A service system guarantee method and device and a storage medium

Abstract

The present application relates to the technical field of IT support, in particular to a business system guarantee method, device, equipment and storage medium. The present application screens common behavior data from user historical behavior data, effectively avoids accidental data from affecting the precision of model training; extracts the key fields of common behavior data, differentially processes the training set, effectively simplifies the data complexity, and guarantees the data integrity and stability of the training set, improves the accuracy of the prediction result; adopts the integrated moving average autoregressive model framework to train the user behavior prediction model, the model framework is simple, and the technical difficulty can be effectively reduced; the present application can output operation and maintenance guarantee list according to the importance of the business system, can alert operation and maintenance personnel to develop operation and maintenance plan and reasonably allocate manpower as needed, and also provides different operation and maintenance guarantee suggestions for different business systems, effectively assists operation and maintenance personnel to efficiently operate, and improves the user experience of the business system.

Description

Technical Field

[0001] This invention relates to the field of IT support technology, and in particular to a business system assurance method, apparatus, equipment, and storage medium. Background Technology

[0002] As people become increasingly reliant on online life, more and more IT systems are emerging, making the operation and maintenance (O&M) of these systems and scenarios extremely important. IT systems are generally maintained through routine inspections, with intensive O&M support provided to certain systems during holidays and important events. Traditional business system O&M methods involve periodically checking network performance, business access, office peripherals, data centers, and servers to identify and eliminate potential risks. However, the sheer number of business systems, the vast amount of data to be checked, and the high computational complexity often lead to a failure to promptly identify and resolve problems, resulting in a poor user experience.

[0003] The closest prior art to this application is Chinese patent CN201811427877.3, which discloses a technical solution: collecting user operation behavior data through bypass monitoring traffic interception, obtaining traffic data from a metropolitan area network switch through a mirror port, transmitting the traffic data to an SDN server, using SDN technology to strip away the MPLS multi-layer protocol, and then transmitting the user operation behavior data to the collection server, and finally using a decision tree to model based on the user operation behavior data to achieve fault and performance prediction. This prior art focuses on using traffic data to collect user operation behavior data and the preliminary data processing, but it does not describe the process of modeling through decision trees in detail. Moreover, building a decision tree model is prone to overfitting, resulting in low generalization ability of the model, and the calculation result of the decision tree is a local optimum rather than a global optimum, making it difficult to guarantee the accuracy of the prediction results.

[0004] The disadvantages of existing technology CN201811427877.3:

[0005] 1. It fails to effectively guarantee the accuracy of the prediction results, and the existing technology requires training multiple weak classifiers separately and going through N iterations to obtain a strong classifier, which is computationally complex and computationally intensive.

[0006] 2. The technical solution was not designed in accordance with the actual operation and maintenance scenario. Although the background technology section mentions operation and maintenance, the technical solution is actually divorced from the operation and maintenance scenario.

[0007] 3. No effective safeguards are provided for the predicted results. The existing technology does not list the possible predicted results, nor does it provide different operation and maintenance safeguards for different results. Summary of the Invention

[0008] Therefore, the technical problem to be solved by the present invention is to overcome the problems in the prior art that fail to effectively guarantee the accuracy of prediction results, fail to build technical solutions in combination with actual operation and maintenance scenarios, and fail to provide effective guarantee suggestions for prediction results.

[0009] To address the aforementioned technical problems, this invention provides a business system assurance method, comprising:

[0010] Obtain user historical behavior data;

[0011] The user's historical behavior data is divided into occasional behavior data and frequent behavior data, and a relationship database between user behavior and business system is established for the frequent behavior data.

[0012] Arima parameters are determined based on the frequently occurring behavior data, and a pre-built user behavior prediction model is trained using an integrated moving average autoregressive model framework.

[0013] The trained user behavior prediction model is used to predict future user behavior, and the usage trends of each business system are obtained based on the prediction results and the relationship database between the user behavior and the business system.

[0014] Based on the usage trends and importance ranking of each business system, a list of business systems to be supported and operational support recommendations are generated.

[0015] Preferably, obtaining user historical behavior data includes:

[0016] By crawling or querying logs, user behavior data, the business systems involved in user behavior, and the main time periods occupied by the business systems involved in user behavior are obtained from each business system, thus obtaining historical user behavior data.

[0017] Preferably, dividing the user's historical behavior data into incidental behavior data and frequent behavior data includes:

[0018] The time window size is determined based on the user's historical behavior data;

[0019] Based on the size of the time window, the user's historical behavior data is grouped in chronological order;

[0020] Calculate the frequency of a specific user behavior in each time window and compare the frequency with a preset threshold;

[0021] When the frequency is not less than the preset threshold, the user-specific behavior is considered a recurring behavior.

[0022] Preferably, establishing a database linking user behavior and business systems based on the frequently occurring behavior data includes:

[0023] Based on the frequently occurring behavior data, key fields such as the occurrence time of the frequently occurring behavior, user behavior, number of users, business systems involved in the user behavior, and the main time periods occupied by the business systems are extracted, and the relationship between user behavior and business systems is established through topology graphs, knowledge graphs, or association charts.

[0024] Each relationship chain representing the relationship between user behavior and the business system is configured with a unique identifier, and a relationship database is established.

[0025] Preferably, the step of determining the Arima parameter based on the frequent behavior data and training the pre-built user behavior prediction model using an integrated moving average autoregressive model framework includes:

[0026] The frequently occurring behavior data is cleaned, and a time index is set to obtain the training set by sorting it by time.

[0027] The training set is differentially processed, and the number of differences for the Arima parameter is determined based on the stationarity of the data.

[0028] The autoregressive coefficients and moving average coefficients of the Arima parameters are determined based on the AIC, BIC, and HQIC information criteria.

[0029] Based on the determined Arima parameters, after differential processing of the training set, the integrated moving average autoregressive model framework is input for time series algorithm training to predict user behavior and the corresponding number of users at a future time.

[0030] Preferably, the step of using a trained user behavior prediction model to predict future user behavior, and obtaining the usage trends of each business system based on the prediction results and the relationship database between user behavior and business systems, includes:

[0031] Predict future user behavior and the corresponding number of users using a trained user behavior prediction model;

[0032] Based on the prediction results, and according to the database of the relationship between user behavior and business systems, the business systems involved in the user behavior and the main time periods occupied by the business systems are obtained.

[0033] Based on the business systems involved in the user behavior and the main time periods occupied by the business systems, the future access volume and peak time of each business system can be obtained.

[0034] Preferably, generating the business system support list and operation and maintenance support recommendations based on the usage trends and importance of each business system includes:

[0035] The importance of a business system is determined based on its business value, customer needs, system reliability, technical feasibility, and security.

[0036] Based on the usage trends and importance of each business system, the business systems are sorted and filtered to obtain a list of business systems to be supported.

[0037] Based on the aforementioned list of business systems to be supported, corresponding operation and maintenance support recommendations are provided for each type of business system.

[0038] The present invention also provides a business system protection device, comprising:

[0039] The data acquisition module is used to acquire users' historical behavior data;

[0040] The data processing module is used to divide the user's historical behavior data into occasional behavior data and frequent behavior data, and to establish a relationship database between user behavior and business system for the frequent behavior data.

[0041] The model training module is used to determine the Arima parameters based on the frequent behavior data and train the pre-built user behavior prediction model using an integrated moving average autoregressive model framework.

[0042] The model prediction module is used to predict future user behavior using a trained user behavior prediction model, and to obtain the usage trends of each business system based on the prediction results and the relational database between the user behavior and the business system.

[0043] The business assurance module is used to generate a list of business systems to be assured and operation and maintenance assurance recommendations based on the usage trends and importance of each business system.

[0044] The present invention also provides a business system support device, comprising:

[0045] Memory, used to store computer programs;

[0046] A processor is used to implement the steps of the above-described business system assurance method when executing the computer program.

[0047] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described business system protection method.

[0048] The technical solution of the present invention has the following advantages compared with the prior art:

[0049] The business system support method described in this invention filters frequently occurring behavior data from historical user behavior data, effectively avoiding the impact of random data on the accuracy of model training; it extracts key fields from frequently occurring behavior data and performs differencing on the training set, effectively simplifying data complexity and ensuring the integrity and stability of the training set data, while improving the accuracy of prediction results; it uses an integrated moving average autoregressive model framework to train the user behavior prediction model, which has a simple framework and only requires endogenous variables without relying on other exogenous variables, effectively reducing technical difficulty; this invention can output an operation and maintenance support list based on the importance of the business system, which can remind operation and maintenance personnel to formulate operation and maintenance plans as needed and allocate manpower reasonably, and also provides different operation and maintenance support suggestions for different business systems, effectively assisting operation and maintenance personnel in efficient operation and maintenance and improving the user experience of the business system. Attached Figure Description

[0050] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein:

[0051] Figure 1 This is a flowchart illustrating the implementation of a business system protection method provided by the present invention;

[0052] Figure 2 This is a diagram illustrating the principles for optimizing parameters in a user behavior prediction model. Detailed Implementation

[0053] The core of this invention is to provide a method, apparatus, device, and computer storage medium that effectively improves prediction accuracy, reduces technical complexity, and combines a technical solution with actual operation and maintenance scenarios to provide effective assurance suggestions for prediction results.

[0054] To enable those skilled in the art to better understand the present invention, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] Please refer to Figure 1 , Figure 1 The flowchart illustrates the implementation of a business system assurance method provided by this invention; the specific operation steps are as follows:

[0056] S101: Obtain user's historical behavior data;

[0057] S102: Divide the user's historical behavior data into occasional behavior data and frequent behavior data, and establish a relationship database between user behavior and business system for the frequent behavior data;

[0058] S103: Determine the Arima parameters based on the frequently occurring behavior data, and train the pre-built user behavior prediction model using the integrated moving average autoregressive model framework;

[0059] S104: Predict future user behavior using a trained user behavior prediction model, and obtain the usage trends of each business system based on the prediction results and the database of the relationship between the user behavior and the business system.

[0060] S105: Based on the usage trends and importance ranking of each business system, generate a business system support list and operation and maintenance support recommendations.

[0061] Based on the above embodiments, this embodiment will provide a detailed description of step S101:

[0062] This invention obtains historical user behavior data from various business systems through methods such as web crawling or log querying. It analyzes user behavior, the business systems involved, and the main time periods occupied by these systems. This clarifies what actions users performed in the past, which business systems were involved, and the approximate duration of each system's usage. For example, a user making a phone call involves multiple business systems such as wireless networks, wired transmission, core networks, CRM systems, and BOSS billing systems; a user accessing websites to obtain information or shopping online involves business systems such as network protocols, domain name systems, servers, operating systems, and browsers; a user downloading files or applications from the internet involves business systems such as download protocols, network protocols, servers, operating systems, and applications. In the telecommunications industry, different domains have different business systems. For example, the B domain (business domain) has unified payment gateway systems and self-built air-charging systems; the M domain (management domain) has OA office systems such as unified information platforms and E-Palm Communication; the M domain ERP systems have procurement supply chain systems and human resource management platforms; and the domain-specific big data systems have event centers and PaaS platform systems.

[0063] Based on the above embodiments, this embodiment will provide a detailed description of step S102:

[0064] To more accurately predict future user behavior, this invention filters historical user behavior data based on preset classification criteria, distinguishing between incidental and recurring behavior data. Only recurring behavior data is used as input data for training the user behavior prediction model, avoiding bias from incidental behavior data and improving the model's robustness. Specifically, this application proposes, based on historical user behavior data, first selecting a time period as a time window, such as a week or a month, and grouping the historical user behavior data chronologically, with each group spanning the selected time period. Then, the frequency of historical user behavior is calculated; that is, for each time window, the frequency of a specific user behavior within that time window is calculated (e.g., the number of searches per day, the number of purchases per week, etc.). The frequency of the specific user behavior within that time window is then compared with a preset threshold (which can be set according to the maintenance team's daily work experience). If the frequency exceeds or equals the threshold, the behavior is considered a recurring behavior; if it is below the threshold, it is considered an incidental behavior (e.g., more than ten searches per day constitute a recurring behavior; otherwise, it is an incidental behavior).

[0065] In addition, the present invention can also use the K-MEANS clustering algorithm to cluster frequently occurring behavior data, while sparse data that deviates from the population is considered as random behavior data.

[0066] After obtaining frequently occurring behavioral data, this invention extracts key fields such as the occurrence time of the frequently occurring behavior, user behavior, number of users, business systems involved in the user behavior, and the main time periods occupied by the involved business systems. Then, it establishes a relationship representation between user behavior and business systems through topology graphs, knowledge graphs, or association diagrams, as shown in the table below. If the data shows that the same user behavior involves different business systems, the principle of majority rule is adopted, and the business systems involved in the majority of the user behavior are taken as the final result.

[0067] Table 1. Relationship between User Behavior and Business Systems

[0068]

[0069] After obtaining the relationship representation between user behavior and business systems, a unique identifier is configured for each relationship chain, and a relational database is established. For example, the relationship chain between user behavior A and its involved business systems and the main time periods occupied by the business systems is configured with the identifier mark(A); the relationship chain between user behavior B and its involved business systems is configured with the identifier mark(B); and the relationship chain between user behavior X and its involved business systems is configured with the identifier mark(X). When you want to search for the business systems involved in a user behavior later, you can search for the identifier.

[0070] Based on the above embodiments, this embodiment will provide a detailed description of step S103:

[0071] Before training the user behavior prediction model, this invention first performs data cleaning and exploration. The acquired data, data = {behavior occurrence time, user behavior, number of users, involved business systems, main time period occupied by the business systems}, is indexed by time and sorted to obtain the training set, denoted as: TrainSet = {value(time), value(user behavior), value(number of users), value(business system), value(main time period occupied by the business system), value(date_index)}.

[0072] The goal is to predict future user behavior based on historical user behavior, such as how many users will perform what actions at what times in the future, in order to obtain access information for each business system, including usage time and number of users. Therefore, user behavior should be grouped, i.e.

[0073] TrainSetA = {value(time), value(user behavior A), value(number of users), value(business system), value(main time period occupied by business system), value(date_index)},

[0074] TrainSetB = {value(time), value(user behavior B), value(number of users), value(business system), value(main time period occupied by business system), value(date_index)},

[0075] TrainSetC = {value(time), value(user behavior C), value(number of users), value(business system), value(main time period occupied by business system), value(date_index)},

[0076] TrainSetD = {value(time), value(user behavior N), value(number of users), value(business system), value(main time period occupied by business system), value(date_index)},

[0077] TrainSet={TrainSetA, TrainSetB, TrainSetC,..., TrainSetN}.

[0078] This invention integrates a moving average autoregressive model framework to train a user behavior prediction model. First, it analyzes the stationarity of the training set (TrainSet) and determines the Arima parameter d (difference order). Then, it performs differencing on the TrainingSet to improve the data's stationarity and interval stability. Specifically, this application proposes to depict a data distribution map based on the time of behavior occurrence, user behavior, and the number of users. When it is found that the data for a certain user behavior (e.g., browsing a website) is not stable and has large fluctuations (e.g., more than 1000 people browsed the website in the previous time window, but only 10 people browsed the same website in the next time window), the data is differencing to make the data tend to be stable. The difference is denoted as C, and the original dataset (TrainSet) is Y = {y1, y2, y3, y4, ... y n If the differenced dataset is F, then...

[0079]

[0080] Where C=0, no differencing is performed; C=1, first-order differencing is performed, but data exploration revealed that first-order differencing is unsuitable for the data distribution characteristics of this application; C=2, second-order differencing is performed, which is more consistent with this application, showing good data stationarity and interval characteristics, and the data tends to have certain regularity after second-order differencing. Therefore, the differencing order d of the Arima parameter is 2, and this invention performs second-order differencing on the training set TrainSet.

[0081] Then, the parameters of the user behavior prediction model are tuned. The model expression used in this invention is: Where Φ represents the AR (autoregressive) coefficient, θ represents the MA (moving average) coefficient, p represents the autoregressive coefficient of the TrainSet, q represents the moving average coefficient of the TrainSet, μ represents a constant, and φ represents the white noise sequence. Specifically, this invention determines the Arima parameters p and q based on the AIC, BIC, and HQIC information criteria. By inputting a stationary training set TrainSet into the statistical library statsmodels, the parameters max_ar and max_ma are adjusted to minimize the output AIC, BIC, and HQIC values. The max_ar and max_ma values ​​at this point are the autoregressive coefficient p and the moving average coefficient q of the TrainSet. The solution process is as follows: Figure 2 As shown.

[0082] After performing differential processing on TrainSet and confirming the Arima parameters p (autoregressive coefficient) and q (moving average coefficient), TrainSet is put into the integrated moving average autoregressive model framework for time series algorithm training, i.e., model = arima(TrainSet, p, d, q), which is used to predict user behavior and the corresponding number of users at a certain time in the future.

[0083] Based on the above embodiments, this embodiment will provide a detailed description of step S104:

[0084] This invention utilizes a trained user behavior prediction model to predict user behavior and the corresponding number of users for the next week, month, or quarter. Based on the prediction results, and combined with a relational database, the invention searches for the business systems involved in the user behavior and the main time periods occupied by these systems, thereby outputting the future access volume and peak time for each business system. For example, a user behavior prediction model predicts that 568 users will call to inquire about services in the coming week, and 332 users will recharge their accounts. Based on the identifier of the call inquiry service relationship chain, it can be determined that the business systems involved in the call inquiry service are business systems A, B, and C, and the main time periods occupied by these business systems are 10:00-12:00 and 14:30-17:30. Based on the identifier of the recharge service relationship chain, it can be determined that the business systems involved in the recharge service are business systems A and C, and the main time periods occupied by these business systems are 9:30-11:30 and 14:30-17:30. Therefore, the total number of visits to business systems A and C in the coming week will be 568 + 332, with peak visits from 9:30-12:00 and 14:30-17:30. The total number of visits to business system B will be 568, with peak visits from 10:00-12:00 and 14:30-17:30, and so on.

[0085] Based on the above embodiments, this embodiment will provide a detailed description of step S105:

[0086] This invention combines future usage trends of various business systems with their importance ranking to generate a business system toplist and provides operational support recommendations. Specifically, the importance ranking criteria for business systems can be set according to different dimensions, such as business value, customer needs, system reliability, technical feasibility, and security. Regarding business value, it can be measured by key indicators such as revenue, market share, and profit. Regarding customer needs, it can be based on indicators such as customer frequency of use, usage patterns, and customer satisfaction with different business systems. Regarding system reliability, it can be achieved by evaluating the reliability, fault tolerance, and disaster recovery mechanisms of the business systems. Regarding technical feasibility, it can be achieved by evaluating the technical difficulty, resource requirements, and technical support of the business systems. Regarding security, it can be achieved by evaluating the system's security, protection capabilities, and data privacy protection. In this application proposal, the external business systems in Domain B and the Economic Sub-domain are classified as Level 1 in importance. The remaining business systems in Domain B and the Economic Sub-domain, as well as the business systems in Domain M, are categorized into Level A, B, and C based on historical login activity and leadership usage levels. Therefore, Level A business systems are classified as Level 2, Level B as Level 3, and Level C as Level 4. Combining future access volume, peak access times, and the importance of each business system, with access volume weighted at 0.55 and the importance weighted at 0.45, a list of the top ten business systems requiring maintenance and support is obtained.

[0087] Based on the business system support list, this application proposal can also provide different operation and maintenance support suggestions for different types of business systems. For example, for network operation systems: real-time monitoring of network performance and availability, including signal strength, coverage, and data transmission rate; configuration of automated alarm systems to promptly identify and resolve potential problems; implementation of load balancing and fault tolerance mechanisms to ensure high availability and reliability of services; regular backup of important data and system configurations to prevent data loss; and the use of appropriate firewalls, antivirus software, and encryption protocols to protect user data security. For communication switching systems: real-time monitoring of switch status and performance indicators, such as call success rate, connection latency, and signal quality; configuration of automated alarm systems to promptly identify and resolve potential problems; implementation of fault tolerance mechanisms to ensure uninterrupted operation of the switching system; regular backup of important data and system configurations to prevent data loss; and the use of appropriate firewalls, antivirus software, and encryption protocols to protect user data security. For business support systems: real-time monitoring of business system performance and availability, including data processing speed, service response time, and error rate; and configuration of automated alarms. The system promptly identifies and resolves potential problems, implements fault tolerance mechanisms to ensure uninterrupted operation of business support systems, regularly backs up important data and system configurations to prevent data loss, and employs appropriate firewalls, antivirus software, and encryption protocols to protect user data security. For business management systems: Real-time monitoring of the performance and availability of business management systems, including business data analysis and report generation, is implemented. An automated alarm system is configured to promptly identify and resolve potential problems, and important data and system configurations are regularly backed up to prevent data loss. Appropriate firewalls, antivirus software, and encryption protocols are employed to protect user data security. For data center equipment management systems: Real-time monitoring of the data center environment and equipment status, including temperature, humidity, and power supply, is implemented. An automated alarm system is configured to promptly identify and resolve potential problems, implements fault tolerance mechanisms to ensure uninterrupted operation of the data center equipment management system, and important data and system configurations are regularly backed up. Appropriate firewalls, antivirus software, and encryption protocols are employed to protect user data security.

[0088] The present invention provides operational support recommendations in the form of an encrypted file containing the business system support list and corresponding operational support recommendations, which is then uploaded to a public cloud for authorized operations and maintenance personnel to decrypt and view (e.g., through identity verification). This helps operations and maintenance personnel to plan ahead and ensures the security of the output data. This application also includes a message reminder function. Before a business system experiences a peak in access, a scheduled operational support reminder will pop up, including prompts for optimizing system performance, strengthening monitoring, regular backup and recovery, enhancing security, and clarifying team collaboration. This alerts the operations and maintenance team to prepare support plans in advance, improving the scientific decision-making and foresight of business system support.

[0089] This invention filters frequently occurring behavior data from users' historical behavior data, effectively avoiding the impact of random data on the accuracy of model training; it extracts key fields from frequently occurring behavior data and performs second-order differencing on the training set, effectively simplifying data complexity and ensuring the integrity and stability of the training set data, while improving the accuracy of prediction results; it uses an integrated moving average autoregressive model framework to train the user behavior prediction model, which has a simple framework that only requires endogenous variables without relying on other exogenous variables, effectively reducing technical difficulty.

[0090] This invention can generate an operation and maintenance support list based on the importance of the business system, which can remind operation and maintenance personnel to formulate operation and maintenance plans as needed and allocate manpower reasonably. Moreover, this application proposal also provides different operation and maintenance support suggestions for different business systems, effectively assisting operation and maintenance personnel to carry out operation and maintenance efficiently and improving the user experience of business systems.

[0091] This invention constructs a technical solution based on actual operation and maintenance scenarios. For example, it lists the behaviors that users may perform in the communications industry, the business systems involved in a user's behavior, the order of importance of the business systems, etc. This application proposal is specific to the operation and maintenance assurance of business systems in the communications industry.

[0092] This invention employs a differential integrated moving average autoregressive model framework to train a user behavior prediction model. By combining the prediction results with the relationship between user behavior and business systems, it derives the future usage trends of each business system. Based on the importance of the business systems requiring support, it outputs a support list for each system and provides operational support recommendations. This provides forward-looking support guidance for operations and maintenance teams, improving work efficiency and reducing technical costs. This invention can be developed into a standardized product, providing intelligent services for the operation and maintenance of industry systems, and has broad market prospects.

[0093] This invention is applicable to scenarios requiring the maintenance of multiple systems, such as the centralized maintenance service project for multi-domain systems of China Mobile Guangdong Company, which requires unified maintenance of a large number of IT systems such as B domain and M domain. This invention can provide the operation and maintenance team with forward-looking support guidance, improve work efficiency and reduce technical costs, thereby increasing efficiency and reducing costs for the company.

[0094] This invention also provides a business system protection device; the specific device may include:

[0095] The data acquisition module is used to acquire users' historical behavior data;

[0096] The data processing module is used to divide the user's historical behavior data into occasional behavior data and frequent behavior data, and to establish a relationship database between user behavior and business system for the frequent behavior data.

[0097] The model training module is used to determine the Arima parameters based on the frequent behavior data and train the pre-built user behavior prediction model using an integrated moving average autoregressive model framework.

[0098] The model prediction module is used to predict future user behavior using a trained user behavior prediction model, and to obtain the usage trends of each business system based on the prediction results and the relational database between the user behavior and the business system.

[0099] The business assurance module is used to generate a list of business systems to be assured and operation and maintenance assurance recommendations based on the usage trends and importance of each business system.

[0100] The business system protection device in this embodiment is used to implement the aforementioned business system protection method. Therefore, the specific implementation of the business system protection device can be found in the embodiment section of the business system protection method above. For example, the data acquisition module, data processing module, model training module, model prediction module, and business protection module are used to implement steps S101, S102, S103, S104, and S105 in the above-mentioned business system protection method, respectively. Therefore, its specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0101] A specific embodiment of the present invention also provides a business system protection device, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the aforementioned business system protection method.

[0102] A specific embodiment of the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described business system protection method.

[0103] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0104] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations 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, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0105] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0106] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0107] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for ensuring the operation of a business system, characterized in that, include: Obtain user historical behavior data; The user's historical behavior data is divided into occasional behavior data and frequent behavior data, and a relationship database between user behavior and business system is established for the frequent behavior data. Arima parameters are determined based on the frequently occurring behavior data, and a pre-built user behavior prediction model is trained using an integrated moving average autoregressive model framework. The trained user behavior prediction model is used to predict future user behavior, and the usage trends of each business system are obtained based on the prediction results and the relationship database between the user behavior and the business system. Based on the usage trends and importance ranking of each business system, a list of business systems to be supported and operational support recommendations are generated.

2. The business system assurance method according to claim 1, characterized in that, The acquisition of user historical behavior data includes: By crawling or querying logs, user behavior data, the business systems involved in user behavior, and the main time periods occupied by the business systems involved in user behavior are obtained from each business system, thus obtaining historical user behavior data.

3. The business system assurance method according to claim 1, characterized in that, The division of the user's historical behavior data into incidental behavior data and frequent behavior data includes: The time window size is determined based on the user's historical behavior data; Based on the size of the time window, the user's historical behavior data is grouped in chronological order; Calculate the frequency of a specific user behavior in each time window and compare the frequency with a preset threshold; When the frequency is not less than the preset threshold, the user-specific behavior is considered a recurring behavior.

4. The business system assurance method according to claim 1, characterized in that, The process of establishing a database linking user behavior to the business system based on the frequently occurring behavior data includes: Based on the frequently occurring behavior data, key fields such as the occurrence time of the frequently occurring behavior, user behavior, number of users, business systems involved in the user behavior, and the main time periods occupied by the business systems are extracted, and the relationship between user behavior and business systems is established through topology graphs, knowledge graphs, or association charts. Each relationship chain representing the relationship between user behavior and the business system is configured with a unique identifier, and a relationship database is established.

5. The business system assurance method according to claim 1, characterized in that, The step of determining the Arima parameter based on the frequent behavior data and training the pre-built user behavior prediction model using an integrated moving average autoregressive model framework includes: The frequently occurring behavior data is cleaned, and a time index is set to obtain the training set by sorting it by time. The training set is differentially processed, and the number of differences for the Arima parameter is determined based on the stationarity of the data. The autoregressive coefficients and moving average coefficients of the Arima parameters are determined based on the AIC, BIC, and HQIC information criteria. Based on the determined Arima parameters, after differential processing of the training set, the integrated moving average autoregressive model framework is input for time series algorithm training to predict user behavior and the corresponding number of users at a future time.

6. The business system assurance method according to claim 1, characterized in that, The process involves using a trained user behavior prediction model to predict future user behavior, and based on the prediction results and a database of relationships between user behavior and business systems, obtaining usage trends for each business system, including: Predict future user behavior and the corresponding number of users using a trained user behavior prediction model; Based on the prediction results, and according to the database of the relationship between user behavior and business systems, the business systems involved in the user behavior and the main time periods occupied by the business systems are obtained. Based on the business systems involved in the user behavior and the main time periods occupied by the business systems, the future access volume and peak time of each business system can be obtained.

7. The business system assurance method according to claim 1, characterized in that, The process of generating a business system support list and operation and maintenance support recommendations based on the usage trends and importance of each business system includes: The importance of a business system is determined based on its business value, customer needs, system reliability, technical feasibility, and security. Based on the usage trends and importance of each business system, the business systems are sorted and filtered to obtain a list of business systems to be supported. Based on the aforementioned list of business systems to be supported, corresponding operation and maintenance support recommendations are provided for each type of business system.

8. A business system protection device, characterized in that, include: The data acquisition module is used to acquire users' historical behavior data; The data processing module is used to divide the user's historical behavior data into occasional behavior data and frequent behavior data, and to establish a relationship database between user behavior and business system for the frequent behavior data. The model training module is used to determine the Arima parameters based on the frequent behavior data and train the pre-built user behavior prediction model using an integrated moving average autoregressive model framework. The model prediction module is used to predict future user behavior using a trained user behavior prediction model, and to obtain the usage trends of each business system based on the prediction results and the relational database between the user behavior and the business system. The business assurance module is used to generate a list of business systems to be assured and operation and maintenance assurance recommendations based on the usage trends and importance of each business system.

9. A business system support device, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of a business system assurance method as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the business system assurance method as described in any one of claims 1 to 7.

Images