Platform management method and system based on low-code development

Abstract

The present invention relates to a platform management method and system based on low-code development, the method comprising: obtaining multiple heterogeneous data sources, extracting first data from the multiple heterogeneous data sources, and preprocessing the first data to obtain standard data for constructing a knowledge graph. A plurality of different entities and entity relationships between different entities are obtained from the standard data, and a knowledge graph is constructed based on the plurality of different entities and entity relationships between different entities. The entity nodes in the knowledge graph and the entity relationships between different entity nodes are mapped to a low-dimensional space for vector representation through a graph embedding model to extract global knowledge including entity relationships between different entity nodes and network topology structures. A predefined business process and a plurality of decision points corresponding to the execution of the business process are obtained, and the execution results of the business process are evaluated by executing global knowledge and current business process variables to obtain corresponding guiding decisions.

Description

Technical Field

[0001] The present invention relates to the field of low-code development and data management technology, and in particular to a platform management method and system based on low-code development. Background Art

[0002] In today's rapidly advancing digitalization and automation landscape, enterprises face increasingly complex data management and business process optimization demands. Low-code development platforms have become increasingly popular in recent years due to their simplified programming requirements and improved development efficiency. These platforms allow non-technical users to quickly build applications through graphical interfaces, significantly shortening development cycles and reducing development costs. However, while low-code platforms offer the convenience of building applications, they often lack the capabilities to handle complex data integration, in-depth data analysis, and advanced business process automation.

[0003] With the rise of big data and artificial intelligence, businesses urgently need advanced tools that can integrate massive data sets from multiple data sources and extract useful insights to support decision-making. Currently, most low-code platforms struggle to provide effective tools to support this need, which has limited their application in data-driven decision support systems. However, future trends are likely to place greater emphasis on integrating artificial intelligence and machine learning technologies, enabling low-code platforms to move beyond rapid application development to become support systems for intelligent decision-making. This includes leveraging advanced data processing and analysis technologies, such as natural language processing and predictive analytics, to improve the accuracy and timeliness of decision-making. Furthermore, with the increasing popularity of cloud computing, more businesses are choosing cloud infrastructure to deploy their low-code development platforms, which not only provides enhanced data processing capabilities but also greater scalability and flexibility. In this environment, data security and privacy protection will also become key considerations when developing these platforms.

[0004] Despite their significant advantages, low-code development platforms still face challenges in handling complex business logic and maintaining system flexibility and scalability. Currently, existing low-code platforms often struggle to fully meet enterprise needs for data integration and automated advanced business processes. Their capabilities for automating large-scale data processing and providing deep data insights are generally limited, limiting their application in certain highly specialized industries. Mainstream low-code platforms currently on the market provide a visual development environment, allowing developers to design user interfaces, configure data models, and implement business logic by dragging and dropping components. These platforms typically integrate basic data processing capabilities and third-party API support to facilitate integration with existing IT infrastructure and external services. However, these platforms have limited support for complex data operations. While they can handle some basic data integration tasks, their functionality is often insufficient for large-scale data integration, advanced data cleansing and transformation, and complex data relationship analysis. Consequently, these existing low-code platforms often lack sufficient flexibility and configuration depth to automate complex and non-standardized business processes. Summary of the Invention

[0005] Based on this, it is necessary to address the above technical issues and provide a platform management method and system based on low-code development that has high flexibility and configuration depth when executing complex and non-standardized automated business processes.

[0006] The present invention provides a platform management method based on low-code development, the method comprising:

[0007] Acquire multiple heterogeneous data sources, extract first data from the multiple heterogeneous data sources, and preprocess the first data to obtain standard data for constructing a knowledge graph, wherein the preprocessing includes data cleaning and data standardization;

[0008] Acquire a plurality of different entities and entity relationships between different entities from the standard data, and construct a knowledge graph based on the plurality of different entities and entity relationships between different entities;

[0009] Mapping the entity nodes and entity relationships between different entity nodes in the knowledge graph to a low-dimensional space for vector representation through a graph embedding model to extract global knowledge including the entity relationships between the different entity nodes and the network topology structure;

[0010] Obtaining a predefined business process and a plurality of decision points corresponding to the execution of the business process, and evaluating the execution results of the business process by executing the global knowledge and current business process variables to obtain corresponding guiding decisions;

[0011] Among them, the first data is useful data or key data that meets the set conditions, the standard data is composed of multiple different entities and entity relationships between different entities, the current business process variables are the business environment variables when the predefined business process is currently executed, and the guiding decision is used to optimize the management of the predefined business process.

[0012] In one embodiment, the step of obtaining multiple heterogeneous data sources, extracting first data from the multiple heterogeneous data sources, and preprocessing the first data to obtain standard data for constructing a knowledge graph includes:

[0013] Extracting a plurality of first data sources from the plurality of heterogeneous data sources according to the set conditions, and extracting a plurality of data columns that meet the set conditions from the plurality of first data sources;

[0014] A fusion algorithm is called to integrate the multiple data columns to obtain a data set, and different data sets are merged according to a set key, and the data in the data set is the first data.

[0015] In one embodiment, the step of obtaining multiple heterogeneous data sources, extracting first data from the multiple heterogeneous data sources, and preprocessing the first data to obtain standard data for constructing a knowledge graph further includes:

[0016] The missing values in the first data are replaced, and abnormal data in each data column whose values do not meet the set threshold are deleted to clean the first data, and the cleaned first data are standardized to obtain the standard data.

[0017] In one embodiment, obtaining a plurality of different entities and entity relationships between different entities from the standard data, and constructing a knowledge graph based on the plurality of different entities and entity relationships between different entities, includes:

[0018] Determine entity attributes, entity relationship types, and entity relationship weights based on the multiple different entities and the entity relationships between the different entities;

[0019] Based on the entity attributes, entity relationship types and entity relationship weights, the knowledge graph is constructed with multiple different entities as nodes and the entity relationships between the different entities as edges;

[0020] Among them, each entity corresponds to a node, and the entity relationship between each two different entities corresponds to an edge.

[0021] In one embodiment, mapping the entity nodes and the entity relationships between different entity nodes in the knowledge graph to a low-dimensional space for vector representation through a graph embedding model to extract global knowledge including the entity relationships between the different entity nodes and the network topology structure includes:

[0022] Selecting a first entity node and a second entity node from the knowledge graph, and obtaining an entity relationship weight between the first entity node and the second entity node, where the first entity node and the second entity node are directly connected neighbor nodes and have an entity relationship;

[0023] The graph embedding model is called to map the first entity node, the second entity node, and the entity relationship between the first entity node and the second entity node to a vector representation in a low-dimensional space according to the entity relationship weight.

[0024] In one embodiment, obtaining a predefined business process and a plurality of decision points corresponding to the execution of the business process, and evaluating the execution result of the business process by executing the global knowledge and current business process variables to obtain a corresponding guiding decision includes:

[0025] Executing the predefined business process according to the set business rules to obtain the multiple decision points and determine the business environment variables when the business process is currently executed;

[0026] Obtaining different execution states and execution results by executing the business process under the global knowledge and the business environment variables, and evaluating the multiple decision points based on the different execution states and execution results to generate the guiding decision;

[0027] The predefined business process is updated and optimized based on the guiding decision to generate a new business process.

[0028] In one embodiment, the method further comprises:

[0029] Obtaining the current resource configuration corresponding to the predefined business process and the constraints satisfied by the resource configuration optimization, and calling the minimization cost function to perform cost minimization processing on the current resource configuration;

[0030] The historical data of the business process execution corresponding to the current resource configuration is analyzed and evaluated based on the global knowledge, and the optimized resource configuration is generated by combining the constraint conditions and the minimization cost function.

[0031] The present invention also provides a platform management system based on low-code development, the system comprising:

[0032] A data preprocessing module, configured to obtain multiple heterogeneous data sources, extract first data from the multiple heterogeneous data sources, and preprocess the first data to obtain standard data for constructing a knowledge graph, wherein the preprocessing includes data cleaning and data standardization;

[0033] A knowledge graph construction module, configured to obtain a plurality of different entities and entity relationships between different entities from the standard data, and to construct a knowledge graph based on the plurality of different entities and entity relationships between different entities;

[0034] A global knowledge extraction module is used to map the entity nodes and the entity relationships between different entity nodes in the knowledge graph to a low-dimensional space for vector representation through a graph embedding model, so as to extract global knowledge including the entity relationships between the different entity nodes and the network topology structure;

[0035] An evaluation and guidance module is used to obtain a predefined business process and a plurality of decision points corresponding to the execution of the business process, and evaluate the execution results of the business process by executing the global knowledge and current business process variables to obtain corresponding guidance decisions;

[0036] Among them, the first data is useful data or key data that meets the set conditions, the standard data is composed of multiple different entities and entity relationships between different entities, the current business process variables are the business environment variables when the predefined business process is currently executed, and the guiding decision is used to optimize the management of the predefined business process.

[0037] The present invention also provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the platform management method based on low-code development as described in any one of the above.

[0038] The present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, implements the platform management method based on low-code development as described above.

[0039] The aforementioned low-code platform management method and system obtains data from multiple heterogeneous data sources, extracts data from these sources, and pre-processes the extracted data through data cleansing and standardization to obtain standard data for constructing a knowledge graph. Subsequently, multiple entities and their relationships are extracted from the obtained standard data, and a knowledge graph is constructed based on these entities and their relationships. A graph embedding model is then used to map the entity nodes and their relationships in the knowledge graph into a low-dimensional space for vector representation, thereby extracting global knowledge, including the entity relationships between the different entity nodes and the network topology. Finally, a predefined business process and multiple decision points corresponding to the business process execution are obtained. The execution results of the business process are evaluated by applying this global knowledge and current business process variables to obtain corresponding guiding decisions, ultimately achieving optimized management of the predefined business process. This method integrates data processing capabilities, deep learning and intelligent algorithms, sophisticated business process automation tools, and business process optimization capabilities within a single low-code platform, achieving highly efficient data integration, in-depth data analysis and intelligent decision-making, flexible business process automation optimization, and cost-effective resource management. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0041] Figure 1 This is one of the flow charts of the platform management method based on low-code development provided by the present invention;

[0042] Figure 2 A schematic diagram of the overall platform management process of the platform management method based on low-code development in the specific embodiment provided by the present invention;

[0043] Figure 3 This is the second flow chart of the platform management method based on low-code development provided by the present invention;

[0044] Figure 4 This is the third flow chart of the platform management method based on low-code development provided by the present invention;

[0045] Figure 5 This is the fourth flow chart of the platform management method based on low-code development provided by the present invention;

[0046] Figure 6This is the fifth flow chart of the platform management method based on low-code development provided by the present invention;

[0047] Figure 7 This is the sixth flow chart of the platform management method based on low-code development provided by the present invention;

[0048] Figure 8 A schematic diagram of the structure of the platform management system based on low-code development provided by the present invention;

[0049] Figure 9 This is a diagram of the internal structure of the computer device provided by the present invention. DETAILED DESCRIPTION

[0050] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0051] The following combination Figures 1-9 Describe the platform management method and system based on low-code development of the present invention.

[0052] like Figure 1 As shown, in one embodiment, a platform management method based on low-code development includes the following steps:

[0053] Step S110: obtain multiple heterogeneous data sources, extract first data from the multiple heterogeneous data sources, and preprocess the first data to obtain standard data for constructing a knowledge graph. The preprocessing includes data cleaning and data standardization.

[0054] Specifically, the server obtains multiple heterogeneous data sources, i.e., multi-source data, extracts useful data or key data that meets the set conditions, i.e., first data, from the multiple heterogeneous data sources, and performs data cleaning and data standardization preprocessing on the extracted first data to obtain standard data for constructing a knowledge graph.

[0055] Among them, data standardization includes data fusion or integration and the merging or combining of data columns or data sets. Standard data is composed of multiple different entities and entity relationships between different entities.

[0056] Combine Figure 2As shown, in a specific embodiment, the platform management method based on low-code development provided by the present invention first requires data integration and preprocessing. During this process, the system automatically extracts data from multiple data sources and optimizes data quality and consistency by applying complex cleaning and fusion algorithms. Data integration first involves extracting relevant data from various databases, file systems or online interfaces, and then removing incomplete, erroneous or irrelevant data through a data cleaning process to ensure the accuracy and availability of the data. Ultimately, data from different sources are integrated through data fusion technology to create a unified data view, which provides a basis for subsequent data analysis and decision support.

[0057] In this embodiment, data integration and preprocessing include three main steps: data extraction, data cleaning, and data fusion. These steps need to deal with data inconsistencies from different sources, such as format differences, duplicate data, and missing values.

[0058] The algorithm formula for data extraction is:

[0059] data_extracted=CONCATENATE(F_extract(data_source_i,SELECT(c1,c2,...,cn)FROM conditions)).

[0060] Where data_source_i represents the i-th data source, such as a database, CSV file, or API. SELECT(c1,c2,...,cn) represents the specific columns extracted from the data source, such as user ID, timestamp, sales volume, etc. conditions represents the conditions that must be met when extracting data, such as a specific date range or specific data values. CONCATENATE is a function used to combine data extracted from different data sources to form a unified dataset. data_extracted represents the extracted and consolidated data.

[0061] The algorithm formula for data cleaning is:

[0062] Cleaned_data=F_clean(data_extracted,REPLACE(Null,'Unknown'),DELETEFROM data_extracted WHERE c1>threshold).

[0063] REPLACE(Null,'Unknown') replaces null values in the data with 'Unknown', preventing errors caused by missing values in subsequent processing. DELETE FROM data_extracted WHERE c1>threshold deletes records whose data values in a column exceed the set threshold, removing outliers and ensuring data quality. Cleaned_data is the data after null values are replaced and outliers are removed.

[0064] The algorithm formula for data fusion is:

[0065] Integrated_data=MERGE(Cleaned_data1,Cleaned_data2 ON key1=key2USINGmethod).

[0066] MERGE combines two datasets into one based on a specified key. ON key1 = key2 specifies the key to merge the two datasets, such as a user ID or product code. USING method specifies the merge method, such as averaging, taking the maximum value, or taking the latest value. Integrated_data provides a comprehensive view of the merged dataset for in-depth data analysis and reporting.

[0067] In practical applications, such as in the retail industry, sales data from different stores may be stored in different database systems, and each store may have different data entry standards, such as inconsistent date formats and currency units. The data integration and preprocessing process first aggregates this data into a central warehouse through data extraction. The data cleaning process standardizes the date format, converts the currency unit, and cleans up illogical sales records (such as data with abnormally high sales). Finally, the cleaned data from all stores is merged through the data fusion process to form a unified data set that can be used to generate national sales reports. This process not only improves the accuracy and consistency of the data, but also greatly enhances the analyzability of the data, providing companies with timely and accurate business insights and supporting more effective decision-making.

[0068] Step S120: Acquire multiple different entities and entity relationships between different entities from standard data, and construct a knowledge graph based on the multiple different entities and entity relationships between different entities.

[0069] Specifically, the server obtains multiple different entities and entity relationships between different entities from the standard data obtained in step S110, and completes the construction of the knowledge graph based on the multiple different entities and entity relationships between different entities.

[0070] Combine Figure 2 As shown in the following example, in a specific embodiment of the present invention, the platform management method based on low-code development is provided. Knowledge graphs, as a powerful means of data organization and representation, structuredly integrate information from different sources to form a graphical representation of entities and their relationships. This representation not only reveals complex relationships between data but also supports complex queries and data analysis, forming the foundation of intelligent systems such as recommendation systems, automated question-answering systems, and complex decision support systems.

[0071] The algorithm formula for constructing the knowledge graph is:

[0072] G=create_graph((entity1,attr1,attr2,...,attrN),(relationship(entity1,entity2),type_of_relationship,strength_of_relationship)).

[0073] G is the constructed knowledge graph, which is a graph consisting of nodes (entities) and edges (relationships). entity1 and entity2 are entities in the graph, which can be people, places, objects, concepts, etc. attr1, attr2,..., attrN are attributes of the entities, which describe their characteristics, such as name, date, and location. relationship(entity1, entity2) represents the relationship between entities, such as "works at," "is located in," and "has," etc. type_of_relationship is the type of entity relationship, indicating the nature of the relationship between entities. strength_of_relationship is the strength or weight of the entity relationship, which can indicate the importance, frequency, or confidence of the relationship.

[0074] For example, a large multinational company might construct a knowledge graph to optimize resource allocation and enhance internal collaboration. Internal data sources might include human resources databases, project management systems, and customer relationship management systems. Entities might include employees, departments, projects, and customers. Employee attributes might include name, position, skills, department, and project experience. Project attributes might include project name, related departments, responsible person, and timeline. The relationship between an employee and a project might be "participated in," with a relationship type of "participated," and strength determined by participation time or contribution. The relationship between an employee and a department might be "belongs to," simply indicating the affiliation between the employee and the department. Using this information to construct a knowledge graph allows the company to more effectively plan its human resources and optimize team composition by analyzing employee skills and project requirements. When employees depart or new projects are launched, quickly querying the knowledge graph can help users find potential employees with the required skills and experience. The knowledge graph also supports more complex queries, such as "Find all employees who have participated in at least three projects and possess data science skills," which is crucial for staffing new projects. Furthermore, by analyzing the strength of relationships between employees and projects, management can assess employees' core skills and potential career paths, further supporting career development and training initiatives. In this way, knowledge graphs not only enhance the visualization and queryability of data, but also greatly improve the decision-making quality and operational efficiency of organizations.

[0075] In step S130, the entity nodes in the knowledge graph and the entity relationships between different entity nodes are mapped to a low-dimensional space for vector representation through a graph embedding model to extract global knowledge including the entity relationships between different entity nodes and the network topology structure.

[0076] Specifically, the server uses a graph embedding model to map the entity nodes in the knowledge graph and the entity relationships between different entity nodes to a low-dimensional space for vector representation, so as to extract global knowledge including the entity relationships between different entity nodes and the network topology structure, that is, the extracted global knowledge is the vector representation of the low-dimensional space.

[0077] Combine Figure 2As shown, in a specific embodiment, the platform management method based on low-code development provided by the present invention, after the knowledge graph is built, the system already has a rich entity and relationship network, which lays the foundation for extracting valuable information from the graph. The global knowledge extraction process converts the complex graph structure into a vector form that can be used for machine learning through the graph embedding model. This process directly depends on the structure and quality of the knowledge graph, and the quality and effect of the graph embedding are directly affected by the quality of the graph. Global knowledge extraction is the process of abstracting and extracting useful information from the knowledge graph using graph embedding technology. Graph embedding technology maps the nodes (entities) and edges (relationships) in the graph to vector representations in a low-dimensional space, so that these vectors can capture and retain the key information of the original graph structure, including the relationships between nodes and the network topology. Such vector representations can be used for various machine learning tasks, such as classification, prediction, and recommendation systems, thereby enhancing the performance of decision support systems.

[0078] The algorithm formula for global knowledge extraction is:

[0079] global_knowledge=embed_graph(G,model,SUM_OVER(neighbors(v)*weight(v,u)FOR u IN V)).

[0080] Where G is the constructed knowledge graph, represented as a graph consisting of nodes and edges. model is a graph embedding model, such as Node2Vec and GraphSAGE, which is used to convert graph data into points in a vector space. neighbors(v) is the set of neighbor nodes of node v, which are other nodes directly connected to node v. weight(v,u) is the weight of the edge between nodes v and u, indicating the strength or importance of the relationship between the two nodes. V is the set of all nodes in the graph. global_knowledge is the global knowledge extracted by the graph embedding model, represented as a vector representation of the node.

[0081] Taking social network analysis as an example, a typical application scenario is using graph embedding technology to identify influential users or communities. A social network can be modeled as a knowledge graph, with individual users as nodes and interactions between users, such as comments, shares, or likes, as edges. Edge weights can be based on the frequency or depth of interactions. In a social network, each user is a node, and interactions between users, such as friendships or shared discussions, constitute edges. Edge weights can be based on the quantity or quality of interactions, such as the number of shared likes. Using graph embedding models such as GraphSAGE, each user's social network pattern is converted into a vector. By aggregating neighbor information for each user node (taking into account edge weights), a vector representing the user's social behavior is generated. Based on similarity in the vector space, the community structure in the social network is identified, identifying core users and influential nodes. These nodes are central in the vector space or connected to multiple important communities. Furthermore, these vectors can be used to predict user behavior, such as likely purchase decisions and new product acceptance. By analyzing the extracted global knowledge, marketing teams can design more effective advertising strategies, targeting specific influential users to promote products and leveraging their social influence to increase product visibility and appeal. It makes it feasible and effective to extract useful information from huge and complex social network data, supporting enterprises to make accurate data-based decisions in a highly competitive market environment.

[0082] Step S140 , obtaining a predefined business process and a plurality of decision points corresponding to the execution of the business process, and evaluating the execution result of the business process by executing global knowledge and current business process variables to obtain corresponding guiding decisions.

[0083] Specifically, the server obtains the predefined business process and multiple decision points corresponding to the execution of the business process, and evaluates the execution results of the business process by executing global knowledge and the business environment variables of the predefined business process during its current execution (i.e., the current business process variables), and obtains corresponding guiding decisions. The guiding decisions are used to optimize the management of the predefined business process and update the business process in real time.

[0084] Combine Figure 2 As shown, in a specific embodiment, the platform management method based on low-code development provided by the present invention also includes a business process automation and optimization process, which improves various aspects of the business through intelligent tools, making the entire operation more efficient and accurate. In this process, the data and performance feedback generated by the system can be further analyzed to provide real-time business insights for resource management. These insights flow directly into the resource prediction and optimization configuration process. The system uses the knowledge obtained from the business process to predict resource requirements and formulate a more reasonable resource configuration, including the allocation of physical resources and the optimization of human and financial resources.

[0085] In this embodiment, the previously extracted global knowledge is used to guide or optimize the execution of the business process. The automated process can not only perform routine tasks, but also make intelligent decisions based on dynamic data and contextual environment at complex decision points.

[0086] The algorithm formula for business process automation and optimization is:

[0087] executed_process=execute_process(defined_process,IF condition THENaction ELSE alternative,EVALUATE(global_knowledge,context)).

[0088] Among them, defined_process is a predefined business process, which can be a flowchart or step description of any business operation, such as order processing, customer service response, etc. IF condition THEN action ELSE alternative describes the decision point encountered in the execution of the business process, and selects to execute action or alternative according to the condition. These conditions may be based on business rules or dynamic data analysis. EVALUATE(global_knowledge,context) represents a function that uses extracted global knowledge and current business environment variables (context) to evaluate and guide decisions. It can involve complex logic and machine learning model outputs, such as risk assessment, customer preference prediction, etc. context is the current business environment variables. These variables include but are not limited to customer data, market conditions, resource status, etc., which are crucial for the evaluation of decision points. executed_process is the process state after execution, which indicates the result of the process execution. It can be an updated database record, a delivered order, or other business output.

[0089] For example, in an online retailer's order processing system, order processing is a complex business process involving multiple processes, including inventory management, payment processing, logistics, and customer communications. Automation and optimization within this process can significantly improve efficiency and customer satisfaction. First, define a standard order processing workflow, including order receipt, inventory check, payment confirmation, shipping, and customer notification. During the inventory check phase, if a product is out of stock, the system automatically decides whether to recommend an alternative product or notify the customer of a delayed shipment. Using the "IF condition THEN action ELSE alternative" construct, if the inventory quantity is less than the order quantity, a similar product is recommended; otherwise, the normal process continues. "EVALUATE(global_knowledge,context)" can be used to evaluate the acceptance rate of product substitutions based on historical data or to personalize recommendations based on customer purchase history. Global knowledge can include inventory trends, seasonal demand fluctuations, and other aspects, helping the system make more intelligent decisions. The process is executed based on the evaluation results, and the results of each step are fed back into the system for monitoring and further optimization. Additionally, the execution result "executed_process" can be a confirmation of successful order processing or a report of any processing anomalies. Through this automation and optimization, retailers can process orders more efficiently, reduce customer churn due to out-of-stock situations, optimize inventory management, and improve the quality of customer service. In addition, the automatically recorded data can be used for subsequent business strategy adjustments and system improvements.

[0090] This embodiment also includes a resource forecasting and optimization process, which is a key component of enterprise resource management, especially in resource-constrained and cost-sensitive business environments. This process involves using global knowledge extracted from data analysis and knowledge graphs, combined with current resource conditions, to predict future demand and optimize resource allocation accordingly. The goal is to maximize resource utilization efficiency and minimize operating costs, while ensuring that business requirements and constraints are met.

[0091] The algorithm formula for resource prediction and optimal configuration is:

[0092] optimized_resources=optimize(global_knowledge,current_resources,MINIMIZE(cost_function)SUBJECT TO constraints).

[0093] Among them, global_knowledge is the extracted global knowledge, which includes historical data analysis, trend forecasts, demand patterns and other information, which is used to guide resource allocation decisions. Current_resources is the current available resource status, including inventory levels, human resources, financial resources, etc. MINIMIZE(cost_function) means that the goal is to minimize the cost function, which usually involves reducing resource waste, reducing excessive inventory, optimizing manpower allocation, etc. SUBJECT TO constraints are the constraints that need to be met in the optimization process, such as budget constraints, supply chain capabilities, market demand, regulatory requirements, etc. optimized_resources is the optimized resource allocation, and the result aims to provide the most effective resource utilization strategy to support business operations.

[0094] For example, a manufacturing company faces the challenge of effectively managing production line operations amidstability in raw material supply and volatile market demand. In this context, resource forecasting and optimal allocation become crucial. Global knowledge, including analysis of market trends, review of historical sales data, and consideration of seasonal variations, can help companies predict product demand for the coming months. Furthermore, detailed records of current raw material inventory, machine utilization, and staffing levels are used to assess financial resources and determine available funds for procurement and investment. Cost functions can include raw material costs, labor costs, storage costs, and opportunity costs (revenue lost due to production line downtime). Constraints include production capacity, delivery time, quality standards, and environmental impact. Mathematical programming methods (such as linear programming and integer programming) are used to solve optimization problems. Algorithms are used to try different resource allocations, calculate costs under various scenarios, and identify the lowest-cost configuration. The optimization results guide adjustments to the production plan, such as increasing production of high-demand products, reducing inventory of low-demand products, and rescheduling worker shifts. The optimized resource allocation is implemented, and any deviations during execution are monitored. The strategy is adjusted based on actual performance, and feedback is collected to fine-tune the global knowledge base and refine subsequent forecasting and optimization models. In this way, manufacturing companies can better cope with market uncertainties, optimize production efficiency, and reduce resource waste, thereby maintaining competitiveness in a highly competitive market. This not only improves economic benefits but also enhances the company's sustainable development capabilities.

[0095] The aforementioned low-code platform management method acquires and extracts data from multiple heterogeneous data sources, pre-processing the extracted data through data cleansing and standardization to generate standard data for constructing a knowledge graph. Subsequently, multiple entities and their relationships are extracted from the obtained standard data, and a knowledge graph is constructed based on these entities and their relationships. A graph embedding model is then used to map the entity nodes and their relationships in the knowledge graph into a low-dimensional space for vector representation, extracting global knowledge, including the relationships between entity nodes and the network topology. Finally, a predefined business process and multiple decision points corresponding to its execution are obtained. The execution results of the business process are evaluated using global knowledge and current business process variables to obtain guiding decisions, ultimately achieving optimized management of the predefined business process. This method integrates data processing capabilities, deep learning and intelligent algorithms, sophisticated business process automation tools, and business process optimization capabilities within a single low-code platform, achieving highly efficient data integration, in-depth data analysis and intelligent decision-making, flexible automated business process optimization, and cost-effective resource management.

[0096] like Figure 3 As shown, in one embodiment, the platform management method based on low-code development provided by the present invention obtains multiple heterogeneous data sources, extracts first data from the multiple heterogeneous data sources, and preprocesses the first data to obtain standard data for constructing a knowledge graph, specifically including the following steps:

[0097] Step S112 : extracting a plurality of first data sources from a plurality of heterogeneous data sources according to a set condition, and extracting a plurality of data columns that meet the set condition from the plurality of first data sources.

[0098] Step S114 , calling a fusion algorithm to integrate multiple data columns to obtain a data set, and merging different data sets according to a set key, where the data in the data set is the first data.

[0099] Step S116, replacing missing values in the first data, and deleting abnormal data in each data column whose values do not meet the set threshold, so as to clean the first data, and standardize the cleaned first data to obtain standard data.

[0100] like Figure 4 As shown, in one embodiment, the platform management method based on low-code development provided by the present invention obtains multiple different entities and entity relationships between different entities from standard data, and constructs a knowledge graph based on the multiple different entities and entity relationships between different entities, specifically including the following steps:

[0101] Step S122 : determining entity attributes, entity relationship types, and entity relationship weights based on a plurality of different entities and entity relationships between different entities.

[0102] Step S124: Based on entity attributes, entity relationship types, and entity relationship weights, a knowledge graph is constructed with multiple different entities as nodes and entity relationships between different entities as edges.

[0103] Among them, each entity corresponds to a node, and the entity relationship between each two different entities corresponds to an edge.

[0104] like Figure 5 As shown, in one embodiment, the platform management method based on low-code development provided by the present invention maps entity nodes in the knowledge graph and entity relationships between different entity nodes to a low-dimensional space for vector representation through a graph embedding model to extract global knowledge including entity relationships between different entity nodes and network topology structures, specifically including the following steps:

[0105] Step S132: Select a first entity node and a second entity node from the knowledge graph, and obtain an entity relationship weight between the first entity node and the second entity node. The first entity node and the second entity node are directly connected neighbor nodes and have an entity relationship.

[0106] In step S134 , the graph embedding model is called to map the first entity node, the second entity node, and the entity relationship between the first entity node and the second entity node into a vector representation in a low-dimensional space according to the entity relationship weight.

[0107] like Figure 6 As shown, in one embodiment, the platform management method based on low-code development provided by the present invention obtains a predefined business process and multiple decision points corresponding to the execution of the business process, and evaluates the execution results of the business process by executing global knowledge and current business process variables to obtain corresponding guiding decisions, which specifically includes the following steps:

[0108] Step S142 : executing a predefined business process according to the set business rules to obtain multiple decision points and determine the business environment variables when the business process is currently being executed.

[0109] Step S144 , obtaining different execution states and execution results by executing the business process under the global knowledge and business environment variables, and evaluating multiple decision points based on the different execution states and execution results to generate a guiding decision.

[0110] Step S146: updating and optimizing the predefined business process based on the guiding decision to generate a new business process.

[0111] like Figure 7 As shown, in one embodiment, the platform management method based on low-code development provided by the present invention further includes the following steps:

[0112] Step S710 , obtaining the current resource configuration corresponding to the predefined business process and the constraints satisfied by the resource configuration optimization, and calling the minimization cost function to perform cost minimization processing on the current resource configuration.

[0113] Step S720 , analyzing and evaluating historical data of the business process execution corresponding to the current resource configuration based on global knowledge, and generating an optimized resource configuration by combining the constraints and the minimization cost function.

[0114] The platform management system based on low-code development provided by the present invention is described below. The platform management system based on low-code development described below and the platform management method based on low-code development described above can be referenced to each other.

[0115] like Figure 8 As shown, in one embodiment, a platform management system based on low-code development includes a data preprocessing module 810, a knowledge graph construction module 820, a global knowledge extraction module 830 and an evaluation guidance module 840.

[0116] The data preprocessing module 810 is used to obtain multiple heterogeneous data sources, extract first data from the multiple heterogeneous data sources, and preprocess the first data to obtain standard data for constructing a knowledge graph. The preprocessing includes data cleaning and data standardization.

[0117] The knowledge graph construction module 820 is used to obtain multiple different entities and entity relationships between different entities from standard data, and to construct a knowledge graph based on the multiple different entities and entity relationships between different entities.

[0118] The global knowledge extraction module 830 is used to map the entity nodes in the knowledge graph and the entity relationships between different entity nodes to a low-dimensional space for vector representation through a graph embedding model, so as to extract global knowledge including the entity relationships between different entity nodes and the network topology structure.

[0119] The evaluation and guidance module 840 is used to obtain predefined business processes and multiple decision points corresponding to the execution of the business processes, and evaluate the execution results of the business processes by executing global knowledge and current business process variables to obtain corresponding guidance decisions.

[0120] Among them, the first data is useful data or key data that meets the set conditions, the standard data is composed of multiple different entities and entity relationships between different entities, and the current business process variable is the business environment variable when the predefined business process is currently executed, which guides decision-making for optimizing the management of the predefined business process.

[0121] In this embodiment, the data preprocessing module of the platform management system based on low-code development provided by the present invention is specifically used to:

[0122] A plurality of first data sources are extracted from a plurality of heterogeneous data sources according to set conditions, and a plurality of data columns that meet the set conditions are extracted from the plurality of first data sources.

[0123] A fusion algorithm is called to integrate multiple data columns to obtain a data set, and different data sets are merged according to a set key. The data in the data set is the first data.

[0124] In this embodiment, the data preprocessing module of the platform management system based on low-code development provided by the present invention is further used to:

[0125] The missing values in the first data are replaced, and abnormal data in each data column whose values do not meet the set threshold are deleted to clean the first data, and the cleaned first data are standardized to obtain standard data.

[0126] In this embodiment, the platform management system based on low-code development provided by the present invention, the knowledge graph construction module is specifically used to:

[0127] Entity attributes, entity relationship types, and entity relationship weights are determined based on multiple different entities and entity relationships between different entities.

[0128] Based on entity attributes, entity relationship types and entity relationship weights, a knowledge graph is constructed with multiple different entities as nodes and entity relationships between different entities as edges.

[0129] Among them, each entity corresponds to a node, and the entity relationship between each two different entities corresponds to an edge.

[0130] In this embodiment, the global knowledge extraction module of the platform management system based on low-code development provided by the present invention is specifically used to:

[0131] A first entity node and a second entity node are selected from a knowledge graph, and an entity relationship weight between the first entity node and the second entity node is obtained. The first entity node and the second entity node are directly connected neighbor nodes and have an entity relationship.

[0132] The call graph embedding model maps the first entity node, the second entity node, and the entity relationship between the first entity node and the second entity node to a vector representation in a low-dimensional space according to the entity relationship weight.

[0133] In this embodiment, the platform management system based on low-code development provided by the present invention, the evaluation guidance module is specifically used to:

[0134] Execute predefined business processes according to set business rules to obtain multiple decision points and determine the business environment variables when the business process is currently executing.

[0135] By executing business processes under global knowledge and business environment variables, different execution states and execution results are obtained, and multiple decision points are evaluated based on different execution states and execution results to generate guiding decisions.

[0136] Based on guided decisions, predefined business processes are updated and optimized to generate new business processes.

[0137] In this embodiment, the platform management system based on low-code development provided by the present invention further includes a resource configuration optimization module for:

[0138] Obtain the current resource configuration corresponding to the predefined business process and the constraints satisfied by the resource configuration optimization, and call the minimization cost function to minimize the cost of the current resource configuration.

[0139] Based on global knowledge, historical data of the business process execution corresponding to the current resource configuration is analyzed and evaluated, and the optimal resource configuration is generated by combining constraints and minimizing cost functions.

[0140] Figure 9 The following is a schematic diagram of the physical structure of an electronic device. The electronic device may be a smart terminal, and its internal structure diagram may be as follows: Figure 9 As shown. The electronic device includes a processor, a memory, and a network interface connected via a system bus. The processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the electronic device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a platform management method based on low-code development is implemented, which includes:

[0141] Acquire multiple heterogeneous data sources, extract first data from the multiple heterogeneous data sources, and preprocess the first data to obtain standard data for constructing a knowledge graph, where the preprocessing includes data cleaning and data standardization;

[0142] Obtain multiple different entities and entity relationships between different entities from standard data, and build a knowledge graph based on multiple different entities and entity relationships between different entities;

[0143] Through the graph embedding model, the entity nodes in the knowledge graph and the entity relationships between different entity nodes are mapped to a low-dimensional space for vector representation, so as to extract global knowledge including the entity relationships between different entity nodes and the network topology structure;

[0144] Obtain predefined business processes and multiple decision points corresponding to the execution of the business processes, and evaluate the execution results of the business processes by executing global knowledge and current business process variables to obtain corresponding guiding decisions;

[0145] Among them, the first data is useful data or key data that meets the set conditions, the standard data is composed of multiple different entities and entity relationships between different entities, and the current business process variable is the business environment variable when the predefined business process is currently executed, which guides decision-making for optimizing the management of the predefined business process.

[0146] Those skilled in the art will understand that Figure 9 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention, and does not constitute a limitation on the electronic device to which the solution of the present invention is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0147] On the other hand, the present invention further provides a computer storage medium storing a computer program, which, when executed by a processor, implements a platform management method based on low-code development, the method comprising:

[0148] Acquire multiple heterogeneous data sources, extract first data from the multiple heterogeneous data sources, and preprocess the first data to obtain standard data for constructing a knowledge graph, where the preprocessing includes data cleaning and data standardization;

[0149] Obtain multiple different entities and entity relationships between different entities from standard data, and build a knowledge graph based on multiple different entities and entity relationships between different entities;

[0150] Through the graph embedding model, the entity nodes in the knowledge graph and the entity relationships between different entity nodes are mapped to a low-dimensional space for vector representation, so as to extract global knowledge including the entity relationships between different entity nodes and the network topology structure;

[0151] Obtain predefined business processes and multiple decision points corresponding to the execution of the business processes, and evaluate the execution results of the business processes by executing global knowledge and current business process variables to obtain corresponding guiding decisions;

[0152] Among them, the first data is useful data or key data that meets the set conditions, the standard data is composed of multiple different entities and entity relationships between different entities, and the current business process variable is the business environment variable when the predefined business process is currently executed, which guides decision-making for optimizing the management of the predefined business process.

[0153] In another aspect, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and when the processor executes the computer instructions, implements a platform management method based on low-code development, the method comprising:

[0154] Acquire multiple heterogeneous data sources, extract first data from the multiple heterogeneous data sources, and preprocess the first data to obtain standard data for constructing a knowledge graph, where the preprocessing includes data cleaning and data standardization;

[0155] Obtain multiple different entities and entity relationships between different entities from standard data, and build a knowledge graph based on multiple different entities and entity relationships between different entities;

[0156] Through the graph embedding model, the entity nodes in the knowledge graph and the entity relationships between different entity nodes are mapped to a low-dimensional space for vector representation, so as to extract global knowledge including the entity relationships between different entity nodes and the network topology structure;

[0157] Obtain predefined business processes and multiple decision points corresponding to the execution of the business processes, and evaluate the execution results of the business processes by executing global knowledge and current business process variables to obtain corresponding guiding decisions;

[0158] Among them, the first data is useful data or key data that meets the set conditions, the standard data is composed of multiple different entities and entity relationships between different entities, and the current business process variable is the business environment variable when the predefined business process is currently executed, which guides decision-making for optimizing the management of the predefined business process.

[0159] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory.

[0160] By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0161] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0162] The above-described embodiments merely illustrate several implementations of the present invention, and while their descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make numerous variations and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A platform management method based on low-code development, characterized in that: The method comprises: Acquire multiple heterogeneous data sources, extract first data from the multiple heterogeneous data sources, and preprocess the first data to obtain standard data for constructing a knowledge graph, wherein the preprocessing includes data cleaning and data standardization; Acquire a plurality of different entities and entity relationships between different entities from the standard data, and construct a knowledge graph based on the plurality of different entities and entity relationships between different entities; Mapping the entity nodes and entity relationships between different entity nodes in the knowledge graph to a low-dimensional space for vector representation through a graph embedding model to extract global knowledge including the entity relationships between the different entity nodes and the network topology structure; Obtaining a predefined business process and a plurality of decision points corresponding to the execution of the business process, and evaluating the execution results of the business process by executing the global knowledge and current business process variables to obtain corresponding guiding decisions; The first data is useful data or key data that meets set conditions, the standard data is composed of multiple different entities and entity relationships between different entities, the current business process variables are business environment variables when the predefined business process is currently executed, and the guiding decision is used to optimize the management of the predefined business process; The acquiring of a plurality of different entities and entity relationships between different entities from the standard data, and constructing a knowledge graph based on the plurality of different entities and entity relationships between different entities, includes: Determine entity attributes, entity relationship types, and entity relationship weights based on the multiple different entities and the entity relationships between the different entities; Based on the entity attributes, entity relationship types and entity relationship weights, the knowledge graph is constructed with multiple different entities as nodes and the entity relationships between the different entities as edges; Among them, each entity corresponds to a node, and the entity relationship between each two different entities corresponds to an edge; The graph embedding model is used to map the entity nodes and the entity relationships between different entity nodes in the knowledge graph to a low-dimensional space for vector representation to extract global knowledge including the entity relationships between the different entity nodes and the network topology structure, including: Selecting a first entity node and a second entity node from the knowledge graph, and obtaining an entity relationship weight between the first entity node and the second entity node, where the first entity node and the second entity node are directly connected neighbor nodes and have an entity relationship; The graph embedding model is called to map the first entity node, the second entity node, and the entity relationship between the first entity node and the second entity node to a vector representation in a low-dimensional space according to the entity relationship weight.

2. The platform management method based on low-code development according to claim 1 is characterized in that: The acquiring of multiple heterogeneous data sources, extracting first data from the multiple heterogeneous data sources, and preprocessing the first data to obtain standard data for constructing a knowledge graph includes: Extracting a plurality of first data sources from the plurality of heterogeneous data sources according to the set conditions, and extracting a plurality of data columns that meet the set conditions from the plurality of first data sources; A fusion algorithm is called to integrate the multiple data columns to obtain a data set, and different data sets are merged according to a set key, and the data in the data set is the first data.

3. The platform management method based on low-code development according to claim 2 is characterized in that: The step of obtaining multiple heterogeneous data sources, extracting first data from the multiple heterogeneous data sources, and preprocessing the first data to obtain standard data for constructing a knowledge graph further includes: The missing values in the first data are replaced, and abnormal data in each data column whose values do not meet the set threshold are deleted to clean the first data, and the cleaned first data are standardized to obtain the standard data.

4. The platform management method based on low-code development according to claim 1 is characterized in that: The obtaining of a predefined business process and a plurality of decision points corresponding to the execution of the business process, and evaluating the execution result of the business process by executing the global knowledge and current business process variables to obtain corresponding guiding decisions, includes: Executing the predefined business process according to the set business rules to obtain the multiple decision points and determine the business environment variables when the business process is currently executed; Obtaining different execution states and execution results by executing the business process under the global knowledge and the business environment variables, and evaluating the multiple decision points based on the different execution states and execution results to generate the guiding decision; The predefined business process is updated and optimized based on the guiding decision to generate a new business process.

5. The platform management method based on low-code development according to any one of claims 1 to 4, characterized in that: The method further comprises: Obtaining the current resource configuration corresponding to the predefined business process and the constraints satisfied by the resource configuration optimization, and calling the minimization cost function to perform cost minimization processing on the current resource configuration; The historical data of the business process execution corresponding to the current resource configuration is analyzed and evaluated based on the global knowledge, and the optimized resource configuration is generated by combining the constraint conditions and the minimization cost function.

6. A platform management system based on low-code development, characterized in that: For implementing the platform management method based on low-code development according to any one of claims 1 to 5, the system comprises: A data preprocessing module, configured to obtain multiple heterogeneous data sources, extract first data from the multiple heterogeneous data sources, and preprocess the first data to obtain standard data for constructing a knowledge graph, wherein the preprocessing includes data cleaning and data standardization; A knowledge graph construction module, configured to obtain a plurality of different entities and entity relationships between different entities from the standard data, and to construct a knowledge graph based on the plurality of different entities and entity relationships between different entities; A global knowledge extraction module is used to map the entity nodes and the entity relationships between different entity nodes in the knowledge graph to a low-dimensional space for vector representation through a graph embedding model, so as to extract global knowledge including the entity relationships between the different entity nodes and the network topology structure; An evaluation and guidance module is used to obtain a predefined business process and a plurality of decision points corresponding to the execution of the business process, and evaluate the execution results of the business process by executing the global knowledge and current business process variables to obtain corresponding guidance decisions; Among them, the first data is useful data or key data that meets the set conditions, the standard data is composed of multiple different entities and entity relationships between different entities, the current business process variables are the business environment variables when the predefined business process is currently executed, and the guiding decision is used to optimize the management of the predefined business process.

7. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

8. A computer storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images