Management decision-making method and system based on knowledge base construction technology

Abstract

The invention discloses a management decision-making method and system based on a knowledge base construction technology. The method comprises the following steps: acquiring business demand information input by a target object; determining a target business type corresponding to the business demand information, and determining a target knowledge graph corresponding to the target business type from a preset knowledge base; performing semantic analysis and intention recognition on the business demand information by using a pre-trained demand analysis model to obtain a target demand feature corresponding to the business demand information; analyzing the target knowledge graph and the target demand features by using a graph neural network to obtain multi-dimensional decision suggestions for the service demand information; and feeding back the multi-dimensional decision suggestions to the target object. The technical problem that user experience is affected due to the fact that decision suggestions generated by a system in related business decision scenes are not highly matched with user requirements is solved.

Description

Technical Field

[0001] The present application relates to the field of financial services technology, and in particular, to a management decision-making method and system based on knowledge base construction technology. Background Art

[0002] In the business management of the financial field, with the increasing complexity of business scenarios and the personalized trend of user needs, decision support systems are facing unprecedented challenges. When dealing with user business needs, traditional systems mainly rely on text input and provide services through simple keyword matching or predefined query logic. However, with the development of technology, users' expectations of the system have far exceeded this scope. They are eager to express their needs in a more natural and convenient way, and require the system to provide more accurate and comprehensive decision suggestions. Although in recent years, some financial management decision systems have begun to try to introduce natural language processing and knowledge graph technology to improve service efficiency, these attempts still expose obvious deficiencies in actual applications and cannot meet the growing demand for intelligence in financial business management decisions.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiment of the present application provides a management decision-making method and system based on knowledge base construction technology, so as to at least solve the technical problem that the decision suggestions generated by the system in the relevant business decision-making scenarios are not highly matched with the user needs, thus affecting the user experience.

[0005] According to one aspect of an embodiment of the present application, a management decision-making method based on knowledge base construction technology is provided, including: obtaining business demand information input by a target object; determining a target business type corresponding to the business demand information, and determining a target knowledge graph corresponding to the target business type from a preset knowledge base; using a pre-trained demand analysis model to perform semantic analysis and intent recognition on the business demand information to obtain target demand features corresponding to the business demand information; using a graph neural network to analyze the target knowledge graph and the target demand features to obtain multi-dimensional decision recommendations for the business demand information; and feeding back multi-dimensional decision recommendations to the target object.

[0006] Optionally, obtaining business requirement information input by the target object includes: obtaining business requirement information in text mode input by the target object; or obtaining voice data input by the target object, recognizing the voice data using voice recognition technology, and obtaining business requirement information in text mode; or obtaining image data input by the target object, recognizing the image data using image recognition technology, and obtaining business requirement information in text mode.

[0007] Optionally, before determining the target knowledge graph corresponding to the target business type from a preset knowledge base, the method also includes: obtaining the identity information of the target object and determining whether the identity information has the authority to access the data of the target business type; if the identity information has the authority to access the data of the target business type, continuing to determine the target knowledge graph corresponding to the target business type from the preset knowledge base; if the identity information does not have the authority to access the data of the target business type, generating a decision failure prompt message, wherein the decision failure prompt message is used to prompt that the target object has insufficient authority.

[0008] Optionally, the knowledge base includes multiple knowledge graphs corresponding to multiple business types, and the knowledge graph generation process includes: for each business type, obtaining business knowledge corresponding to the business type, wherein the type of business knowledge includes at least one of the following: business rules and regulations, business operation procedures, market rules, historical business cases, and customer information; constructing multiple first knowledge entities based on the business knowledge, and determining the relationship between each first knowledge entity to obtain a knowledge graph corresponding to the business type; periodically obtaining new business knowledge corresponding to the business type, and updating the knowledge graph based on the new business knowledge.

[0009] Optionally, the training process of the demand analysis model includes: constructing an initial model, wherein the type of the initial model includes at least one of the following: a transformer bidirectional encoder representation model, a bidirectional long short-term memory network and a conditional random field combination model; obtaining multiple groups of historical demand texts, taking each group of historical demand texts as a training sample, and annotating the demand features corresponding to each group of historical demand texts as sample labels; using multiple training samples and corresponding sample labels to iteratively train the initial model to obtain a demand analysis model.

[0010] Optionally, before using the graph neural network to analyze the target knowledge graph and target demand characteristics, the method also includes: obtaining a target user portrait corresponding to the target object from a preset user portrait library, wherein the target user portrait includes at least one of the following: attribute information, historical behavior information, and preference information of the target object; constructing multiple second knowledge entities based on the target user portrait, and adding the multiple second knowledge entities to the target knowledge graph.

[0011] Optionally, a graph neural network is used to analyze the target knowledge graph and target demand characteristics to obtain multi-dimensional decision recommendations for business demand information, including: using a graph neural network to determine the target knowledge entity corresponding to the target demand characteristics in the target knowledge graph, and determining the associated knowledge entity associated with the target knowledge entity; generating multi-dimensional decision recommendations for business demand information based on the target knowledge entity, the associated knowledge entity, and the association between the target knowledge entity and the associated knowledge entity, wherein the dimensions of the decision recommendations include at least one of the following: compliance dimension, cost-effectiveness dimension, operating process dimension, personalized service dimension, technology application dimension, time dimension, resource allocation dimension, customer experience dimension, and business optimization dimension.

[0012] Optionally, multi-dimensional decision suggestions are fed back to the target object, including: using a pre-trained risk assessment model to analyze the decision suggestions in each dimension to obtain the risk score and corresponding risk points of the decision suggestions in each dimension; sorting the decision suggestions in each dimension from large to small according to the corresponding risk scores; and displaying the decision suggestions and corresponding risk points in each dimension to the target object based on the sorting results.

[0013] According to another aspect of an embodiment of the present application, a management decision-making device based on knowledge base construction technology is also provided, including: an acquisition module, used to acquire business demand information input by a target object; a graph determination module, used to determine the target business type corresponding to the business demand information, and determine the target knowledge graph corresponding to the target business type from a preset knowledge base; a demand analysis module, used to use a pre-trained demand analysis model to perform semantic analysis and intent recognition on the business demand information, and obtain target demand features corresponding to the business demand information; a decision module, used to use a graph neural network to analyze the target knowledge graph and target demand features, and obtain multi-dimensional decision recommendations for the business demand information; and a feedback module, used to feed back multi-dimensional decision recommendations to the target object.

[0014] According to another aspect of an embodiment of the present application, a computer program product is further provided, the computer program product comprising: a computer program, wherein when the computer program is executed by a processor, the management decision method based on the knowledge base construction technology is implemented.

[0015] According to another aspect of an embodiment of the present application, an electronic device is also provided, which includes: a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to execute the above-mentioned management decision method based on knowledge base construction technology through the computer program.

[0016] In the embodiment of the present application, by using the pre-trained demand analysis model, it is possible to perform deep semantic analysis and intent recognition on the business demand information input by the user, so as to more accurately understand the user's real needs and business scenarios, and provide a solid foundation for subsequent decision analysis; by determining the specific knowledge graph corresponding to the target business type, it is possible to effectively call and integrate relevant knowledge resources, build a knowledge framework required for business decisions, and enhance the logic and comprehensiveness of decisions; the application of graph neural networks can mine multi-level related information from the knowledge graph, and combine the target demand characteristics to generate multi-dimensional decision recommendations covering compliance, cost-effectiveness, operational process optimization, etc., thereby providing more comprehensive and in-depth decision assistance; feeding back these multi-dimensional decision recommendations to the target object can not only help to quickly respond to business needs, but also through a dynamic feedback mechanism, users can continuously adjust and optimize decision-making strategies, improve the flexibility and adaptability of decisions, and thus solve the technical problem that the decision recommendations generated by the system in the relevant business decision scenarios are not highly matched with user needs, affecting user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0018] Figure 1 It is a flowchart of an optional management decision-making method based on knowledge base construction technology according to an embodiment of the present application;

[0019] Figure 2 is a schematic diagram of the structure of an optional management decision system based on knowledge base construction technology according to an embodiment of the present application;

[0020] Figure 3 It is a schematic diagram of the structure of an optional electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0021] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0022] It should be noted that the terms "first", "second", etc. in the specification, claims and drawings of the present application are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0023] In order to better understand the embodiments of the present application, some nouns or terms that appear in the description of the embodiments of the present application are first translated and explained as follows:

[0024] The information collected in the embodiments of the present application is information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data comply with the relevant laws, regulations and standards of the relevant countries and regions, take necessary confidentiality measures, do not violate public order and good customs, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0025] Example 1

[0026] According to an embodiment of the present application, a management decision-making method based on knowledge base construction technology is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0027] Figure 1 is a flow chart of a management decision-making method based on knowledge base construction technology provided in an embodiment of the present application, such as Figure 1 As shown, the method comprises the following steps:

[0028] Step S102, obtaining business requirement information input by the target object;

[0029] Step S104, determining the target business type corresponding to the business demand information, and determining the target knowledge graph corresponding to the target business type from a preset knowledge base;

[0030] Step S106, using the pre-trained demand analysis model to perform semantic analysis and intent recognition on the business demand information to obtain target demand features corresponding to the business demand information;

[0031] Step S108, using a graph neural network to analyze the target knowledge graph and target demand characteristics to obtain multi-dimensional decision suggestions for business demand information;

[0032] Step S110, providing multi-dimensional decision suggestions to the target object.

[0033] The following describes the various steps of the management decision-making method based on knowledge base construction technology in combination with the specific implementation process.

[0034] The following steps can be performed on an application platform. The platform can adopt a microservice architecture design to decompose the platform into multiple independent service modules, such as intelligent services, smart operations, financial products, etc. Each module is deployed independently, and efficient collaboration between modules is achieved through standardized API (Application Programming Interface) interfaces and unified data interaction protocols. In addition, through the cross-system collaboration framework, the platform can be deeply integrated with the head office system, external service platforms, etc. to ensure real-time synchronization of information and optimal allocation of resources. The platform uses knowledge graphs and natural language processing technologies to achieve deep semantic association and dynamic integration of business knowledge, and can automatically capture information such as business rule adjustments and quickly integrate it into the knowledge graph to ensure real-time update and high coverage of knowledge. The platform integrates task decomposition technology, rule engines and real-time data analysis capabilities, and can dynamically plan and optimize processes for complex tasks to ensure efficient use of resources and accurate execution of business. Based on knowledge graphs and deep learning models, the platform can accurately analyze user needs and generate personalized recommendations. The platform maintains the timeliness and accuracy of the knowledge graph through a real-time update mechanism.

[0035] First, obtain the business requirement information input by the target object.

[0036] As an optional implementation, the process can be carried out in the following steps: obtaining business demand information in text mode input by the target object; or obtaining voice data input by the target object, recognizing the voice data using voice recognition technology, and obtaining business demand information in text mode; or obtaining image data input by the target object, recognizing the image data using image recognition technology, and obtaining business demand information in text mode.

[0037] For example, in banking management, the business needs of the target object usually revolve around personal or corporate financial services, which may include the following aspects:

[0038] Account information inquiry: Customers may need to inquire about account balance, transaction records, account status, historical bills and other information; Financial product consultation: Customers have consultation needs on interest rates, limits, terms, application conditions, risk assessment and other details of various financial products (such as deposits, loans, credit cards, funds, insurance, etc.); Business process guidance: When customers implement specific businesses, such as remittances, loan applications, opening company accounts, etc., they may need to understand detailed process steps, required documents, time expectations and other information; Regulatory interpretation: Customers may have questions about the laws and regulations involved (such as foreign exchange management, loan interest rate adjustments, etc.) and need clear and accurate interpretation; Problem solving and suggestions: When customers encounter problems during the use of services, they need quick feedback mechanisms and solutions, or need personalized suggestions when choosing financial products.

[0039] The target object can directly input business demand information through the platform's input box, or obtain voice data and use voice recognition technology to obtain business demand information in text mode. In some scenarios, the target object may need to express business needs by uploading pictures (such as insurance policies, reports, check images, etc.). The platform uses image recognition technology to parse the uploaded image data and extract the text information therein to obtain business demand information. For example, the image is pre-processed, including image enhancement, denoising, and tilt correction, and then the OCR (Optical Character Recognition) technology is used to recognize and extract the text in the image and convert the characters in the image into an editable text format. For complex images containing tables, charts, or specific symbols, the platform may also use deep learning models (such as convolutional neural networks or Transformers) to recognize and parse structured information, so as to more comprehensively understand the business demand information. For uploaded images, the platform can process images of various formats, including scanned copies, photos, etc., and has the ability to automatically adjust the image size, rotation angle, and enhance clarity.

[0040] To this end, the platform has designed a unified interface and data processing flow that can simultaneously process data input in three modalities: text, voice, and image. In the data fusion processing stage, the platform uses multimodal pre-trained models (such as multimodal Transformer) to perform feature extraction and representation learning on data of different modalities, ensuring that no matter how users input information, the platform can accurately understand their business needs.

[0041] After obtaining the business demand information, determine the target business type corresponding to the business demand information, and determine the target knowledge graph corresponding to the target business type from the preset knowledge base.

[0042] As an optional implementation, the knowledge base includes multiple knowledge graphs corresponding to multiple business types, and the knowledge graph generation process may include the following steps: for each business type, obtain the business knowledge corresponding to the business type, wherein the type of business knowledge includes at least one of the following: business rules and regulations, business operation procedures, market rules, historical business cases, and customer information; construct multiple first knowledge entities based on the business knowledge, and determine the relationship between each first knowledge entity to obtain the knowledge graph corresponding to the business type; periodically obtain new business knowledge corresponding to the business type, and update the knowledge graph based on the new business knowledge.

[0043] As an optional implementation, before determining the target knowledge graph corresponding to the target business type from a preset knowledge base, the following steps can also be taken to perform permission verification: obtain the identity information of the target object, and determine whether the identity information has the permission to access the data of the target business type; if the identity information has the permission to access the data of the target business type, continue to determine the target knowledge graph corresponding to the target business type from the preset knowledge base; if the identity information does not have the permission to access the data of the target business type, generate a decision failure prompt message, wherein the decision failure prompt message is used to prompt that the target object has insufficient authority.

[0044] For example, when the platform obtains the business demand information input by the user, it can parse the keywords and topics in the information through deep semantic analysis and intent recognition algorithms to determine the business type to which the business demand information belongs. Business types may include: account information query, financial product consultation, business process guidance, regulatory interpretation and other business types. For example, if the user asks about "personal loan interest rate", the platform will classify it into the business type of "financial product consultation". Once the business type is determined, the platform will search for the knowledge graph corresponding to the business type from the preset knowledge base. The knowledge graph is a structured database that contains entities, relationships and knowledge related to a specific business type. For example, the "financial product consultation" business type may correspond to a knowledge graph containing entities and relationships such as loan products, interest rates, application processes, and compliance requirements. The platform will periodically obtain new business knowledge related to the business type, including market trends, customer feedback and other information, and integrate this new knowledge into the corresponding knowledge graph to maintain the timeliness and accuracy of the knowledge graph. The platform can introduce a streaming data processing framework such as Apache Kafka or Apache Flink is used to continuously monitor and receive new information, automatically identify and integrate new knowledge through graph neural network models, realize dynamic maintenance of knowledge graphs, and ensure the timeliness and accuracy of information.

[0045] Before calling the knowledge graph of a specific business type, the platform will obtain the identity information of the target object (user) and verify whether the user has the authority to access the data of the specific business type. If the user does not have sufficient authority, the platform will generate a prompt message indicating insufficient authority to prevent unauthorized access. After verifying the authority, the platform calls the knowledge graph corresponding to the business type to provide users with accurate business guidance, intelligent recommendations or detailed operation procedures to ensure personalized and efficient services. For example, ordinary employees may only be able to access business knowledge related to their positions, while management can access more comprehensive business data and sensitive information. If the user does not have the authority to access the requested business type data, the system will generate a prompt message indicating insufficient authority to ensure data security and compliance. At the same time, the platform uses encrypted transmission and access control technology to protect the data in the knowledge graph from unauthorized access.

[0046] After obtaining the business demand information, the pre-trained demand analysis model is used to perform semantic analysis and intent recognition on the business demand information to obtain the target demand features corresponding to the business demand information.

[0047] As an optional implementation, the training process of the demand analysis model can take the following steps:

[0048] S1, constructing an initial model, wherein the type of the initial model includes at least one of the following: a transformer bidirectional encoder representation model, a bidirectional long short-term memory network and a conditional random field combination model;

[0049] S2, obtaining multiple groups of historical demand texts, taking each group of historical demand texts as a training sample, and annotating the demand features corresponding to each group of historical demand texts as sample labels;

[0050] S3, iteratively train the initial model using multiple training samples and corresponding sample labels to obtain a demand analysis model.

[0051] After obtaining the demand analysis model using the above method, the model can be used to analyze the business demand model.

[0052] In addition to the above steps, it may also include: obtaining a target user portrait corresponding to the target object from a preset user portrait library, wherein the target user portrait includes at least one of the following: attribute information, historical behavior information, and preference information of the target object; constructing multiple second knowledge entities based on the target user portrait, and adding the multiple second knowledge entities into the target knowledge graph.

[0053] Using the prepared multiple training samples and their demand feature annotations, the platform optimizes the initial model through an iterative training process to obtain a demand analysis model. During the training process, the model learns the ability to identify specific entities, parse semantic relationships, and understand user intent, so that it can accurately extract target demand features from new business demand information. For example, the model can understand the impact of keywords such as "interest rate adjustment" and "loan term" on user demand. For example, a user may query product information through a natural language expression such as "I want to know what the interest rate of XX product is?" The platform can accurately identify the keywords "interest rate" and "XX product", and understand that the user's intention is to query the interest rate information of a specific financial product, thereby providing accurate search results.

[0054] For example, the platform extracts user portrait data of the target object (user) from the preset user portrait library. This process includes but is not limited to the acquisition of the following information: Attribute information: including the user's basic attributes, such as age, gender, occupation, department, etc., and related attributes in branch business, such as account type, credit score, asset status, etc. This information provides the platform with a preliminary understanding of the user's background and demand characteristics; Historical behavior information: historical behavior data of users in branch business, such as transaction records, service requests, complaint records, consultation history, etc., to understand the user's usage habits and behavior patterns; Preference information: analyze user preferences for service methods, product types, operating interfaces, etc., to provide a basis for personalized services. Based on the target user portrait, the platform constructs multiple second knowledge entities, which represent knowledge nodes of user-specific attributes or preferences, including: attribute entity: based on the attribute information in the user portrait (such as occupation, credit score), the corresponding attribute entity is constructed to supplement the personalized information in the target knowledge graph; behavior entity: based on the user's historical behavior information, such as transaction frequency and operation preferences in a specific time period, behavior entity nodes are created to reflect the user's behavioral characteristics; preference entity: by analyzing user preference information, preference entities are constructed, such as service method preferences (online service or counter service), product preferences (time deposits or demand deposits), etc., which are used to construct personalized recommendation logic in the knowledge graph. These second knowledge entities not only enrich the content of the target knowledge graph, but also ensure that the knowledge graph can be dynamically adjusted and optimized according to user characteristics, and provide decision-making recommendations that are more in line with user needs. The platform integrates the constructed second knowledge entity into the target knowledge graph, and deeply integrates the user portrait information with static business knowledge through the embedded learning ability of the graph neural network. For example, for the "personal loan" business scenario, the system will not only consider entities such as regulations and market interest rates, but also optimize the loan approval process based on the attribute entity constructed based on the user's credit score; provide recommendations for online or counter loan services based on the user's preference entity; adjust the risk assessment strategy through the behavior entity constructed through the user's historical behavior. This personalized extension enables the knowledge graph to better adapt to user needs and provide customized decision-making recommendations. After integrating the target user portrait and the target knowledge graph, the platform generates decision-making recommendations based on user attributes, historical behaviors and preferences through the deep learning ability of the graph neural network. For example, in the "cash management" business, the system can analyze user account balances, transaction habits and preference information, and combine cash management rules and market trends to generate personalized cash management strategies, including fund allocation, investment recommendations and risk control measures. Through the integration of user portraits and knowledge graphs, decision-making recommendations not only cover general dimensions such as compliance, cost-effectiveness and operational procedures, but also pay special attention to personalized service dimensions and customer experience dimensions, providing users with more intimate and efficient business guidance and support.

[0055] While extracting demand features, the platform also obtains the user portrait of the target object (user) from the preset user portrait library. The user portrait contains user attribute information (such as position, department), historical behavior information (such as previously proposed demand types, operation preferences) and preference information (such as preferred service methods, business areas of concern). Based on the target user portrait, the platform constructs multiple second knowledge entities, such as user preference nodes and historical behavior nodes. These entities further enrich the target knowledge graph, enabling it to more accurately reflect the personalized characteristics of user needs. The platform integrates user-specific information into the target knowledge graph by associating the second knowledge entity (user portrait information) with the first knowledge entity (business rules and regulations, operation procedures, etc.). This personalized extension enables the knowledge graph to provide more tailored service recommendations and operation guidelines based on user attributes and historical behaviors. For example, the platform can provide optimized loan interest rate consultation or personalized repayment plan recommendations based on the user's loan history.

[0056] After obtaining the above-mentioned target demand characteristics and target knowledge graph, the target knowledge graph and target demand characteristics are analyzed using the graph neural network to obtain multi-dimensional decision suggestions for business demand information. This process can be carried out in the following steps:

[0057] A graph neural network is used to determine the target knowledge entity corresponding to the target demand characteristics in the target knowledge graph, and to determine the associated knowledge entity associated with the target knowledge entity; based on the target knowledge entity, the associated knowledge entity, and the association relationship between the target knowledge entity and the associated knowledge entity, multi-dimensional decision recommendations for business demand information are generated, wherein the dimensions of the decision recommendations include at least one of the following: compliance dimension, cost-effectiveness dimension, operation process dimension, personalized service dimension, technology application dimension, time dimension, resource allocation dimension, customer experience dimension, and business optimization dimension.

[0058] For example, if the target demand feature is "personal loan interest rate consultation", the platform will locate the "personal loan" and "interest rate" related entities in the knowledge graph. Next, based on the target knowledge entity, through relationship chain analysis, it will determine other knowledge entities that are directly or indirectly related to it, that is, related knowledge entities. Related entities may include market laws, historical cases, compliance requirements, etc. related to loan products. These entities together constitute a comprehensive knowledge base for decision-making recommendations.

[0059] Compliance dimension: Based on the compliance requirement entity, analyze whether the business needs meet the relevant laws and regulations, generate compliance assessment suggestions, and ensure the legality of business operations. For example, in loan approval, decision-making suggestions need to consider the latest credit regulations to ensure that business operations are carried out within the legal and compliance framework;

[0060] Cost-benefit dimension: Through the cost-benefit analysis entity, the system evaluates the economic benefits achieved by business needs, provides cost optimization suggestions, and helps users or decision makers make wise choices at the financial level. For example, in the guarantee business, the decision suggestions may include cost estimation, interest rate discount suggestions, and economic benefit analysis for both the bank and the customer.

[0061] Operation process dimension: Based on the operation process entity, the system generates detailed process guidance and optimization suggestions to ensure the efficiency and standardization of business execution. Decision suggestions may include standardized operation steps, process templates, automated processing tools, etc. to support efficient business execution;

[0062] Personalized service dimension: Based on the user profile of the target object, the system provides personalized service solutions to meet the specific needs of users and improve user experience. For users who have a preference for specific financial products, decision suggestions may include customized product recommendations or service customization solutions;

[0063] Technology application dimension: Analyze the relationship between technology application entities and business needs, and systematically recommend suitable technology tools and solutions to promote business innovation and digital transformation. This dimension may involve RPA (Robotic Process Automation), AI (Artificial Intelligence) algorithm calls, and recommendations for data analysis tools.

[0064] Time dimension: Considering time-sensitive entities, such as deadlines, periodic requirements, etc., the system generates time management suggestions to ensure that business needs are effectively responded to within the appropriate time frame. For example, in foreign exchange trading, decision suggestions may include suggestions for the best trading time based on market forecasts;

[0065] Resource allocation dimension: Based on resource entities, the system evaluates resource availability and limitations, and provides suggestions for optimizing resource allocation to ensure efficient execution of business needs. For example, in high-frequency transaction processing scenarios, decision suggestions may include personnel scheduling, equipment configuration optimization, task priority adjustment, etc.

[0066] Customer experience dimension: Through the customer experience entity, the system analyzes the impact of business needs on customer satisfaction and proposes strategies to improve customer experience. For example, in customer service decisions, it may include suggestions such as increasing waiting time, optimizing service processes, and providing additional support;

[0067] Business optimization dimension: Combined with the business optimization entity, the system provides process improvement and business development suggestions to promote continuous business optimization and innovation. For example, decision-making suggestions may include business process reengineering, product innovation suggestions or market development strategies.

[0068] After obtaining multi-dimensional decision suggestions, the multi-dimensional decision suggestions are fed back to the target object.

[0069] As an optional implementation, the process can be carried out in the following steps: use a pre-trained risk assessment model to analyze the decision recommendations of each dimension to obtain the risk score and corresponding risk points of the decision recommendations of each dimension; sort the decision recommendations of each dimension from large to small according to the corresponding risk scores; and display the decision recommendations of each dimension and the corresponding risk points to the target object based on the sorting results.

[0070] In addition, the platform can adjust and optimize the user portrait library and the second knowledge entity in real time through user feedback on the use of decision suggestions, ensuring that the knowledge graph can continue to reflect the user's latest needs and behavior patterns. These feedback information is used for model iteration and dynamic updating of the knowledge graph. For example, the user's adoption and effectiveness evaluation of a cash management strategy can be used to fine-tune the parameters of the graph neural network and optimize the construction logic of the user portrait, thereby providing more accurate and effective suggestions in future decision-making. This efficient feedback mechanism enables the platform to continuously learn and adapt to user needs, and continuously improve the personalization and intelligence of decision suggestions.

[0071] In the embodiment of the present application, by using the pre-trained demand analysis model, it is possible to perform deep semantic analysis and intent recognition on the business demand information input by the user, so as to more accurately understand the user's real needs and business scenarios, and provide a solid foundation for subsequent decision analysis; by determining the specific knowledge graph corresponding to the target business type, it is possible to effectively call and integrate relevant knowledge resources, build a knowledge framework required for business decisions, and enhance the logic and comprehensiveness of decisions; the application of graph neural networks can mine multi-level related information from the knowledge graph, and combine the target demand characteristics to generate multi-dimensional decision recommendations covering compliance, cost-effectiveness, operational process optimization, etc., thereby providing more comprehensive and in-depth decision assistance; feeding back these multi-dimensional decision recommendations to the target object can not only help to quickly respond to business needs, but also through a dynamic feedback mechanism, users can continuously adjust and optimize decision-making strategies, improve the flexibility and adaptability of decisions, and thus solve the technical problem that the decision recommendations generated by the system in the relevant business decision scenarios are not highly matched with user needs, affecting user experience.

[0072] Example 2

[0073] According to the embodiment of the present application, a management decision system based on knowledge base construction technology is also provided for implementing the management decision method based on knowledge base construction technology in embodiment 1, such as Figure 2As shown, the management decision system based on the knowledge base construction technology at least includes: an acquisition module 21, a graph determination module 22, a demand analysis module 23, a decision module 24 and a feedback module 25, wherein:

[0074] An acquisition module 21 is used to acquire business demand information input by a target object;

[0075] A graph determination module 22, used to determine a target business type corresponding to the business demand information, and to determine a target knowledge graph corresponding to the target business type from a preset knowledge base;

[0076] The demand analysis module 23 is used to perform semantic analysis and intent recognition on the business demand information using a pre-trained demand analysis model to obtain target demand features corresponding to the business demand information;

[0077] The decision module 24 is used to analyze the target knowledge graph and target demand characteristics using a graph neural network to obtain multi-dimensional decision suggestions for business demand information;

[0078] The feedback module 25 is used to provide multi-dimensional decision suggestions to the target object.

[0079] The following describes the functions of each module of the management decision-making system based on knowledge base construction technology in conjunction with the specific implementation process.

[0080] First, the acquisition module obtains the business requirement information input by the target object.

[0081] As an optional implementation, the process can be carried out in the following steps: obtaining business demand information in text mode input by the target object; or obtaining voice data input by the target object, recognizing the voice data using voice recognition technology, and obtaining business demand information in text mode; or obtaining image data input by the target object, recognizing the image data using image recognition technology, and obtaining business demand information in text mode.

[0082] After obtaining the business demand information, the graph determination module determines the target business type corresponding to the business demand information, and determines the target knowledge graph corresponding to the target business type from a preset knowledge base.

[0083] As an optional implementation, the knowledge base includes multiple knowledge graphs corresponding to multiple business types, and the knowledge graph generation process includes: for each business type, obtaining business knowledge corresponding to the business type, wherein the type of business knowledge includes at least one of the following: business rules and regulations, business operation procedures, market rules, historical business cases, and customer information; constructing multiple first knowledge entities based on the business knowledge, and determining the relationship between each first knowledge entity to obtain a knowledge graph corresponding to the business type; periodically obtaining new business knowledge corresponding to the business type, and updating the knowledge graph based on the new business knowledge.

[0084] As an optional implementation, before determining the target knowledge graph corresponding to the target business type from a preset knowledge base, it also includes: obtaining the identity information of the target object and determining whether the identity information has the authority to access the data of the target business type; if the identity information has the authority to access the data of the target business type, continuing to determine the target knowledge graph corresponding to the target business type from the preset knowledge base; if the identity information does not have the authority to access the data of the target business type, generating a decision failure prompt message, wherein the decision failure prompt message is used to prompt that the target object has insufficient authority.

[0085] After obtaining the business demand information, the demand analysis module uses the pre-trained demand analysis model to perform semantic analysis and intent recognition on the business demand information to obtain the target demand features corresponding to the business demand information.

[0086] As an optional implementation, the training process of the demand analysis model can take the following steps:

[0087] S1, constructing an initial model, wherein the type of the initial model includes at least one of the following: a transformer bidirectional encoder representation model, a bidirectional long short-term memory network and a conditional random field combination model;

[0088] S2, obtaining multiple groups of historical demand texts, taking each group of historical demand texts as a training sample, and annotating the demand features corresponding to each group of historical demand texts as sample labels;

[0089] S3, iteratively train the initial model using multiple training samples and corresponding sample labels to obtain a demand analysis model.

[0090] In addition to the above steps, it may also include: obtaining a target user portrait corresponding to the target object from a preset user portrait library, wherein the target user portrait includes at least one of the following: attribute information, historical behavior information, and preference information of the target object; constructing multiple second knowledge entities based on the target user portrait, and adding the multiple second knowledge entities into the target knowledge graph.

[0091] After obtaining the above-mentioned target demand characteristics and target knowledge graph, the decision module uses the graph neural network to analyze the target knowledge graph and target demand characteristics to obtain multi-dimensional decision recommendations for business demand information. This process can be carried out in the following steps:

[0092] A graph neural network is used to determine the target knowledge entity corresponding to the target demand characteristics in the target knowledge graph, and to determine the associated knowledge entity associated with the target knowledge entity; based on the target knowledge entity, the associated knowledge entity, and the association relationship between the target knowledge entity and the associated knowledge entity, multi-dimensional decision recommendations for business demand information are generated, wherein the dimensions of the decision recommendations include at least one of the following: compliance dimension, cost-effectiveness dimension, operation process dimension, personalized service dimension, technology application dimension, time dimension, resource allocation dimension, customer experience dimension, and business optimization dimension.

[0093] After obtaining the multi-dimensional decision suggestions, the feedback module feeds back the multi-dimensional decision suggestions to the target object.

[0094] As an optional implementation, the process can be carried out in the following steps: use a pre-trained risk assessment model to analyze the decision recommendations of each dimension to obtain the risk score and corresponding risk points of the decision recommendations of each dimension; sort the decision recommendations of each dimension from large to small according to the corresponding risk scores; and display the decision recommendations of each dimension and the corresponding risk points to the target object based on the sorting results.

[0095] It should be noted that each module in the management decision system based on knowledge base construction technology in the embodiment of the present application corresponds one by one to each implementation step of the management decision method based on knowledge base construction technology in Example 1. Since a detailed description has been given in Example 1, some details not reflected in this embodiment can be referred to Example 1, and will not be repeated here.

[0096] Example 3

[0097] According to an embodiment of the present application, a computer program product is also provided, which includes a computer program, wherein when the computer program is executed by a processor, the management decision-making method based on the knowledge base construction technology in Example 1 is implemented.

[0098] According to an embodiment of the present application, a non-volatile storage medium is also provided, which includes a stored computer program, wherein the device where the non-volatile storage medium is located executes the management decision-making method based on the knowledge base construction technology in Example 1 by running the computer program.

[0099] According to an embodiment of the present application, a processor is also provided, which is used to run a computer program, wherein the management decision-making method based on the knowledge base construction technology in Example 1 is executed when the computer program is running.

[0100] According to an embodiment of the present application, an electronic device is also provided, which includes: a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to execute the management decision-making method based on the knowledge base construction technology in Example 1 through the computer program.

[0101] Specifically, the computer program executes the following steps when it is running: obtaining business requirement information input by the target object; determining the target business type corresponding to the business requirement information, and determining the target knowledge graph corresponding to the target business type from a preset knowledge base; using a pre-trained requirement analysis model to perform semantic analysis and intent recognition on the business requirement information to obtain target requirement features corresponding to the business requirement information; using a graph neural network to analyze the target knowledge graph and target requirement features to obtain multi-dimensional decision recommendations for the business requirement information; and providing feedback of multi-dimensional decision recommendations to the target object.

[0102] As an optional implementation, the electronic device may be in the form of a mobile terminal, a computer terminal or a similar computing device. Figure 3 FIG. 1 shows a hardware structure block diagram of an electronic device for implementing a management decision-making method based on knowledge base construction technology. Figure 3 As shown, the electronic device 30 may include one or more (302a, 302b, ..., 302n are used to illustrate) processors 302 (the processor 302 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 304 for storing data, and a transmission device 306 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 3 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 3 More or fewer components as shown, or with Figure 3 Different configurations are shown.

[0103] It should be noted that the one or more processors 302 and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the electronic device 30. As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0104] The memory 304 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the management decision method based on the knowledge base construction technology in the embodiment of the present application. The processor 302 executes various functional applications and data processing by running the software programs and modules stored in the memory 304, that is, realizing the vulnerability detection method of the above-mentioned application. The memory 304 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 304 may further include a memory remotely arranged relative to the processor 302, and these remote memories may be connected to the electronic device 30 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0105] The transmission device 306 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the electronic device 30. In one example, the transmission device 306 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 306 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0106] The display may be, for example, a touch screen liquid crystal display (LCD) that may enable a user to interact with a user interface of the electronic device 30 .

[0107] The serial numbers of the above embodiments are only for description and do not represent the advantages or disadvantages of the embodiments.

[0108] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0109] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0110] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed over multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0111] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0112] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0113] The above are only preferred implementations of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A management decision-making method based on knowledge base construction technology, characterized in that: include: Obtain business requirement information input by the target object; Determine the target business type corresponding to the business demand information, and determine the target knowledge graph corresponding to the target business type from a preset knowledge base; Using a pre-trained demand analysis model to perform semantic analysis and intent recognition on the business demand information, and obtain target demand features corresponding to the business demand information; Analyze the target knowledge graph and the target demand characteristics using a graph neural network to obtain multi-dimensional decision suggestions for the business demand information; The multi-dimensional decision suggestions are fed back to the target object.

2. The method according to claim 1, characterized in that Obtain business requirement information input by the target object, including: Acquire the business requirement information of the text modality input by the target object; or, Acquire the voice data input by the target object, recognize the voice data using voice recognition technology, and obtain the business demand information in text mode; or, The image data input by the target object is obtained, and the image data is recognized using image recognition technology to obtain business demand information in text mode.

3. The method according to claim 1, characterized in that Before determining the target knowledge graph corresponding to the target business type from a preset knowledge base, the method further includes: Obtaining the identity information of the target object, and determining whether the identity information has the authority to access the data of the target business type; In the case where the identity information has the authority to access the data of the target business type, continue to determine the target knowledge graph corresponding to the target business type from the preset knowledge base; In the case that the identity information does not have the authority to access the data of the target business type, a decision failure prompt message is generated, wherein the decision failure prompt message is used to prompt that the target object has insufficient authority.

4. The method according to claim 1, characterized in that: The knowledge base includes multiple knowledge graphs corresponding to multiple business types, and the generation process of the knowledge graph includes: For each business type, obtain business knowledge corresponding to the business type, wherein the types of business knowledge include at least one of the following: business rules and regulations, business operation procedures, market rules, historical business cases, and customer information; Constructing a plurality of first knowledge entities according to the business knowledge, and determining the association relationship between each of the first knowledge entities, to obtain a knowledge graph corresponding to the business type; New business knowledge corresponding to the business type is periodically acquired, and the knowledge graph is updated according to the new business knowledge.

5. The method according to claim 1, characterized in that The training process of the demand analysis model includes: Constructing an initial model, wherein the type of the initial model includes at least one of the following: a transformer bidirectional encoder representation model, a bidirectional long short-term memory network and a conditional random field combination model; Acquire multiple groups of historical demand texts, use each group of the historical demand texts as a training sample, and annotate the demand features corresponding to each group of the historical demand texts as sample labels; The initial model is iteratively trained using a plurality of the training samples and corresponding sample labels to obtain the demand analysis model.

6. The method according to claim 4, characterized in that Before analyzing the target knowledge graph and the target demand characteristics by using the graph neural network, the method further includes: Acquire a target user portrait corresponding to the target object from a preset user portrait library, wherein the target user portrait includes at least one of the following: attribute information, historical behavior information, and preference information of the target object; Construct multiple second knowledge entities based on the target user portrait, and add the multiple second knowledge entities into the target knowledge graph.

7. The method according to claim 1, characterized in that The target knowledge graph and the target demand characteristics are analyzed using a graph neural network to obtain multi-dimensional decision suggestions for the business demand information, including: Determine a target knowledge entity corresponding to the target demand feature in the target knowledge graph by using a graph neural network, and determine an associated knowledge entity associated with the target knowledge entity; A multi-dimensional decision suggestion for the business demand information is generated based on the target knowledge entity, the associated knowledge entity, and the association relationship between the target knowledge entity and the associated knowledge entity, wherein the dimensions of the decision suggestion include at least one of the following: compliance dimension, cost-effectiveness dimension, operation process dimension, personalized service dimension, technology application dimension, time dimension, resource allocation dimension, customer experience dimension, and business optimization dimension.

8. The method according to claim 1, characterized in that: Feedback of the multi-dimensional decision suggestions to the target object includes: Use the pre-trained risk assessment model to analyze the decision suggestions for each dimension and obtain the risk score and corresponding risk points of the decision suggestions for each dimension; Sort the decision suggestions for each dimension from large to small according to the corresponding risk scores; Based on the sorting results, decision suggestions and corresponding risk points in each dimension are displayed to the target object.

9. A management decision system based on knowledge base construction technology, characterized in that: include: The acquisition module is used to obtain the business requirement information input by the target object; A graph determination module, used to determine the target business type corresponding to the business demand information, and determine a target knowledge graph corresponding to the target business type from a preset knowledge base; A demand analysis module is used to perform semantic analysis and intent recognition on the business demand information using a pre-trained demand analysis model to obtain target demand features corresponding to the business demand information; A decision-making module, used to analyze the target knowledge graph and the target demand characteristics by using a graph neural network to obtain multi-dimensional decision suggestions for the business demand information; A feedback module is used to provide the target object with the multi-dimensional decision suggestions.

10. An electronic device, characterized in that: include: A memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the management decision-making method based on the knowledge base construction technology as described in any one of claims 1 to 8 through the computer program.

Images