An intelligent interactive platform serving enterprise policy decision-making

Abstract

The application belongs to the technical field of intelligent enterprise policy decision-making, and discloses an intelligent interaction platform for enterprise policy decision-making. The policy and supporting resource acquisition unit obtains the original text and interpretation of provincial and municipal policies through the government-enterprise collaborative docking module. The cross-level rule crawler module captures the supporting rules of cities and counties and policy conflict processing guidelines. The dynamic updating and scheduling module ensures that the policy is updated within 10 minutes and avoids missing key information. The analysis and disassembly unit identifies policy mandatory requirements, core indicators and other elements through the multi-level semantic analysis module. The execution path mapping module uses a bidirectional mapping algorithm to convert policy quantitative indicators into time parameters and quality standards of business actions. The cross-policy conflict identification module can also generate a mediation plan. The combination of the two makes the policy understandable and executable, solves the problem of policy and business disconnection, and reduces the failure of reporting due to no clear path or missing rules.

Description

Technical Field

[0001] This invention belongs to the field of intelligent technology for enterprise policy decision-making, specifically an intelligent interactive platform that serves enterprise policy decision-making. Background Technology

[0002] In enterprise operation and development, policy compliance and the ability to convert policy dividends have become key factors affecting cost control, market competitiveness, and long-term development. As the policy systems of governments at all levels become increasingly complex, existing technological systems are struggling to support closed-loop management from policy acquisition to implementation, as detailed below: Currently, most mainstream policy service platforms only provide the function of extracting and displaying policy texts, lacking the ability to structurally break down the requirements of macro policies. They are unable to translate policy requirements into specific business actions that enterprises can implement, resulting in enterprises being "unable to understand and unable to act" when faced with policies. Some enterprises, due to a lack of clear implementation paths, have spent several months and still missed the policy application window. Existing technologies have not been able to achieve cross-level policy interpretation, and enterprises are prone to overlooking key information such as local subsidy details and regional special requirements, which leads to problems such as inconsistent application materials and failure to pass compliance review.

[0003] Existing policy risk assessment technologies largely rely on matching static data such as enterprise registration information and historical compliance data with policies, failing to anticipate potential policy risks by combining them with dynamic business plans. For example, some electronics companies fail to consider regional energy consumption quota policies and project construction cycles when planning new production lines, resulting in forced shutdowns after the production lines are completed due to excessive energy consumption. Existing technologies do not consider the compatibility between policy buffer periods and enterprise rectification cycles, making it impossible to deduce a reasonable rectification start time, often leading to situations where enterprises trigger penalties for violations due to overdue rectification.

[0004] Currently, policy service platforms only provide policy advice. On the one hand, when enterprises apply for policy subsidies and implement compliance rectification, the platform cannot track the review progress and rectification compliance in real time, making it difficult for enterprises to respond to problems in the implementation process in a timely manner. On the other hand, when enterprises face similar policy decisions in the future, they need to repeatedly review the process and summarize their experience again, which is not only inefficient, but also prone to repeating mistakes due to failure to reuse historical lessons, and cannot form a continuous optimization of policy decision-making capabilities. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent interactive platform for enterprise policy decision-making, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an intelligent interactive platform for enterprise policy decision-making, the platform comprising: Policy and Supporting Resource Acquisition Unit: Used to acquire industry policy information and regional policy coordination data, and to classify and store implementation cases, resource information, and policy-business mapping template data; Analysis and decomposition unit: used to identify policy elements and rules, execute the mapping template of the policy and business, generate exclusive execution paths, and generate mediation plans and mediation basis by combining historical execution data; Business Scenario Mapping Unit: Used to build enterprise profiles, store core enterprise information and label business characteristics, collect production and supply chain data, model according to industry scenarios and perform tagging processing, and associate enterprise plans with the regional policy coordination data; Matching and Risk Prediction Unit: Based on enterprise profiles, calculate the early warning window period and matching degree, construct a three-dimensional risk transmission chain and quantify the transmission probability, thereby predicting risks, identifying policy dividends and marking deadlines and material lists; Risk Intervention and Implementation Guidance Unit: Develop intervention strategies based on risk levels, output a list of related actions, resources and standards, and integrate and categorize policy-specific resource pools; Decision-making interaction unit: Establish objective function, form multiple decision-making schemes for comparison, obtain review progress through government affairs platform, set progress warning, compare enterprise indicators before and after implementation, form experience base and establish decision optimization knowledge base; Data Security and Operations Unit: Processes enterprise data according to data sensitivity, establishes an attribute and permission mapping rule base and generates audit logs, encrypts and verifies the integrity of interactive data, monitors system operating status and performance, and automatically switches to a backup node when the primary node fails.

[0007] Preferably, the policy and supporting resource acquisition unit are as follows: The government-enterprise collaboration module connects with provincial and municipal government service platforms to obtain original policy texts, official interpretations, and industry policy guidance. When the cross-level detailed rules crawler module is specifically crawling the supporting implementation rules of cities and counties, it simultaneously collects regional policy coordination data on local policy conflict handling guidelines. The Case Studies and Resource Library module stores implementation cases, a directory of compliance service agencies, and equipment supplier information, categorized by industry, policy type, and implementation effect. It also stores policy-business mapping template data. The dynamic update scheduling module triggers synchronization within 10 minutes when policies and supporting resources are updated. The case library is updated weekly. When the frequency of keywords in the policy text exceeds the threshold, the template node generation logic is automatically triggered to supplement the corresponding business nodes for qualification review and process verification.

[0008] Preferably, the analysis and disassembly unit is specifically as follows: The multi-level semantic analysis module identifies mandatory and optional requirements, core and auxiliary indicators, application and rectification nodes in policies, and extracts policy priority rules and industry compliance practices. The execution path mapping module calls the industry policy and business mapping template stored in the policy and supporting resource collection unit, combines the business process modeling symbols to generate a unique execution path, and adopts a two-way mapping algorithm between policy indicators and business actions to transform the quantitative indicators in the policy into time parameters and quality standards for business actions. The cross-policy conflict identification module, when identifying policy conflicts, combines the company's past execution data to generate a mediation plan and output the policy basis. The conflict identification first screens through a preset policy conflict rule library, then calculates the semantic similarity of the conflict clauses, and after determining the conflict type, calls the solution template in the historical execution data.

[0009] Preferably, the business scenario mapping unit is specifically as follows: Construct a comprehensive enterprise profile that includes basic information, operational dynamics, supply chain connections, regional adaptation, and planning. The basic information module stores core information such as enterprise registration, industry classification, and qualification certificates, while also labeling business characteristics by sub-industry. The operational dynamics acquisition module is used to acquire production capacity and energy consumption data, and also collects data from upstream and downstream of the supply chain. After real-time data cleaning, structured storage and field-level indexing, the data is transmitted to the matching and risk prediction unit. The business scenario modeling module models scenarios by industry segmentation. First, the core business of the industry is broken down into indivisible basic links. Then, each link is associated with corresponding compliance indicators and production indicators. Finally, it is mapped to the actual responsible departments of the enterprise, forming tagged data of link-indicator-responsible department. The future planning access module is used to acquire structured future business adjustment data, parse out core elements including changes in production capacity, new links, and resource requirements, and then perform field-level matching with the regional policy coordination data to generate a structured labeled dataset containing comparisons between the present and the future.

[0010] Preferably, the matching and risk prediction unit is specifically as follows: Based on the enterprise panoramic profile constructed by the business scenario mapping unit, the dynamic matching algorithm module associates policy execution nodes with enterprise business plan nodes based on temporal similarity, establishes a rectification cycle prediction function based on the enterprise's historical rectification data, takes into account parameters including policy buffer period and enterprise business adjustment complexity, and outputs the optimal rectification start time. The risk prediction model module constructs a three-dimensional risk transmission chain of enterprises, supply chains, and regions. It uses a risk factor transmission algorithm to quantify the probability of risk transmission at each stage. By analyzing the impact coefficient of enterprise business adjustments on supply chain links and the correlation between supply chain link anomalies and regional policy compliance, it calculates the overall probability of risk transmission from enterprises to regions. At the same time, it introduces a time-series attention mechanism to assign dynamic weights to core indicators within 12 months by calculating the contribution of the indicators in historical risk events. The dividend opportunity mining module identifies policy dividends that can be matched with enterprise plans. It uses a policy dividend keyword matching algorithm to extract dividend elements from policy texts and performs field-level matching with enterprise plans to filter out suitable dividend opportunities. At the same time, it extracts key information from the policy database to form a structured dividend opportunity dataset.

[0011] Preferably, the risk intervention and implementation guidance unit is specifically as follows: Based on the risk prediction results of the matching and risk prediction unit, the graded intervention strategy module formulates measures according to the risk level. For low risk, it outputs rectification suggestion documents; for medium risk, it generates a phased rectification plan and sets progress reminders; and for high risk, it assigns a dedicated compliance consultant. The execution guidance generation module outputs a list of associated actions, resources, and standards. Based on the policy area to which the risk belongs and the industry attributes of the enterprise, it selects the top 3 matching resources from the resource pool. At the same time, it automatically associates the corresponding standard number and technical requirements in the standard database to clarify the quality standard for each rectification action. The third-party resource collaboration module integrates a policy-specific resource pool, categorizes resources by policy type, industry, and service capabilities, and builds a resource retrieval index; it supports enterprises to operate online throughout the entire process of resource filtering, online reservation, and progress inquiry.

[0012] Preferably, the decision-making interaction unit is specifically as follows: Based on the execution progress data of the aforementioned risk intervention and execution guidance unit, the intelligent decision generation module associates the entire policy cycle and calibrates the key time points of the entire policy cycle with the enterprise's execution nodes on a timeline. At the same time, it adopts a multi-objective optimization algorithm with subsidy amount, application cycle, and success rate as optimization objectives, and constructs an objective function by combining historical application data to form 3-5 sets of decision schemes for comparison. The execution progress tracking module obtains the review progress and transforms the unstructured progress information returned by the government affairs platform into a standardized status of material submission - under review - review result - public announcement - payment received; at the same time, it sets progress warning rules to automatically trigger reminders when the time remaining before the application deadline or the review feedback time is less than a preset threshold. The effect evaluation and feedback module compares enterprise indicators before and after implementation, calculates the rate of change of enterprise compliance indicators and operational indicators after implementation, conducts deviation analysis with the target values ​​required by the policy, and generates an effect evaluation report; it uses an experience extraction algorithm to mine key operations of successful cases and core issues of failed cases from historical decision-making implementation data to form a structured experience library; The solution iteration and optimization module establishes a decision optimization knowledge base based on the experience base and adopts a case similarity matching algorithm. When an enterprise initiates a similar decision request, it retrieves the lessons learned from similar cases in the knowledge base and automatically pushes optimization suggestions. It supports manual annotation of new information, which is synchronized to the knowledge base. At the same time, the experience data is transmitted to the parsing and decomposition unit and the matching and risk prediction unit.

[0013] Preferably, the data security and operation and maintenance unit is specifically as follows: The hierarchical data desensitization module processes enterprise data according to data sensitivity. Production energy consumption uses a range desensitization algorithm to convert specific values ​​into ranges. Financial data uses role-based access control technology to only grant access to authorized personnel. Supply chain data is set with different levels of data viewing detail according to the level of cooperation. The dynamic permission adaptation module uses enterprise size, policy confidentiality level, and resource type as permission judgment attributes to establish an attribute-permission mapping rule base; it records key operations in the resource collaboration process and generates structured audit logs. The cross-system security collaboration module constructs a unified interface access point, and uses the national cryptographic SM4 algorithm to encrypt and transmit the interaction data between the platform and ERP and third-party service systems. The integrity of the interaction data is verified by comparing hash values ​​every hour. The intelligent operation and maintenance module monitors the system's operating status in real time through performance monitoring indicators, sets threshold warning rules, and triggers fault warnings when indicators exceed limits; at the same time, it adopts master-slave node cluster technology, which automatically switches to the backup node when the master node fails; and the operation and maintenance logs generated during the operation and maintenance process are classified and stored.

[0014] The beneficial effects of this invention are as follows: 1. This invention acquires the original text and interpretation of provincial and municipal policies through a policy and supporting resource collection unit with the help of a government-enterprise collaboration module. It also captures supporting rules and policy conflict handling guidelines from prefecture-level cities and counties through a cross-level detailed rules crawling module. A dynamic update scheduling module ensures that policy updates are synchronized within 10 minutes to avoid missing key information. The parsing and decomposition unit identifies policy mandatory requirements, core indicators and other elements through a multi-level semantic analysis module. The execution path mapping module uses a two-way mapping algorithm to convert policy quantitative indicators into time parameters and quality standards for business actions. The cross-policy conflict identification module can also generate mediation solutions. The combination of these two approaches transforms policies from "incomprehensible" to "executable," solving the problem of policy and business disconnect and reducing application failures caused by the lack of clear paths or missing details.

[0015] 2. This invention collects production energy consumption and supply chain data through a business scenario mapping unit, integrates it with the enterprise's future business plan, and forms a dynamic panoramic profile of the enterprise. Based on this profile, the matching and risk prediction unit constructs a three-dimensional risk transmission chain of enterprise-supply chain-region through a risk prediction model module. It quantifies the transmission probability using a risk factor transmission algorithm, assigns weights to 12-month core indicators using a time-series attention mechanism, and outputs the optimal rectification start time using a rectification cycle prediction function to avoid rectification delays. At the same time, the dividend opportunity mining module extracts policy dividend elements and matches them with enterprise plan fields at the field level, marking deadlines and material lists to help enterprises avoid risks in advance and accurately grasp dividends.

[0016] 3. This invention transforms the unstructured review progress of the government affairs platform into a standardized status of "materials submitted - under review - review result - public announcement - payment received" through a decision-making interaction unit, and sets progress warnings; the effect evaluation and feedback module compares the indicators before and after implementation to generate reports, and the solution iteration and optimization module establishes a decision optimization knowledge base based on the experience base to avoid enterprises repeatedly sorting out processes; the data security and operation and maintenance unit processes data according to sensitivity, encrypts transmission using the national cryptographic SM4 algorithm and verifies data integrity, and automatically switches between primary and backup nodes to ensure system stability; thus, it achieves full-process controllability and decision optimization of policy implementation, while also ensuring enterprise data and system security. Attached Figure Description

[0017] Figure 1 This invention provides a flowchart of an intelligent interactive platform for enterprise policy decision-making. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] like Figure 1 As shown, this embodiment of the invention provides an intelligent interactive platform for enterprise policy decision-making, the platform comprising: Policy and Supporting Resource Acquisition Unit: Used to acquire industry policy information and regional policy coordination data, and to classify and store implementation cases, resource information, and policy-business mapping template data; Analysis and decomposition unit: used to identify policy elements and rules, execute the mapping template of the policy and business, generate exclusive execution paths, and generate mediation plans and mediation basis by combining historical execution data; Business Scenario Mapping Unit: Used to build enterprise profiles, store core enterprise information and label business characteristics, collect production and supply chain data, model according to industry scenarios and perform tagging processing, and associate enterprise plans with the regional policy coordination data; Matching and Risk Prediction Unit: Based on enterprise profiles, calculate the early warning window period and matching degree, construct a three-dimensional risk transmission chain and quantify the transmission probability, thereby predicting risks, identifying policy dividends and marking deadlines and material lists; Risk Intervention and Implementation Guidance Unit: Develop intervention strategies based on risk levels, output a list of related actions, resources and standards, and integrate and categorize policy-specific resource pools; Decision-making interaction unit: Establish objective function, form multiple decision-making schemes for comparison, obtain review progress through government affairs platform, set progress warning, compare enterprise indicators before and after implementation, form experience base and establish decision optimization knowledge base; Data Security and Operations Unit: Processes enterprise data according to data sensitivity, establishes an attribute and permission mapping rule base and generates audit logs, encrypts and verifies the integrity of interactive data, monitors system operating status and performance, and automatically switches to a backup node when the primary node fails.

[0020] The specific policies and supporting resource collection units are as follows: The government-enterprise collaboration module connects with provincial and municipal government service platforms to obtain original policy texts, official interpretations, and industry-specific policy guidance. When the cross-level detailed rules crawler module is specifically crawling the supporting implementation rules of cities and counties, it simultaneously collects regional policy coordination data on local policy conflict handling guidelines. The Case Studies and Resource Library module stores implementation cases, a directory of compliance service agencies, and equipment supplier information, categorized by industry (manufacturing / service / new energy), policy type, and implementation effect. It also stores policy-business mapping template data. The dynamic update scheduling module triggers synchronization within 10 minutes when policies and supporting resources are updated. The case library is updated weekly. When the frequency of keywords in the policy text exceeds the threshold, the template node generation logic is automatically triggered to supplement the corresponding business nodes for qualification review and process verification.

[0021] The analysis and disassembly unit is specifically as follows: The multi-level semantic analysis module identifies mandatory and optional requirements, core and auxiliary indicators, application and rectification nodes in policies, and extracts policy priority rules such as national policies taking precedence over local policies, as well as industry compliance practices. The execution path mapping module calls the industry policy and business mapping templates stored in the policy and supporting resource collection unit, and generates a unique execution path by combining business process modeling symbols. It adopts a two-way mapping algorithm between policy indicators and business actions to convert the quantitative indicators in the policy (such as energy consumption limits and testing cycles) into time parameters and quality standards for business actions. At the same time, it associates with the existing business process node IDs of the enterprise to ensure that the generated execution path can be directly connected to the enterprise's internal process management system, and marks the responsible departments such as the production department / quality inspection department and the acceptance standards such as GB18483-2020. The cross-policy conflict identification module, when identifying policy conflicts, combines the company's past execution data to generate a mediation plan and output the policy basis. It first conducts an initial screening through a preset policy conflict rule library (such as the "detection cycle - appointment cycle" conflict rule), then calculates the semantic similarity of the conflict clauses, determines the conflict type, and calls the solution template in the historical execution data to generate a mediation plan that is suitable for the current scenario, providing a clear path for the company to match policies and avoid execution conflicts.

[0022] The business scenario mapping unit is specifically as follows: Construct a comprehensive enterprise profile that includes basic information, operational dynamics, supply chain connections, regional adaptation, and planning. The basic information module stores core information such as enterprise registration, industry classification, and qualification certificates, while also labeling business characteristics by sub-industry. The operational dynamics acquisition module connects to the ERP and MES systems in real time to obtain production capacity and energy consumption data. It also collects data from upstream and downstream of the supply chain. After real-time data cleaning, structured storage and field-level indexing, the data is transmitted to the matching and risk prediction unit. The business scenario modeling module models scenarios by industry segment, such as the stamping-welding-painting-final assembly process in automotive parts production, and the warehousing-transportation-delivery process in logistics companies. First, the core business of the industry is broken down into indivisible basic processes, then each process is associated with corresponding compliance indicators and production indicators, and finally mapped to the actual responsible departments of the enterprise, forming labeled data of process-indicator-responsible department. The future planning access module connects to the enterprise strategic planning system and production planning system through standardized APIs to obtain structured future business adjustment data, parse out core elements including capacity changes, new links, and resource requirements, and then perform field-level matching with the regional policy coordination data to generate a structured labeled dataset containing comparisons between the present and the future.

[0023] The matching and risk prediction unit is specifically as follows: Based on the enterprise panoramic profile constructed by the business scenario mapping unit, the dynamic matching algorithm module associates policy execution nodes with enterprise business plan nodes based on time-series similarity, establishes a rectification cycle prediction function based on the enterprise's historical rectification data, takes into account parameters including policy buffer period and enterprise business adjustment complexity, and outputs the optimal rectification start time; the matching degree calculation adopts time-series similarity algorithm, and outputs a matching result of 0-100 points by comparing the overlap and trend consistency between the time series of policy execution nodes and the time series of enterprise business plans. The risk prediction model module constructs a three-dimensional risk transmission chain of enterprises, supply chains, and regions. It uses a risk factor transmission algorithm to quantify the probability of risk transmission at each stage. By analyzing the impact coefficient of enterprise business adjustments on supply chain links and the correlation between supply chain link anomalies and regional policy compliance, it calculates the overall probability of risk transmission from enterprises to regions. At the same time, it introduces a time-series attention mechanism to assign dynamic weights to indicators such as energy consumption and environmental qualifications that have been frequently adjusted in the past 12 months and have large deviations in enterprise execution. These indicators are assigned dynamic weights based on their contribution to historical risk events. The dividend opportunity mining module identifies policy dividends that can be matched with corporate plans. It uses a policy dividend keyword matching algorithm to extract dividend elements from policy texts and performs field-level matching with business adjustment directions, resource inputs, and target indicators in corporate plans to filter out suitable dividend opportunities. At the same time, it extracts key information from the policy database to form a structured dividend opportunity dataset.

[0024] Rectification cycle prediction function: ; In the formula: This indicates the predicted optimal time to initiate rectification, expressed in days. It represents the interval from the current date to when the company should initiate policy compliance rectification. The value must meet the following conditions. ≥0; This indicates the average rectification period for similar business operations in the company's history, expressed in days. The value must meet the following requirements. >0; Data is extracted from the historical execution database of the business scenario mapping unit and statistically analyzed according to business type; This represents the business adjustment complexity coefficient, used to reflect the degree of impact of business complexity on the rectification cycle; 0 < ≤2 is an empirical coefficient based on fitting historical rectification data in the business scenario mapping unit. The more departments involved in the rectification and the more complex the process, the higher the coefficient. The larger the value, the more complex the associated calls are adjusted through business logic. This indicates the complexity of the current business adjustments, used to quantify the complexity of current rectification needs, 0 ≤ ≤1; calculated by the business scenario mapping unit, and normalized to a coefficient of 0-1 by breaking down the number of links, departments and resources involved in the rectification. This represents the policy buffer period utilization coefficient, used to reflect the degree to which the policy buffer period optimizes the rectification initiation time; 0 < ≤1; is an empirical coefficient based on the fitting of policy types in the analysis and decomposition unit. The longer the policy buffer period, the better. The larger the value, the more likely it is to be extracted using the policy buffer period label; This indicates the policy buffer period, expressed in days, representing the interval between the policy's release and the deadline for rectification. >0; Extracted from the policy text by the analysis and decomposition unit, for example, if the policy clearly states that energy consumption rectification should be completed within 30 days, then... =30.

[0025] Formula for matching policy and business time series similarity: ; In the formula: This indicates the time-series similarity and matching degree between policies and business operations, with the final output being a score of 0-100; a higher score indicates a stronger fit between the policy and the company's business. The time series overlap weight is used to reflect the importance of the time overlap between policy and business nodes for matching; 0 < ≤1; This is the weight of the business rules preset by the platform. Because the overlapping time nodes directly affect the efficiency of policy implementation, it has a higher priority and is usually selected as ≤1. =0.6, which can be fine-tuned according to industry characteristics through the experience base of the decision interaction unit; This indicates the degree of overlap between policy implementation milestones and enterprise business plan milestones, used to quantify the proportion of time overlap between the two, 0 ≤ ≤1; calculated by the matching and risk prediction unit, the calculation logic is the number of overlapping nodes / the total number of policy nodes. For example, if a policy contains 5 execution nodes, and 3 of them overlap with the enterprise's business plan nodes, then... =3 / 5=0.6; This represents the time series trend consistency weight, used to reflect the importance of the synchronization between policy and business trends for matching; 0 < ≤1 and + =1; This is the platform's preset business rule weight. Due to the impact of trend synchronization on the long-term adaptability of policies, its priority is secondary, and it is usually taken as 1. =0.4, and Complementary; The coefficient represents the consistency between policy implementation trends and corporate business plan trends, used to quantify the degree of synchronization between the two trends, 0 ≤ ≤1; calculated by the matching and risk prediction unit, using the Pearson correlation coefficient to analyze the time series of policy trends and corporate business plan trends, and normalizing the absolute value of the correlation coefficient to a coefficient of 0-1.

[0026] The specific details of the risk intervention and implementation guidance unit are as follows: Based on the risk prediction results of the matching and risk prediction unit, the graded intervention strategy module formulates measures according to the risk level. For low risk, it outputs rectification suggestion documents; for medium risk, it generates a phased rectification plan and sets progress reminders; and for high risk, it assigns a dedicated compliance consultant and triggers the connection of third-party resources. The high-risk rectification plan is generated using a matrix model of risk impact and rectification difficulty. The rectification phases are divided based on parameters such as the estimated economic losses caused by the risk and the resource input required for rectification. Each phase sets a progress threshold and verification node. Rectification data is collected in real time through the progress tracking API, and node timeout warnings are triggered. The execution guidance generation module outputs a list of associated actions, resources, and standards. Based on the policy area to which the risk belongs and the industry attributes of the enterprise, it selects the top 3 matching resources from the resource pool, including the resource provider's qualifications, service price, response time, etc. At the same time, it automatically associates the corresponding standard number and technical requirements in the standard database, clarifies the quality standard for each rectification action, and ensures that the rectification process complies with policy compliance requirements. The third-party resource collaboration module integrates a policy-specific resource pool, categorizes resources by policy type (environmental protection / taxation / technological upgrading), industry (automotive / electronics), and service capabilities (response within 3 days / national-level qualification), and builds a resource retrieval index; it supports enterprises to operate online throughout the entire process of resource screening, online appointment, and progress tracking.

[0027] The overall probability formula for three-dimensional risk transmission: ; In the formula: This represents the overall probability of risk spreading from a company to a region, 0 ≤ ≤1; This result directly determines the level of risk intervention, for example A value greater than 0.7 triggers a high-risk intervention; a value less than 0.3 triggers a low-risk intervention A medium-risk intervention is triggered when the value is ≤0.7. This represents a month index based on the time series dimension, covering the last 12 months. Indicates the first The time-series attention weight of the monthly core indicators is used to reflect the contribution of the indicator to risk in that month, and the value must satisfy 0 < ≤1 and =1; Calculated by the matching and risk prediction unit, the indicators for the past 3 months have a higher weight due to their stronger timeliness. Indicates the first The initial risk probability of a company in a given month is used to quantify the compliance risk the company faces that month, 0 ≤ ≤1; Extracted from the business scenario mapping unit, by comparing the enterprise's first... The actual indicators and policy standards for a month are calculated, and the compliance deviation rate is normalized to a probability value of 0-1. This represents the impact coefficient of a company's business adjustments on the supply chain, used to reflect the likelihood of risk propagating from the company to the supply chain, 0 ≤ ≤1; calculated by the matching and risk prediction unit, the higher the enterprise's dependence on the supply chain, The larger the value, the more it is calculated based on data from upstream and downstream of the supply chain. This indicates the correlation between supply chain anomalies and regional policy compliance, reflecting the likelihood of risk transmission from the supply chain to the region, 0 ≤ ≤1; Extracted from the regional policy collaboration database of the policy and supporting resource collection unit, the stricter the regional policy requirements for supply chain compliance, the higher the compliance rate. The larger the value.

[0028] The decision-making interaction unit is specifically as follows: Based on the execution progress data of the aforementioned risk intervention and execution guidance unit, the intelligent decision generation module links the entire policy lifecycle, calibrating the key time points of the entire policy lifecycle (application start / end, acceptance time) with the enterprise execution nodes (material preparation, internal review, submission) on a timeline, setting the prerequisite dependencies and time buffer thresholds between nodes; at the same time, it adopts a multi-objective optimization algorithm, with subsidy amount, application cycle, and success rate as optimization objectives, and constructs an objective function in combination with historical application data, outputting a Pareto optimal solution set, forming multiple decision schemes for comparison; The execution progress tracking module obtains the review progress and transforms the unstructured progress information returned by the government affairs platform into a standardized status of material submission - under review - review result - public announcement - payment received; at the same time, it sets progress warning rules, which automatically trigger correction reminders and progress query reminders when the time remaining before the application deadline or the review feedback time is less than the preset threshold, to ensure that the execution progress is controllable; The effect evaluation and feedback module compares enterprise indicators before and after implementation, calculates the rate of change of enterprise compliance indicators and operational indicators after implementation, conducts deviation analysis with the target values ​​required by the policy, and generates an effect evaluation report; at the same time, it uses an experience extraction algorithm to mine key operations of successful cases and core issues of failed cases from historical decision-making implementation data to form a structured experience library. The solution iteration and optimization module establishes a decision optimization knowledge base based on the experience base and adopts a case similarity matching algorithm. When an enterprise initiates a similar decision request, it retrieves the lessons learned from similar cases in the knowledge base and automatically pushes optimization suggestions. It supports manual annotation of new policy requirements, new industry regulations and other information, which are synchronized to the knowledge base. At the same time, the experience data is transmitted to the parsing and decomposition unit and the matching and risk prediction unit.

[0029] The data security and operation and maintenance unit is specifically as follows: The hierarchical data desensitization module processes enterprise data according to data sensitivity. Production energy consumption data uses a range desensitization algorithm to convert specific values ​​into ranges. Financial data uses role-based access control technology to only grant access to authorized personnel. Supply chain data uses hierarchical access control to set different levels of data viewing detail according to the level of cooperation, ensuring that sensitive data is not leaked. The dynamic permission adaptation module uses enterprise size, policy confidentiality level, and resource type as permission judgment attributes to establish an attribute and permission mapping rule base. When a user initiates a resource access request, the system verifies the matching degree between the user's attributes and the requested resource's attributes in real time and assigns the corresponding operation permissions. At the same time, it records key operations in the resource collaboration process, such as access requests, data queries, and operation records, and generates structured audit logs. The cross-system security collaboration module constructs a unified interface access point, and uses the national cryptographic SM4 algorithm to encrypt and transmit the interactive data between the platform and ERP and third-party service systems. At the same time, a data integrity verification mechanism is set up to verify the integrity of the interactive data by comparing hash values ​​every hour. When data inconsistency is detected, the data repair process is automatically triggered to retrieve the correct data from the backup database for retransmission, ensuring that data interaction is secure and reliable. The intelligent operation and maintenance module monitors the system's operating status in real time through performance monitoring indicators (CPU utilization, memory usage, interface response time), sets threshold warning rules, and triggers fault warnings when indicators exceed limits; at the same time, it adopts master-slave node cluster technology, and automatically switches to the backup node when the master node fails; and it classifies and stores the operation and maintenance logs generated during the operation and maintenance process.

[0030] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0031] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent interactive platform for enterprise policy decision-making, characterized in that: The platform includes: Policy and Supporting Resource Acquisition Unit: Used to acquire industry policy information and regional policy coordination data, and to classify and store implementation cases, resource information, and policy-business mapping template data; Analysis and decomposition unit: used to identify policy elements and rules, execute the mapping template of the policy and business, generate exclusive execution paths, and generate mediation plans and mediation basis by combining historical execution data; Business Scenario Mapping Unit: Used to build enterprise profiles, store core enterprise information and label business characteristics, collect production and supply chain data, model according to industry scenarios and perform tagging processing, and associate enterprise plans with the regional policy coordination data; Matching and Risk Prediction Unit: Based on the enterprise panoramic profile constructed by the business scenario mapping unit, the dynamic matching algorithm module associates policy execution nodes and enterprise business plan nodes based on temporal similarity, establishes a rectification cycle prediction function based on the enterprise's historical rectification data, takes into account parameters including policy buffer period and enterprise business adjustment complexity, and outputs the optimal rectification start time; The risk prediction model module constructs a three-dimensional risk transmission chain of enterprises, supply chains, and regions. It uses a risk factor transmission algorithm to quantify the probability of risk transmission at each stage. By analyzing the impact coefficient of enterprise business adjustments on supply chain links and the correlation between supply chain link anomalies and regional policy compliance, it calculates the overall probability of risk transmission from enterprises to regions. At the same time, it introduces a time-series attention mechanism to assign dynamic weights to core indicators within 12 months by calculating the contribution of the indicators in historical risk events. The dividend opportunity mining module identifies policy dividends that can be matched with enterprise plans. It uses a policy dividend keyword matching algorithm to extract dividend elements from policy texts and performs field-level matching with enterprise plans to filter out suitable dividend opportunities. At the same time, it extracts key information from the policy database to form a structured dividend opportunity dataset. Risk Intervention and Implementation Guidance Unit: Develop intervention strategies based on risk levels, output a list of related actions, resources and standards, and integrate and categorize policy-specific resource pools; Decision-making interaction unit: Based on the execution progress data of the risk intervention and execution guidance unit, the intelligent decision generation module associates the entire policy cycle and calibrates the key time points of the entire policy cycle with the enterprise execution nodes on the timeline; at the same time, it adopts a multi-objective optimization algorithm with subsidy amount, application cycle and success rate as optimization objectives, and constructs an objective function in combination with historical application data to form a comparison of 3-5 decision schemes; The execution progress tracking module obtains the review progress and transforms the unstructured progress information returned by the government affairs platform into a standardized status of material submission - under review - review result - public announcement - payment received; at the same time, it sets progress warning rules to automatically trigger reminders when the time remaining before the application deadline or the review feedback time is less than a preset threshold. The effect evaluation and feedback module compares enterprise indicators before and after implementation, calculates the rate of change of enterprise compliance indicators and operational indicators after implementation, conducts deviation analysis with the target values ​​required by the policy, and generates an effect evaluation report; it uses an experience extraction algorithm to mine key operations of successful cases and core issues of failed cases from historical decision-making implementation data to form a structured experience library; The solution iteration and optimization module establishes a decision optimization knowledge base based on the experience base and adopts a case similarity matching algorithm. When an enterprise initiates a similar decision request, it retrieves the lessons learned from similar cases in the knowledge base and automatically pushes optimization suggestions. It supports manual annotation of new information, which is synchronized to the knowledge base. At the same time, the experience data is transmitted to the parsing and decomposition unit and the matching and risk prediction unit. Data Security and Operations Unit: Processes enterprise data according to data sensitivity, establishes an attribute and permission mapping rule base and generates audit logs, encrypts and verifies the integrity of interactive data, monitors system operating status and performance, and automatically switches to a backup node when the primary node fails.

2. The intelligent interactive platform for enterprise policy decision-making according to claim 1, characterized in that: The specific policies and supporting resource collection units are as follows: The government-enterprise collaboration module connects with provincial and municipal government service platforms to obtain original policy texts, official interpretations, and industry policy guidance. When the cross-level detailed rules crawler module is specifically crawling the supporting implementation rules of cities and counties, it simultaneously collects regional policy coordination data on local policy conflict handling guidelines. The Case Studies and Resource Library module stores implementation cases, a directory of compliance service agencies, and equipment supplier information, categorized by industry, policy type, and implementation effect. It also stores policy-business mapping template data. The dynamic update scheduling module triggers synchronization within 10 minutes when policies and supporting resources are updated. The case library is updated weekly. When the frequency of keywords in the policy text exceeds the threshold, the template node generation logic is automatically triggered to supplement the corresponding business nodes for qualification review and process verification.

3. The intelligent interactive platform for enterprise policy decision-making according to claim 2, characterized in that: The specific analysis and disassembly unit is as follows: The multi-level semantic analysis module identifies mandatory and optional requirements, core and auxiliary indicators, application and rectification nodes in policies, and extracts policy priority rules and industry compliance practices. The execution path mapping module calls the industry policy and business mapping template stored in the policy and supporting resource collection unit, combines the business process modeling symbols to generate a unique execution path, and adopts a two-way mapping algorithm between policy indicators and business actions to transform the quantitative indicators in the policy into time parameters and quality standards for business actions. The cross-policy conflict identification module, when identifying policy conflicts, combines the company's past execution data to generate a mediation plan and output the policy basis. The conflict identification first screens through a preset policy conflict rule library, then calculates the semantic similarity of the conflict clauses, and after determining the conflict type, calls the solution template in the historical execution data.

4. The intelligent interactive platform for enterprise policy decision-making according to claim 3, characterized in that: The specific business scenario mapping unit is as follows: Construct a comprehensive enterprise profile that includes basic information, operational dynamics, supply chain connections, regional adaptation, and planning. The basic information module stores core information such as enterprise registration, industry classification, and qualification certificates, while also labeling business characteristics by sub-industry. The operational dynamics acquisition module is used to acquire production capacity and energy consumption data, and also collects data from upstream and downstream of the supply chain. After real-time data cleaning, structured storage and field-level indexing, the data is transmitted to the matching and risk prediction unit. The business scenario modeling module models scenarios by industry segmentation. First, the core business of the industry is broken down into indivisible basic links. Then, each link is associated with corresponding compliance indicators and production indicators. Finally, it is mapped to the actual responsible departments of the enterprise, forming tagged data of link-indicator-responsible department. The future planning access module is used to acquire structured future business adjustment data, parse out core elements including changes in production capacity, new links, and resource requirements, and then perform field-level matching with the regional policy coordination data to generate a structured labeled dataset containing comparisons between the present and the future.

5. The intelligent interactive platform for enterprise policy decision-making according to claim 4, characterized in that: The specific risk intervention and implementation guidance unit is as follows: Based on the risk prediction results of the matching and risk prediction unit, the graded intervention strategy module formulates measures according to the risk level. For low risk, it outputs rectification suggestion documents; for medium risk, it generates a phased rectification plan and sets progress reminders; and for high risk, it assigns a dedicated compliance consultant. The execution guidance generation module outputs a list of associated actions, resources, and standards. Based on the policy area to which the risk belongs and the industry attributes of the enterprise, it selects the top 3 matching resources from the resource pool. At the same time, it automatically associates the corresponding standard number and technical requirements in the standard database to clarify the quality standard for each rectification action. The third-party resource collaboration module integrates a policy-specific resource pool, categorizes resources by policy type, industry, and service capabilities, and builds a resource retrieval index. It supports enterprises to operate the entire process online, including resource screening, online appointment, and progress tracking.

6. The intelligent interactive platform for enterprise policy decision-making according to claim 5, characterized in that: The data security and operation and maintenance unit is specifically as follows: The hierarchical data desensitization module processes enterprise data according to data sensitivity. Production energy consumption uses a range desensitization algorithm to convert specific values ​​into ranges. Financial data uses role-based access control technology to only grant access to authorized personnel. Supply chain data is set with different levels of data viewing detail according to the level of cooperation. The dynamic permission adaptation module uses enterprise size, policy confidentiality level, and resource type as permission judgment attributes to establish an attribute-permission mapping rule base; it records key operations in the resource collaboration process and generates structured audit logs. The cross-system security collaboration module constructs a unified interface access point, and uses the national cryptographic SM4 algorithm to encrypt and transmit the interaction data between the platform and ERP and third-party service systems. The integrity of the interaction data is verified by comparing hash values ​​every hour. The intelligent operation and maintenance module monitors the system's operating status in real time through performance monitoring indicators, sets threshold warning rules, and triggers fault warnings when indicators exceed limits; at the same time, it adopts master-slave node cluster technology, which automatically switches to the backup node when the master node fails; and the operation and maintenance logs generated during the operation and maintenance process are classified and stored.

Images