Risk assessment method and device, storage medium, electronic equipment and product

Abstract

The invention provides a risk assessment method and device, a storage medium, electronic equipment and a product. The method comprises the following steps: acquiring an initial change scheme for a target service and software design information corresponding to the target service; wherein the initial change scheme comprises the change content and the implementation process of the service change to be implemented to the target service; according to the initial change scheme, change reference information matched with the initial change scheme is determined in a pre-constructed knowledge base; generating a cue word based on the change reference information, the software design information and the initial change scheme, and inputting the cue word into a preset scheme evaluation model, so that the scheme evaluation model evaluates the initial change scheme to obtain an evaluation result; wherein the evaluation result is used for representing whether the initial change scheme meets a target change demand reflected by the change reference information or not under the constraint of the software design information; and according to the evaluation result, determining the risk of implementing business change on the target business based on the initial change scheme.

Description

TECHNICAL FIELD

[0001] One or more embodiments of the present specification relate to the technical field of computer technology, and more particularly to a risk assessment method and device, a storage medium, an electronic device and a product. BACKGROUND

[0002] In a fast iteration software development environment, business changes for businesses such as e-commerce transactions, financial payments, cloud computing services, etc. have become the norm, and their change content covers multiple aspects such as new function online, defect repair, configuration update, etc. It is an important technical means for continuous optimization and iterative upgrade of business application.

[0003] Before implementing changes to the target business, a change plan is usually formulated according to the business target, technical architecture and risk prediction, and the feasibility, security and potential impact of the change plan are comprehensively evaluated to reduce various risks that may occur during the change process.

[0004] However, the traditional change plan evaluation method mainly relies on human experience, which is not only inefficient, but also limited by the knowledge reserve and focus of the evaluators, resulting in low accuracy and reliability of the evaluation results. It cannot accurately detect the risks that may exist in the implementation process of the business change plan, and it is difficult to provide reliable protection for business changes. SUMMARY

[0005] Therefore, one or more embodiments of the present specification provide technical solutions as follows: According to a first aspect of one or more embodiments of the present specification, a risk assessment method is provided, comprising: obtaining an initial change plan for a target business and software design information corresponding to the target business; wherein the initial change plan includes change content and implementation process of the business change to be implemented to the target business; determining change reference information matched with the initial change plan in a pre-constructed knowledge base; filling the change reference information, the software design information and the initial change plan into a preset prompt word template to generate a prompt word based on the change reference information, the software design information, the initial change plan and the prompt text in the prompt word template, and inputting the prompt word into a scheme evaluation model based on an attention mechanism, so that the scheme evaluation model determines whether the initial change plan meets the target change requirement reflected by the change reference information under the constraint of the software design information under the instruction of the prompt text, to evaluate the initial change plan and obtain an evaluation result; wherein the target change requirement includes at least one of change release requirement, change monitoring requirement and change recovery requirement. According to the evaluation result, a change risk of implementing the business change on the target business based on the initial change scheme is determined, wherein the change risk includes an abnormality of the target business caused by implementing the initial change scheme without meeting the target change requirement.

[0006] According to a second aspect of one or more embodiments of the present specification, a risk evaluation device is provided, comprising: An acquisition module is configured to acquire an initial change scheme for a target business and software design information corresponding to the target business, wherein the initial change scheme includes change content and an implementation process of a business change to be implemented on the target business. A matching module is configured to determine change reference information matched with the initial change scheme in a pre-constructed knowledge base according to the initial change scheme. An evaluation module is configured to fill the change reference information, the software design information, and the initial change scheme into a preset prompt word template to generate a prompt word based on the change reference information, the software design information, the initial change scheme, and prompt text in the prompt word template, and input the prompt word into a scheme evaluation model based on an attention mechanism, so that the scheme evaluation model determines whether the initial change scheme meets a target change requirement reflected by the change reference information under the constraint of the software design information under the instruction of the prompt text, to evaluate the initial change scheme and obtain an evaluation result, wherein the target change requirement includes at least one of a change release requirement, a change monitoring requirement, and a change recovery requirement. A determination module is configured to determine a change risk of implementing the business change on the target business based on the initial change scheme according to the evaluation result, wherein the change risk includes an abnormality of the target business caused by implementing the initial change scheme without meeting the target change requirement.

[0007] According to a third aspect of one or more embodiments of the present specification, an electronic device is provided, comprising a processor, a memory for storing processor-executable instructions, wherein the processor implements the steps of the above method by running the executable instructions.

[0008] According to a fourth aspect of one or more embodiments of the present specification, a computer-readable storage medium is provided, which stores computer instructions, and the instructions are executed by a processor to implement the steps of the above method.

[0009] According to a fifth aspect of one or more embodiments of the present specification, a computer program product is provided, comprising computer program / instructions, which are executed by a processor to implement the steps of the above method.

[0010] From the above examples, the present specification generates a prompt word by changing the reference information, software design information and initial change scheme, and inputs the whole flow automation processing of the evaluation model, effectively solving the pain points of traditional manual evaluation: the present scheme takes the change demand reflected by the relevant change reference information in the knowledge base as a reference, takes the software design information as the context constraint of the target business, relies on the evaluation model to evaluate whether the initial change scheme meets the target change demand, the whole process does not need to rely on the subjective judgment of the evaluation personnel, greatly improves the evaluation efficiency, and significantly improves the accuracy and reliability of the evaluation result. BRIEF DESCRIPTION OF DRAWINGS

[0011] Figure 1 FIG. 1 is a schematic diagram of the architecture of a risk assessment service system according to an example embodiment; Figure 2 FIG. 2 is a schematic diagram of the flow of a risk assessment method according to an example embodiment; Figure 3 FIG. 3 is a whole flowchart of a risk assessment according to an example embodiment; Figure 4 FIG. 4 is a schematic diagram of the interaction process of a business change system according to an example embodiment; Figure 5 FIG. 5 is a schematic diagram of the structure of a device according to an example embodiment; Figure 6 FIG. 6 is a block diagram of a risk assessment apparatus according to an example embodiment. DETAILED DESCRIPTION

[0012] In order for those skilled in the art to better understand the technical solutions in the present specification, the technical solutions in the present specification will be described clearly and completely below in conjunction with the drawings in the embodiments of the present specification. Obviously, the described embodiments are only part of the embodiments of the present specification, not all embodiments. Based on the embodiments in the present specification, all other embodiments obtained by those skilled in the art without creative labor should be within the scope of protection of the present specification.

[0013] In business scenarios such as financial technology, e-commerce platforms, enterprise-level SaaS systems, etc., publishing business changes (such as new function online, defect repair, configuration update, interface iteration, data structure adjustment, permission system upgrade, etc.) is the norm to support business iteration, but also contains great risks, such as service interruption, data error, user experience degradation, transaction process jam, compliance verification failure, cross-system interaction anomaly, etc.

[0014] Traditional change review relies on human experience, which has significant shortcomings: low efficiency, narrow coverage, easy to miss, and difficult to effectively use historical experience and complex system knowledge. Moreover, there is currently a lack of solutions in the industry that can systematically and automatically assess and optimize the core change requirements (such as change resilience) of change schemes, making it difficult to quickly identify potential risks in changes and to provide actionable optimization recommendations based on historical change data.

[0015] Based on this, the present specification provides a business change scheme that uses the natural language understanding and reasoning capabilities of a large language model (LLM), combined with change reference information retrieved based on semantics in a knowledge base, to automatically evaluate whether the change scheme meets the target change requirements, identify deficiencies and provide specific improvement suggestions, thereby improving the safety, reliability and efficiency of releasing changes and reducing operational risks.

[0016] Figure 1 Figure 1 is a schematic diagram of the architecture of a risk assessment service system according to an example embodiment. As shown in Figure 1, the system can include a server 11, a network 12, and several electronic devices such as a personal computer (PC) 13, a mobile phone 14, etc. Figure 1

[0017] The server 11 can be a physical server containing a standalone host, or it can be a virtual server carried by a host cluster. During operation, the server 11 can run the server-side program of an application to implement the related functions of the application, such as when the server 11 runs the program of the business change service, it can implement the corresponding business change service platform.

[0018] ​PC 13, mobile phone 14 are only part of the types of electronic devices that users can use. In fact, users can obviously also use electronic devices of the following types: tablet devices, notebook computers, personal digital assistants (PDAs), wearable devices (such as smart glasses, smart watches, etc.), and the like, and one or more embodiments of the present specification do not limit this. During operation, the electronic device can run a client-side program of an application to implement the related functions of the application, such as when the electronic device runs the program of a business change service, it can implement a client of the business change service. The application program of the client of the business change service described above can be started and run on the electronic device. The client-side program can be a native application installed on the electronic device, or the client-side program can be an applet, a fast application, or other similar forms. Of course, when using web technologies such as HTML5 or the like, the related functions can be implemented through a page displayed by a browser, and the browser can be a standalone browser application or a browser module embedded in some application.

[0019] As for the network 12 between the electronic devices such as PC 13, mobile phone 14 and the server 11, it can be based on the communication mode supported by the corresponding electronic device, and the wired or wireless network is selected to realize communication, and the present specification does not limit this. For example, PC 13 can support wired and wireless communication at the same time, so wired or wireless network can be used to realize communication as needed, and mobile phone 14 usually only supports wireless communication, so wireless network can be used to realize communication.

[0020] Based on the above service system architecture, the present specification provides a risk assessment method, as shown in Figure 2

[0021] Figure 2 FIG. 1 is a flow diagram of a risk assessment method provided by an exemplary embodiment, including the following steps: S201: Obtain an initial change scheme for a target business and software design information corresponding to the target business; wherein the initial change scheme includes the change content and implementation process of the business change to be implemented to the target business.

[0022] In the present specification, the execution subject for executing the risk assessment method can be a designated device such as a server, of course, it can also be a cloud server cluster, an edge computing node, an embedded terminal device, a special hardware with data processing capability (such as an encryption server in the financial industry, a transaction processing terminal in the e-commerce field) and other devices. In order to facilitate description, the following will take the server as an example to describe the business change method provided by the present specification. ​

[0023] The business change can be a change issued for an online business (such as a commodity transaction business of an e-commerce platform, an online credit business of a financial institution, a collaborative office business, an online consultation business), an offline business (such as a cash register business of a retail store, a warehouse management business of a logistics site, a counter service business of an offline bank, a store operation business of a chain restaurant), or the like. The changes can include: online of new modules, optimization of existing functions, and defect repair; configuration level business rule adjustment, parameter threshold optimization, and permission configuration update; data and interface level data structure adjustment, interface iteration upgrade, and data migration synchronization; hardware and environment level server resource expansion, offline device firmware update, and system running environment optimization.

[0024] Of course, in actual application, other business changes corresponding to other businesses can also be included, which are not limited in the present specification.

[0025] Before the business change is issued, a user (such as an operation and maintenance personnel) usually first formulates an initial change scheme, which is used to describe the change content and implementation process of the business change to be implemented to the target business.

[0026] For example, for the online business of the e-commerce platform-function level-new module online, the change content of the initial change scheme is: online of the function modules of “commodity display”, “real-time interactive ordering”, and “commission settlement”; the implementation process is: deploy the modules in the test environment 3 days in advance and complete the function and stress test, deploy the production environment at 00:00-02:00 on the online day, perform 1 hour of gray verification after deployment, and if there is no exception, fully open, and prepare a rollback plan at the same time.

[0027] For example, for the offline business of the retail store-hardware and environment level-firmware update of the cash register device, the initial change scheme will explicitly specify that the change content is to update the firmware version that fixes the payment security vulnerability, and the implementation process is: select 2 stores as pilot stores, complete the device update of all stores before the next day of operation if there is no problem in the pilot, test 3 simulated transactions for each device after updating, ensure the payment function is normal, and at the same time, prepare the old firmware installation package for rollback.

[0028] In the present specification, the timing at which the server acquires the initial change scheme can be various, for example, the server can acquire the initial change scheme after the user uploads the initial change scheme to the business change system. For another example, when the user completes the initial change scheme online in the business change system and clicks the "submit for review" button, the system automatically synchronizes the scheme data to the server, and the server acquires the scheme in real time. Of course, the user can also transmit the initial change scheme file to the designated server through email, file transfer protocol (FTP) and the like, and the server acquires the initial change scheme after monitoring the completion of file transmission and verifying the format compliance.

[0029] In addition, in order to ensure that accurate software context information is provided to the evaluation model during the process of business change, the server can also acquire software design information of the target business, which can be directly extracted from the system analysis documentation of the target business.

[0030] The system analysis documentation refers to the detailed technical description of the design documentation, interface documentation, data dictionary, dependency graph, etc. of the specific system involved in the change to be released. The software design information extracted therefrom can include: system architecture description of the target business (such as microservice module division logic, module interaction timing diagram of core business process, database hierarchical architecture, etc.), dependency list (such as software running dependent third-party component version, interface list of cross-system interaction, hardware resource dependency, data dependency, etc.), etc. Of course, other software design information related to the software design of the target business can also be included in actual application, which will not be enumerated one by one in the present specification.

[0031] S202: According to the initial change scheme, determine the change reference information matched therewith in the pre-constructed knowledge base.

[0032] After acquiring the initial change scheme, the server can determine the change reference information matched therewith in the pre-constructed knowledge base. In the present specification, a variety of change reference information is pre-stored in the knowledge base, which can include historical change schemes for reflecting different change requirements, failed change schemes that cause change failure in the process of trying business change in the past, and preferred change schemes that are verified to have excellent effect, controllable risk and up-to-standard efficiency. In addition, the change reference information can also include some typical architecture documents, such as software architecture diagram, microservice dependency document, payment module interaction flowchart, system API access specification, distributed database sharding design document, server cluster expansion architecture record, etc., thereby providing technical basis for change evaluation.

[0033] In practical applications, the change reference information is usually stored in the form of a vector in the material information library. In the process of retrieving the change reference information, the server can first perform semantic analysis on the initial change scheme to extract structured change data and convert it into a corresponding semantic vector (e.g., by using a natural language processing model or other semantic analysis engine to extract feature representations and generate a fixed-dimension vector). Then, the similarity (e.g., cosine similarity, Euclidean distance, etc.) between the feature representation of the change data and the feature representation of each change reference information in the knowledge base is calculated, and one or more change reference information with a similarity higher than a preset threshold is filtered out, or the change reference information is sorted according to the similarity, and one or more change reference information before a preset position is filtered out.

[0034] S203: filling the change reference information, the software design information, and the initial change scheme into a preset prompt word template to generate a prompt word based on the change reference information, the software design information, the initial change scheme, and the prompt text in the prompt word template, and inputting the prompt word into a scheme evaluation model based on an attention mechanism, so that the scheme evaluation model determines whether the initial change scheme meets the target change requirement reflected by the change reference information under the constraint of the software design information under the instruction of the prompt text, to evaluate the initial change scheme and obtain an evaluation result; wherein the target change requirement includes at least one of a change release requirement, a change monitoring requirement, and a change recovery requirement; S204: determining a change risk of implementing business change based on the initial change scheme for the target business according to the evaluation result, wherein the change risk includes an abnormality of the target business caused by implementing the initial change scheme without meeting the target change requirement.

[0035] The server can generate a prompt word based on the change reference information, the software design information, and the initial change scheme to instruct the scheme evaluation model to evaluate the initial change scheme through the prompt word.

[0036] In practical applications, the server can fill the change reference information, the software design information, and the initial change scheme into a preset prompt word template, which sets corresponding prompt text for prompting the scheme evaluation model to refer to which information reflecting the target change requirement and which information as a constraint, so as to generate a prompt word based on the change reference information, the software design information, the initial change scheme, and the prompt text in the prompt word template. An example of the prompt word will be described in detail below, and the present specification will not be described in detail here.

[0037] The scheme evaluation model can be a large language model based on an attention mechanism. Precise semantic analysis is achieved through token-level attention distribution. The prompt words are disassembled into discrete tokens. The associated weights between different tokens are calculated, and the feature vectors are weighted and summed to make the determined feature representation accurately express the semantic information of the prompt words, assisting the model in identifying core change requirements in the change reference information, key constraint clauses in the software design information, implementation points in the initial change scheme, and generation instructions expressed by the prompt text.

[0038] During this process, the model first disassembles the input prompt words into discrete tokens (which can be understood as the smallest semantic unit of text, such as words, characters, or subwords). Then, the self-attention (Self-Attention) mechanism is used to calculate the correlation weights between different tokens. This step can mine the semantic correlation between words in the text. Then, the feature vectors corresponding to each token are weighted and summed according to these weights, so that the final feature representation accurately represents the semantic information of the prompt words.

[0039] In addition, the scheme evaluation model can also be a domain-specific fine-tuned model, a multi-modal evaluation model, and a rule and learning hybrid model, and other artificial intelligence models, wherein: The domain-specific fine-tuned model is fine-tuned based on general large language models and professional data in a specific field (such as software development). For example, in the software development field, after fine-tuning a large amount of software requirement documents, design specifications, and code change records, the model has a more accurate understanding of professional terms (such as "microservice architecture" and "API interface"), business logic (such as "user registration process" and "order payment link"), and change scenarios (such as "function addition in version iteration" and "online bug fixing"). When used for scheme evaluation, it can more efficiently identify the core elements in software change tasks, such as whether a software module refactoring scheme meets the design principles of the field and whether there are compatibility issues with existing systems, making up for the shortcomings of general models in domain details.

[0040] The multi-modal evaluation model can process various types of inputs such as text, images, and audio at the same time. In some complex scenario evaluation scenarios, it is often necessary to combine multiple types of information. For example, when evaluating a software adaptation scheme for an intelligent hardware product, there may be not only text-form functional requirement descriptions, but also image information such as hardware appearance diagrams and interface schematics, and even voice instructions for product operation processes. The multi-modal evaluation model can analyze and fuse these different types of information, identify the structural characteristics of the hardware from the images, extract the operation logic points from the voice, and then combine the text requirements to comprehensively evaluate whether the software adaptation scheme can meet the various requirements of the hardware, ensuring the rationality and feasibility of the scheme in multiple dimensions.

[0041] The rule and learning hybrid model combines artificially formulated rules and the adaptive ability of machine learning. On the one hand, artificial rules can be formulated according to industry standards, regulatory requirements, etc., such as rules related to data encryption, transaction security verification, etc. in financial software scheme evaluation. When the scheme involves these aspects, the model will first make a preliminary judgment based on the rules. On the other hand, through machine learning, the model can learn some hidden patterns and rules from a large amount of historical scheme evaluation data, which may be difficult to accurately cover through artificial rules, such as the potential preferences of users for functional priorities in different business scenarios. In actual evaluation, the model first uses rules for rigid constraint checking, and then uses learned patterns to evaluate the flexibility and innovation of the scheme, so as to achieve both compliance with basic norms and adaptation to complex and variable actual needs, making the evaluation results more comprehensive and accurate.

[0042] Before using the above scheme evaluation model, the server can first train or fine-tune it to enable the scheme evaluation model to accurately identify the semantic information of the keywords and obtain the evaluation results. The training or fine-tuning process of the above scheme evaluation model is as follows: A structured training data set is constructed, which includes: Sample input data: covering sample change schemes (such as version iteration plans of historical systems, functional module adjustment schemes, etc.), sample software design information (such as system architecture documents, interface specifications, performance constraint indicators, etc.), and sample change reference information (such as explicit change release window period, monitoring indicators that must be covered, and historical change schemes that require recovery timeliness).

[0043] Label data: the actual evaluation results of the sample change scheme by the developers or professional evaluation team are used as labels. The label form can be set according to the scene, such as binary judgment results, quantitative scores, comprehensive evaluation reports containing "release feasibility", "monitoring integrity", "recovery effectiveness", etc.

[0044] The model training process is as follows: The server constructs a training prompt word according to the same logic as the actual reasoning, fills the sample change scheme, sample software design information, sample change reference information into a preset prompt word template, and generates an input text containing evaluation instructions.

[0045] The training prompt word is input into the scheme evaluation model to be trained to obtain a predicted evaluation result output by the model. The process is the same as the actual reasoning process, and the specific content will be described in detail in the actual reasoning process below. This specification will not be described in detail here.

[0046] The loss value is calculated according to the label data (actual evaluation result) and the predicted evaluation result. For example, cross-entropy loss can be used for binary judgment, mean square error loss can be used for evaluation, and weighted mixed loss function can be used for multi-dimensional evaluation.

[0047] The optimization goal is to minimize the loss value, and the model parameters (such as attention weight matrix, feature fusion layer parameters, etc.) are iteratively adjusted through the back propagation algorithm until the evaluation index (such as accuracy, F1 score, score error rate) of the model on the validation set reaches the preset standard, and the training process is completed.

[0048] Through the training mechanism, the scheme evaluation model can learn the matching rules of change requirements and design constraints from historical cases, and gradually improve the evaluation accuracy of new change schemes.

[0049] Of course, in actual application, the above scheme evaluation model can also not be trained, but a mature large language model is directly used as a risk evaluation model to evaluate the initial change scheme.

[0050] Specifically, the above target change requirement can include a change requirement that meets the Change Resilience Triad (CRT), which includes a change release requirement corresponding to Progressive Deployment, a change monitoring requirement corresponding to Comprehensive Observability, and a change recovery requirement corresponding to Rapid Rollback / Circuit-Breaker Capability.

[0051] Among them, the change release requirement needs to ensure that the change is released online in batches, flows, and regions to control the impact range and facilitate problem discovery and rapid loss-making; The change monitoring requirement needs to ensure that there are sufficient monitoring indicators, logs, and tracking data during or after the implementation of the change to real-time understand the change status, business impact, and system health; Change recovery requirements need to ensure that when changes cause unexpected problems, there is the ability to quickly, reliably, one-click, or with minimal human intervention, to roll back to a stable version, or to melt abnormal traffic / functionality.

[0052] In addition, in order to ensure the accuracy of the evaluation results, in the process of evaluating the initial change scheme, software design information can also be considered, which can provide software context information of the target business for the scheme evaluation model and serve as a constraint to assist the model to achieve multi-dimensional verification, such as: At the architecture adaptability level, it is judged whether the progressive deployment scheme is compatible with the existing service registration and discovery mechanism, such as whether the gray strategy can rely on the existing gateway to realize traffic splitting; at the resilience element verification level, combined with the coverage blind area of the software architecture evaluation observability scheme, the effectiveness of the fault isolation design is judged according to the service fuse component; at the risk reasoning level, the typical problems in the same architecture in the history change are associated, and the resource occupation estimation of the current scheme is integrated to identify potential bottlenecks.

[0053] Through the deep integration of software design information, the evaluation results are closely bound to the technical characteristics of the business system, significantly improving the accuracy of risk prediction and the operability of improvement suggestions.

[0054] In addition, the scheme evaluation model can also use software design information to extract system component dependency relationships, deployment topology, and other core information as the underlying basis for evaluating the initial change scheme. For example, by analyzing the call links between microservices in the software design information, it is judged whether the new interface in the change scheme will break the stability of the existing service dependency; according to the database sharding strategy, it is checked whether the data operation involved in the change complies with the sharding rules to avoid cross-sharding transaction risks; combined with performance benchmarking indicators (such as maximum concurrency per node), it is evaluated whether the resource occupation caused by the change scheme is within the system carrying capacity, thereby strengthening the technical adaptability judgment of the evaluation.

[0055] Thus, an evaluation result for indicating whether the initial change information meets the target change requirements reflected by the change reference information under the constraint of the software design information can be obtained.

[0056] In this specification, the change reference information can contain several historical change schemes, and each historical change scheme can correspond to one or more change requirements. If a single historical scheme cannot cover all target change requirements (i.e., cannot completely map the change resilience three elements or other requirement dimensions), the server can determine multiple historical change schemes so that their requirement coverage ranges are complementary, and together they completely reflect the full amount of requirements required by the change reference information, providing comprehensive historical experience reference for the evaluation of the change scheme.

[0057] Specifically, the scheme evaluation model can perform semantic analysis on the prompt word, and based on the semantic analysis result, evaluate whether the implementation process meets the change release demand under the constraint of the software design information under the indication of the prompt text, with the historical change scheme reflecting the corresponding change release demand as a reference. The semantic analysis result can be a feature representation determined by the scheme evaluation model based on an attention mechanism.

[0058] It should be noted that different historical change schemes can be referred to when evaluating different change demands, and of course, the same historical change scheme reflecting all change demands can also be referred to. In addition, in actual application, the evaluation of which change demand can be determined according to the actual evaluation target, such as evaluating only one or two of the change release demand, the change monitoring demand, the change recovery demand, or evaluating all change demands.

[0059] For the failed change scheme contained in the change reference information, the scheme evaluation model can extract the related information that causes the change failure (such as the abnormal traffic split caused by the inadaptation of the gray release strategy to the registration center mechanism, the fault discovery delay caused by the non-coverage of the monitoring index to the core business link, the data consistency destruction caused by the incompatibility of the rollback script to the historical data format, etc.) from the failed change scheme, and then detect whether the related information is involved in the initial change scheme to evaluate whether the initial change scheme has a change risk that causes the change failure.

[0060] The change risk can include an abnormality that occurs in the target business when the initial change scheme is implemented without meeting the at least one target change demand. For example, if a failed change scheme fails due to "not setting gray data isolation rules in the database read-write separation scenario, causing cross-pollution of new and old version data", the scheme evaluation model can extract this related information, and if it is found that the initial change scheme also does not mention the gray data isolation design in the read-write separation scenario, it can be determined that there is a similar failure risk.

[0061] When the scheme evaluation model evaluates the initial change scheme, one or more change demands can be evaluated, and an evaluation result (i.e., whether it meets a certain evaluation demand) or an overall evaluation result (i.e., whether it meets all change demands) for each change demand can be given.

[0062] Of course, in actual application, the evaluation result can also be reflected in the form of a specific score, so as to more accurately quantify the degree to which the initial change scheme meets the change demand.

[0063] Specifically, the scheme evaluation model can determine a final score according to the satisfaction of the change release requirement, the satisfaction of the change monitoring requirement, the satisfaction of the change recovery requirement, whether there is a situation leading to change failure, and the matching with the software design information, with the higher score indicating a higher satisfaction of the change requirement and a lower change risk brought by the change scheme.

[0064] In the formula, the server can set a weight for each dimension, and each dimension is divided into two cases of "satisfaction" and "non-satisfaction". If it is satisfied, the value is 1, and if it is not satisfied, the value is 0. When the weight corresponding to each dimension is the same, that is, 0.2, if the results of all dimensions are satisfied, that is, the change release requirement, the change monitoring requirement, the change recovery requirement, and the matching with the software design information are satisfied, the final score is 1*5*0.2*100% = 100. If one of the dimensions is not satisfied, the final score is 1*4*0.2*100% = 80.

[0065] For example, the server can first determine whether the initial change scheme will lead to change failure. If it will lead to change failure, the score is directly output as 0, and the other dimensions are not scored. If it will not lead to change failure, the final score is obtained by weighted summation according to the results of each dimension and the weight.

[0066] In addition, the server can also set a set score threshold (such as 60), and if it is lower than the threshold, it is determined that the initial evaluation scheme does not satisfy the requirement as a whole.

[0067] Further, before using the above scheme evaluation model, the server can first train it. In this process, the server can obtain a sample change scheme, sample software design information, and sample change reference information, and obtain the actual evaluation result of the sample change scheme as a label. The actual evaluation result can be the expert evaluation result made by the developer for the sample change scheme (such as whether it satisfies a certain change reference information or a score).

[0068] Then the server can construct the prompt word in the same way as above, and output the predicted evaluation result through the scheme evaluation model to be trained. Then, according to the expert evaluation result and the predicted evaluation result, a loss value is determined, and the scheme evaluation model is optimized with the minimum loss value as the optimization goal.

[0069] Of course, in actual application, the scheme evaluation model can also not be trained, but the already open source evaluation model can be directly applied to evaluate the initial change scheme using the prompt word. After obtaining the evaluation result, the scheme evaluation model can also refer to the preferred change scheme in the change reference information, and generate improvement suggestions for the initial change scheme according to the evaluation result.

[0070] For example, if the evaluation result is: “does not meet the requirements of the corresponding change release of progressive deployment”, the scheme evaluation model can determine the verified batch release process from the preferred change scheme and extract the key control nodes (such as the automated verification link after each batch release, the gray ratio threshold of traffic switching) as the basis for improvement to generate specific suggestions.

[0071] For ease of understanding, the present specification provides a whole flowchart of risk assessment, as shown in Figure 3

[0072] Figure 3 A whole flowchart of risk assessment provided by an exemplary embodiment.

[0073] Among them, the server can obtain an initial change scheme and software design information, and perform semantic analysis on the initial change scheme based on a preset analysis engine to obtain structured text, and at the same time, extract software design information from the system analysis document; Then the server can retrieve the change reference information in the knowledge base through the analysis result (structured text) of the initial change scheme; The server can further construct prompt words based on the analysis result of the initial change scheme, the change reference information and the software design information, and input the scheme evaluation model, so that the scheme evaluation model generates an evaluation result, which can include whether the initial change scheme meets the target change requirement, whether the initial change scheme has change risk, and improvement suggestions for the initial change scheme; After generating the evaluation result, the server can determine the risk of implementing business changes based on the initial change scheme for the target business.

[0074] Exemplarily, the above prompt words can be as follows: “As a professional software change scheme evaluation expert, please conduct a comprehensive evaluation of the initial change scheme based on the following three types of key information: Change reference information (target change requirement): Requirement type: change release requirement (progressive release), change monitoring requirement (data observability), change recovery requirement (data recoverability).

[0075] Change content: [upgrade the existing “user points system” function and add “points exchange coupon” function]; ​Implementation process: [Adopt a gradual release strategy: first release to 10% of users for 24 hours, then expand to 50% for 12 hours, and finally release to all users; Real-time monitoring of core indicators is required: interface call success rate (≥99.9%), exchange operation response time (≤500ms), and consistency between point deduction and coupon issuance (100% match); Emergency recovery capabilities are required: if there are indicator abnormalities during the gray period, the system can be rolled back to the original version within 10 minutes, and the user's point data after rollback should be consistent with that before the change.].

[0076] Software design information (design side constraints): Existing system architecture: [Front-end and back-end separation architecture, front-end uses xxx framework, back-end uses xxx micro-service, order data is stored in xxx database, file export service depends on existing "file processing micro-service"]; Design specification constraints: [Back-end interfaces must comply with xxx, request parameters must pass validation, Excel file export must use xxx toolkit (version ≥3.1.0), and new code must comply with the company's development specifications]; Potential conflict points: [For example: the existing file-service service has a maximum of 800 lines per file, and the "order time" field in the t_order table in the xxx database is of type datetime, which needs to be consistent with the Excel export format].

[0077] Initial change scheme: Change content: [Add a "Redeem Coupon" button on the "My Points" page, display the redemption rules popup after clicking, and call the back-end redemption interface after confirmation; Add a point redemption interface; Add daily redemption limit logic]; Implementation process: [Perform gradual release according to requirements, achieve user diversion through configuration; Add log printing in the redemption interface, collect logs through ELK and configure alarms; Execute version rollback scripts through Jenkins].

[0078] Based on the above information, output the evaluation results, which should include the following content: Determine whether the initial change scheme meets the target change requirements reflected in the change reference information under the constraints of the software design information; Determine whether the initial change scheme has risks that could lead to change failure; Finally, give a summary conclusion: whether the initial change scheme "meets the target change requirements reflected in the change reference information under the constraints of the software design information, and based on the evaluation results and the preferred change scheme in the change reference information, give improvement suggestions for the initial change scheme.

[0079] In actual application, the evaluation result output by the scheme evaluation model can not meet the preset evaluation condition, such as incomplete information of the evaluation result, obvious judgment error of the evaluation result, lack of operability of the improvement suggestion, and the like. Therefore, if the evaluation result does not meet the preset evaluation condition, a review task is automatically created, the original evaluation result output by the model, the change reference information, and the key fragments of the software design information need to be attached in the task, and the task is sent to a reviewer who has the technical background of the target business. After obtaining the review result returned by the reviewer, the initial change scheme is adjusted in combination with the review opinion.

[0080] The server can adjust the initial change scheme based on the evaluation result and the improvement suggestion, so as to obtain a target change scheme.

[0081] In the process of implementing the business change on the target business, the server can generate a change code for the target business based on the adjusted change scheme, so as to implement the business change on the target business by using the change code.

[0082] Specifically, the server can input the change scheme and the related original code of the target business into a preset code generation model, so as to parse the change scheme by the code generation model, and generate the change code for the target business based on the parsing result and the related original code, and then implement the business change on the target business by using the change code.

[0083] The related original code can be a local code associated with the module or interface to be changed (such as the source code of the user point management module and the coupon generation module when the point exchange function is changed, or the code segment of the order creation and payment processing module when the order flow is adjusted), or the entire code of the target business. The code generation model can first determine the function points and interaction logic to be modified in combination with the change scheme, and then refer to the existing architecture and coding specification of the related original code to identify the code units (such as interface methods, data processing logic, and page interaction components) to be added, modified, or deleted, so as to obtain incremental change code (including newly added function implementation code, adjustment fragments of existing code, and corresponding configuration updates) compatible with the related original code.

[0084] In addition, the server can also obtain an initial change code of the target service, which can be written by an engineer or generated by an automatic tool (such as a code generation plug-in or a low-code platform). Then, the adjusted change plan and the change code are input into a preset code generation model to optimize the change code (such as adjusting the code structure to adapt to the existing architecture of the target service, supplementing abnormal processing logic to improve robustness, and simplifying redundant code to improve execution efficiency) through the code generation model, and then implementing business changes on the target service using the optimized change code.

[0085] It should be noted that the above code generation model can also be constructed using LLM. The model can be used as a built-in function module of the scheme evaluation model and be linked with the evaluation logic (for example, after the evaluation finds that the scheme needs to supplement rollback code, the built-in model generates the corresponding code fragment directly); or it can be an independently running model that receives change schemes and related original codes through an interface and completes the code generation or optimization task independently.

[0086] Further, the above business change method can be implemented through a corresponding business change system. In order to facilitate understanding, the present specification provides an interaction process diagram of a business change system, as shown in Figure 4 .

[0087] Figure 4 is an interaction process diagram of a business change system provided by an exemplary embodiment, which includes the following steps: S401: a user (developer) submits a document of an initial change scheme on a user end; S402: after the server receives the initial change scheme, the server performs semantic analysis on the initial change scheme; S403: the server queries change reference information related to the analysis result in the knowledge base based on the analysis result; S404: the server generates prompt words based on the change reference information returned by the knowledge base, the initial change scheme, and software design information, and inputs the prompt words into a scheme evaluation model; S405: the scheme evaluation model evaluates the initial change scheme based on the prompt words to obtain an evaluation result; S406: the scheme evaluation model returns the generated evaluation result to the server; S407: the server integrates the evaluation result and feeds back to the user.

[0088] As can be seen from the above method, the present specification realizes deep semantic understanding and complex logical reasoning of the change scheme by means of the core capability of the LLM, effectively overcomes the limitations that static list, keyword matching and simple rule engine cannot parse the context of natural language, cannot identify potential dependencies and risks, and can accurately capture the details, intention and hidden problems of the change scheme.

[0089] At the same time, by fusing the knowledge base and the system analysis document, the short board of the limited knowledge of the artificial review, the inability to fully associate the historical lessons and the current system complexity is made up, and the problem of lack of field knowledge support of other schemes is also solved, so that the evaluation is more in line with the actual business environment of the enterprise.

[0090] On this basis, the present specification realizes the automation and scaling of change evaluation, breaks through the bottleneck of low efficiency of artificial review and difficulty in dealing with a large number of high-frequency changes, and the LLM can further evaluate the rationality, feasibility and completeness of the CRT elements on the basis of judging whether the CRT elements are mentioned, generate specific and context-related improvement suggestions, and avoid the deficiency of other schemes that can only give pass / fail or simple prompts.

[0091] Figure 5 is a schematic structural diagram of a device provided by an exemplary embodiment. As shown in Figure 5 The device 500 mainly consists of a communication interface 501, a user interface 502, a processor 503 and a data storage 504, which are connected and communicate with each other through a system bus, a network or other connection mechanism 505. The communication interface 501 enables the device 500 to communicate with other devices, access networks and transmission networks through analog or digital modulation. For example, the communication interface 501 can include a chipset and an antenna for wireless communication with a radio access network or an access point. In addition, the communication interface 501 can also be a wired interface such as an Ethernet, token ring or USB port, or a wireless interface such as Wifi, Bluetooth, Global Positioning System (GPS) or Wide Area Wireless Interface (e.g. WiMAX or LTE). Of course, the communication interface 501 can also support other forms of physical layer interfaces and standard or proprietary communication protocols. The communication interface 501 can also include multiple physical communication interfaces, such as a Wifi interface, a Bluetooth interface and a wide area wireless interface.

[0092] The user interface 502 includes input components such as a keypad, keyboard, touch- sensitive or presence-sensitive panel, computer mouse, trackball, joystick, microphone, still camera, and video camera, among others, and output components such as a display screen (which can be combined with a touch-sensitive panel), CRT, LCD, LED, display using DLP technology, printer, other known or future developed equivalent devices, among others. The user interface 502 can also generate auditory output through a speaker, speaker jack, audio output port, audio output device, headphones, and other known or future developed equivalent devices. In some embodiments, the user interface 502 can include software, circuitry, or other forms of logic that enables the device 500 to transmit data to and receive data from external user input / output devices. In addition or alternatively, the device 500 can support remote access from other devices through the communication interface 501 or another physical interface (not shown). The user interface 502 can be configured to receive user input, the location and movement of which can be indicated by a pointer or cursor as described herein. The user interface 502 can also be configured as a display device for rendering or displaying a text segment.

[0093] The processor 503 can include one or more general processors and / or dedicated processors.

[0094] The data storage 504 can include one or more volatile and / or non-volatile storage components, and can be integrated in whole or in part with the processor 503. The data storage 504 can include removable and / or non-removable components.

[0095] The processor 503 is capable of executing program instructions 511 (e.g., compiled or interpreted program logic and / or machine code) stored in the data storage 504 to implement various functionality described herein. The data storage 504 can include non-transitory computer-readable media having stored thereon program instructions that, when executed by the device 500, enable the device 500 to carry out any of the methods, processes, or functions disclosed in the specification and / or drawings. Execution of the program instructions 511 by the processor 503 can cause the processor 503 to utilize the data 506.

[0096] For example, the program instructions 511 can include an operating system 510 (e.g., operating system kernel, device drivers, and / or other modules) installed on the device 500, as well as one or more application programs 509 (e.g., a browser, social application, or game application). Similarly, the data 506 can include operating system data 508 and application data 507. The operating system data 508 is primarily accessible to the operating system 510, while the application data 507 is primarily accessible to the one or more application programs 509. The application data 507 can be located in a file system that is visible or hidden to a user of the device 500.

[0097] The application programs 509 can communicate with the operating system 510 through one or more application programming interfaces (APIs). These APIs can facilitate the application programs 509 reading and / or writing to the application data 507, communicating or receiving information via the communication interface 501, receiving or displaying information on the user interface 502, and so on.

[0098] In some terminology, the application programs 509 can be referred to simply as "apps." Furthermore, the application programs 509 can be downloaded to the device 500 through one or more online application stores or application markets. However, the application programs can also be installed on the device 500 through other means, such as through a web browser or a physical interface on the device 500 (e.g., a USB port).

[0099] Reference should be made to Figure 6 The risk assessment device can be applied to the device as shown in Figure 5 , to implement the technical solutions of the present specification. Wherein, the risk assessment device can include: The acquisition module 601 is configured to acquire an initial change scheme for a target service and software design information corresponding to the target service, wherein the initial change scheme includes change content and an implementation process of a service change to be implemented on the target service. The matching module 602 is configured to determine change reference information matched with the initial change scheme in a pre-constructed knowledge base according to the initial change scheme. The evaluation module 603 is configured to fill the change reference information, the software design information, and the initial change scheme into a preset prompt word template, to generate a prompt word based on the change reference information, the software design information, the initial change scheme, and a prompt text in the prompt word template, and to input the prompt word into a scheme evaluation model based on an attention mechanism, so that the scheme evaluation model determines whether the initial change scheme meets target change requirements reflected by the change reference information under the constraint of the software design information under the instruction of the prompt text, to evaluate the initial change scheme and obtain an evaluation result, wherein the target change requirements include at least one of change release requirements, change monitoring requirements, and change recovery requirements. The determination module 604 is configured to determine a change risk of implementing a service change on the target service based on the initial change scheme according to the evaluation result, wherein the change risk includes an abnormality of the target service caused by implementing the initial change scheme without meeting the target change requirements.

[0100] Optionally, the change reference information includes a historical change scheme. The initial change scheme is evaluated, specifically including at least one of the following: With reference to a historical change scheme reflecting corresponding change release requirements, it is evaluated whether the implementation process meets the change release requirements of the incremental release under the constraint of the software design information; With reference to a historical change scheme reflecting corresponding change monitoring requirements, it is evaluated whether the change content can generate externally observable data that meets the change monitoring requirements after being implemented under the constraint of the software design information; With reference to a historical change scheme reflecting corresponding change recovery requirements, it is evaluated whether the implementation process has a data rollback operation path that meets the change recovery requirements under the constraint of the software design information.

[0101] Optionally, the change reference information includes a failed change scheme; The evaluation module 603 is specifically configured to extract relevant information causing change failure from the failed change scheme; It is detected whether the initial change scheme involves the relevant information, so as to evaluate whether the initial change scheme has a risk of causing change failure.

[0102] Optionally, the change reference information includes a preferred change scheme; The evaluation module 603 is further configured to generate improvement suggestions for the initial change scheme according to the evaluation result, with reference to the preferred change scheme; The apparatus further includes a change module 605 configured to adjust the initial change scheme according to the improvement suggestions to obtain an adjusted change scheme, determine a target change code for the target business according to the adjusted change scheme, and implement business change for the target business by using the target change code.

[0103] Optionally, the change module 605 is configured to input the adjusted change scheme and relevant original code of the target business into a preset code generation model, analyze the change scheme by using the code generation model, and generate a target change code for the target business based on an analysis result and on the basis of the relevant original code.

[0104] Optionally, the change module 605 is configured to input the adjusted change scheme and relevant original code of the target business into a preset code generation model, analyze the change scheme by using the code generation model, and generate a target change code for the target business based on an analysis result and on the basis of the relevant original code.

[0105] Optionally, the changing module 605 is specifically configured to: acquire an initial change code for the target service; and input the adjusted change scheme and the initial change code into a preset code generation model, so as to optimize the initial change code by the code generation model, and obtain the target change code.

[0106] For ease of description, the above apparatus is described in various modules or units in terms of functions. Of course, the functions of each module or unit can be implemented in one or more software and / or hardware in implementing one or more of the present description, and the modules having the same functions can be combined or integrated into another system, or some features can be ignored or not implemented.

[0107] Based on the same idea as the above method, the present description also provides an electronic device, comprising: a processor; a memory for storing processor executable instructions; wherein the processor implements the steps of the method according to any one of the above embodiments by running the executable instructions.

[0108] Based on the same idea as the above method, the present description also provides a computer readable storage medium, which stores computer instructions, and the instructions are executed by a processor to implement the steps of the method according to any one of the above embodiments.

[0109] Based on the same idea as the above method, the present description also provides a computer program product, which comprises computer program / instructions, and the computer program / instructions are executed by a processor to implement the steps of the method according to any one of the above embodiments.

[0110] Those skilled in the art can understand that: In the present description, the terms "comprise", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that processes, methods, products or devices comprising a series of elements do not only include those elements, but also include other elements not explicitly listed, or further include elements inherent to such processes, methods, products or devices. Without more limitations, it does not exclude the presence of other same or equivalent elements in the processes, methods, products or devices comprising the elements.

[0111] In the present description, "one", "a" and "the" do not refer to the singular, but also include the plural.

[0112] In the present specification, the first, second, and the like ordinal numbers do not necessarily mean a sequence or order, and in many cases, are used for the purpose of distinguishing an object from another. For example, a first server and a second server are used for the purpose of distinguishing the two servers from each other. Of course, the two servers can be the same in some cases.

[0113] In the present specification, unless explicitly stated otherwise, "receiving and transmitting data" is not necessarily direct reception and transmission, and can be indirect reception and transmission. For example, A receiving data transmitted by B can be understood as A directly receiving data transmitted by B, or can be understood as A indirectly receiving data transmitted by B through C or the like. Similarly, B transmitting data to A can be understood as B directly transmitting data to A, or can be understood as B indirectly transmitting data to A through C or the like. Here, C can be one subject, or two or more subjects.

[0114] In the present specification, unless explicitly stated otherwise, the association relationship generated between structures can be a direct association relationship or an indirect association relationship. For example, when "A is connected to B" is described, unless it is explicitly stated that A is directly connected to B, it should be understood that A can be directly connected to B or indirectly connected to B; for another example, when "A is on B", unless it is explicitly stated that A is directly above B (AB is adjacent and A is above B), it should be understood that A can be directly above B or A can be indirectly above B (AB is separated by other elements and A is above B). By analogy.

[0115] The present specification uses specific words to describe the embodiments of the present specification. As "one embodiment", "an embodiment", and / or "some embodiments" means a certain feature, structure, or characteristic related to at least one embodiment of the present specification. Therefore, it should be emphasized and noted that the "an embodiment" or "one embodiment" or "one alternative embodiment" mentioned in different places in the present specification does not necessarily refer to the same embodiment. In addition, the skilled in the art can combine and combine the different embodiments or examples described in the present specification and the features of the different embodiments or examples without contradiction.

[0116] Although the present specification provides method steps as described in the embodiments or flowcharts, it can be understood that the order of steps listed in the embodiments or flowcharts is only one of the many execution orders, and does not represent the only execution order. Therefore, when the claims involve method steps, the adjustment of the order of the steps or the parallelism between the steps is also within the scope of protection of the claims.

Claims

1. A risk assessment method, characterized in that: include: Obtaining an initial change plan for a target business and software design information corresponding to the target business; wherein the initial change plan includes: the change content and implementation process of the business change to be implemented for the target business; According to the initial change plan, matching change reference information is determined in a pre-built knowledge base; Filling the change reference information, the software design information, and the initial change plan into a preset prompt word template to generate a prompt word based on the change reference information, the software design information, the initial change plan, and the prompt text in the prompt word template, and inputting the prompt word into a solution evaluation model based on an attention mechanism, so that the solution evaluation model, under the guidance of the prompt text, determines whether the initial change plan meets the target change requirement reflected by the change reference information under the constraints of the software design information, thereby evaluating the initial change plan and obtaining an evaluation result; wherein the target change requirement includes: at least one of a change release requirement, a change monitoring requirement, and a change recovery requirement; According to the evaluation result, the change risk of implementing a business change on the target business based on the initial change plan is determined, wherein the change risk includes: anomalies in the target business caused by implementing the initial change plan without meeting the target change requirements.

2. The method according to claim 1, wherein The change reference information includes historical change plans; Evaluate the initial change plan, including at least one of the following: Using historical change plans reflecting corresponding change release requirements as a reference, assess whether the implementation process meets the change release requirements for progressive release under the constraints of the software design information; Using historical change plans that reflect corresponding change monitoring requirements as a reference, assess whether the changes, when implemented under the constraints of the software design information, can generate externally observable data that meets the change monitoring requirements; With reference to the historical change plan reflecting the corresponding change recovery requirements, an evaluation is performed on whether the implementation process has a data rollback operation path that meets the change recovery requirements under the constraints of the software design information.

3. The method according to claim 1, wherein The change reference information includes a failed change plan; Evaluate the initial change proposal, including: extracting relevant information causing the change failure from the failed change plan; Detect whether the initial change plan involves the relevant information to assess whether the initial change plan has a risk of causing change failure.

4. The method according to claim 1, wherein The change reference information includes a preferred change plan; The evaluation of the initial change proposal also includes: With reference to the preferred change plan, an improvement suggestion for the initial change plan is generated according to the evaluation result.

5. The method according to claim 4, wherein The method further comprises: Adjusting the initial change plan according to the improvement suggestions to obtain an adjusted change plan; According to the adjusted change plan, a target change code for the target business is determined, and the business change is implemented on the target business using the target change code.

6. The method according to claim 5, wherein Determine the target change code for the target business based on the adjusted change plan, specifically including: The adjusted change plan and the relevant original code of the target business are input into a preset code generation model, so that the change plan is parsed by the code generation model, and based on the parsing result, the target change code for the target business is generated on the basis of the relevant original code.

7. The method according to claim 5, wherein Determine the change code for the target business based on the adjusted change plan, specifically including: Obtaining an initial change code for the target business; The adjusted change plan and the initial change code are input into a preset code generation model, so that the initial change code is optimized by the code generation model to obtain the target change code.

8. A service change device, characterized in that: include: An acquisition module is configured to acquire an initial change plan for a target business and software design information corresponding to the target business; wherein the initial change plan includes: the change content and implementation process of the business change to be implemented for the target business; A matching module, configured to determine, based on the initial change plan, change reference information that matches the initial change plan in a pre-built knowledge base; An evaluation module is configured to fill a preset prompt word template with the change reference information, the software design information, and the initial change plan, generate a prompt word based on the change reference information, the software design information, the initial change plan, and the prompt text in the prompt word template, and input the prompt word into a solution evaluation model based on an attention mechanism, so that the solution evaluation model, under the guidance of the prompt text, determines whether the initial change plan meets the target change requirement reflected by the change reference information under the constraints of the software design information, thereby evaluating the initial change plan and obtaining an evaluation result; wherein the target change requirement includes at least one of a change release requirement, a change monitoring requirement, and a change recovery requirement; A determination module is used to determine the change risk of implementing business changes to the target business based on the initial change plan according to the evaluation results, wherein the change risk includes: anomalies in the target business caused by implementing the initial change plan without meeting the target change requirements.

9. An electronic device comprising: A processor; characterized by a memory for storing processor-executable instructions; wherein the processor implements the steps of the method according to any one of claims 1 to 7 by running the executable instructions.

10. A computer-readable storage medium, characterized in that Computer instructions are stored thereon, and when the instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

11. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images