obeys the normal distribution; After obtaining the output matrix, add the bias of the hidden layer and perform nonlinear transformation through the Sigmoid function to obtain the output value of the hidden layer. The formula of the Sigmoid function is as follows: Where, e is the base of natural logarithm, and x is the input value; S44, multiplying the output value of the hidden layer by the weight matrix from the hidden layer to the output layer, adding the bias of the output layer, and obtaining the output value of the output layer through the Sigmoid function again; S45, calculate the loss function, compare the output value of the BP neural network with the actual label value, and obtain the error value; the loss function calculation formula is as follows: Among them, X i ,Y i are the total measured and predicted parameters when the number of elements is i; n is the number of elements in the test sample; when E RMSE When the value is minimum, the number of neurons in the hidden layer is determined; S46. According to the value of the loss function, the gradient of each layer is calculated by the chain rule; first, the gradient of the output layer is calculated, that is, the derivative of the output error with respect to the input of the output layer; then, the gradient of the output layer is multiplied by the transpose of the weight matrix from the output layer to the hidden layer to obtain the gradient of the hidden layer; finally, the value of the weight matrix and the bias is updated according to the gradient and the learning rate; S47, repeating steps S42 to S46 until the change in the loss function is less than a set threshold; S48. Use the trained BP neural network model to perform fault diagnosis, input the test samples into the neural network, obtain the output value through forward propagation, and establish a stable BP neural network model through continuous training to obtain the fault category, fault location and fault assessment coefficient; wherein the value of the fault assessment coefficient is a number greater than 0 and less than or equal to 1, which is used to quantify the severity of the fault.

6. The method for monitoring the operation of IT operation and maintenance management software based on artificial intelligence according to claim 1, characterized in that: The step five is as follows: S51, scanning the system operation and maintenance fault database data and fault data, and counting the support of each item; wherein the support is the frequency of occurrence of the item set; S52, selecting a first-level frequent item set according to the support threshold; S53, generating a candidate second-level option set based on the first-level frequent item set, and counting its support; S54, selecting secondary frequent itemsets according to the support threshold; S55. Generate candidate third-level option sets based on the second-level frequent item sets, and count their support; S56, repeat the above process until no more candidate item sets can be generated; S57. Generate association rules based on frequent item sets and calculate their confidence; wherein the confidence is the degree of association between the fault data and the fault causes in the system operation and maintenance fault database data; S58. Select an association rule that meets the requirements according to the confidence threshold; S59. Use the generated association rules to quickly locate the cause of the fault and guide the fault repair work; wherein the association rules are association rules that meet the minimum support and minimum confidence.

7. The method for monitoring the operation of IT operation and maintenance management software based on artificial intelligence according to claim 1, characterized in that: The performance optimization in step six includes adjusting resource configuration and optimizing code logic.

8. The method for monitoring the operation of IT operation and maintenance management software based on artificial intelligence according to claim 1, characterized in that: The preventive measures described in step six include regularly evaluating the health of the system, predicting potential risk points, and taking measures in advance to avoid failures.

Images