Intelligent cloud projection system, method and electronic device

Abstract

The application provides a smart cloud bidding system, method and electronic equipment, and relates to the technical field of digital bidding management, and comprises: a bidding database used for storing historical bidding documents and feature labeling information thereof; an intelligent recommendation module used for constructing a project feature vector and generating a candidate document recommendation list based on a hybrid recommendation model; a risk identification engine used for performing multidimensional risk automatic identification on the candidate document and the bidding document, and outputting risk items and risk levels; a collaborative workbench used for real-time collaborative editing and associated display of the risk items, and persisting the approval flow and version trace in a directed acyclic graph structure; and a compliance dynamic adaptation module used for listening to external policy changes and dynamically updating verification rules. Through deep integration of intelligent recommendation, risk identification and compliance adaptation, the application solves the problems of redundant traditional manual bidding management process, information lag and difficult traceability.

Description

Technical Field

[0001] This invention relates to the field of digital management technology for bidding and tendering, and in particular to a smart cloud bidding system, method, and electronic device. Background Technology

[0002] With the acceleration of enterprise digital transformation, information technology in bidding management has become the mainstream trend in the industry. According to the "2024 Enterprise Procurement and Bidding Digitalization Report," more than 65% of large and medium-sized enterprises in China have deployed bidding management systems, and 80% of those that have not deployed such systems suffer from redundant processes and information lag. Meanwhile, government procurement and large-scale project bidding have increasingly higher requirements for the standardization and traceability of bidding processes, making manual management models insufficient to meet policy compliance needs.

[0003] The current bidding process relies on manual data processing, which generally suffers from the following core pain points: First, redundant processes lead to slow response times for urgent bidding projects, and excessively long manual processes result in missed market opportunities; Second, human error leads to a high error rate in bidding data, resulting in significant losses due to failed bids each year; Third, bidding information is scattered across local documents in various departments, causing delays in cross-departmental information synchronization, and requiring manual review of paper records to trace approval nodes, which is time-consuming and prone to losing key information.

[0004] Existing bidding management systems primarily focus on online processes, but they exhibit significant shortcomings in areas such as intelligent decision support, dynamic adaptation to policy compliance, automatic risk identification, and real-time data synchronization across heterogeneous systems. For example, most systems only offer fixed-field retrieval and simple process approvals, lacking intelligent recommendation capabilities based on historical bidding data; static permissions and encryption measures are insufficient to meet dynamic security requirements; and policy updates require manual code modification, resulting in delayed responses and a high risk of errors.

[0005] Therefore, a smart cloud projection system, method, and electronic device are proposed. Summary of the Invention

[0006] This manual provides a smart cloud bidding system, method, and electronic device that solves the problems of redundancy, information lag, and difficulty in traceability in traditional manual bidding management processes by deeply integrating intelligent recommendation, risk identification, and compliance adaptation.

[0007] This manual provides a smart cloud projection system, including: A bidding database is used to store historical bidding documents and their feature annotation information, which includes at least project type, bidding party attributes, technical parameter tags, and historical bidding results. The intelligent recommendation module is used to obtain the demand information of new bidding projects and construct project feature vectors. Based on the hybrid recommendation model, it calculates the matching degree between the project feature vectors and each historical bidding document in the bidding data database, and generates a candidate data recommendation list. The risk identification engine is used to automatically obtain the candidate material recommendation list generated by the intelligent recommendation module and the tender document prepared by the user based on the recommendation list, and to automatically identify the risks of the candidate materials and the tender document in multiple dimensions, and output the risk items and risk levels. The collaborative workbench is used to provide real-time collaborative editing of the tender documents by multiple users. The risk items output by the risk identification engine are displayed in association on the editing interface, and the approval process and version traceability are persisted in a directed acyclic graph structure. The compliance dynamic adaptation module is used to monitor external policy release sources, capture compliance requirement change events, identify the affected verification rules and update them dynamically. The updated verification rules directly apply to the identification logic of the risk identification engine.

[0008] Optionally, the hybrid recommendation model includes: The collaborative filtering scoring unit calculates the similarity between the new bidding project and the historical projects based on the latent semantic model of the historical winning bids. The content matching scoring unit uses a semantic matching model to calculate the semantic similarity between the tender requirements text and the historical data text; The feedback weighting scoring unit dynamically adjusts the weight of materials based on the adoption status of historical recommended materials and the bidding results of the adopted projects; The fusion unit integrates the collaborative filtering score, content matching score, and feedback weight score according to adjustable parameters to generate a final recommendation score and output the candidate material recommendation list in a sorted manner.

[0009] Optionally, the risk identification engine includes: The qualification validity detection unit uses a pre-trained named entity recognition model to perform sequence annotation on the OCR text of the qualification certificate images in the candidate material recommendation list, extracts the validity period entity and compares it with the current date to identify the risk of qualification expiration. The parameter conflict detection unit performs graph matching between the structured conditions in the tender document and the technical parameters in the bid document, and identifies parameter conflict risks based on a preset bid knowledge graph. The comprehensive scoring unit extracts multi-dimensional features, including qualification completeness and parameter conflict number, and inputs them into an ensemble learning scoring model trained with historical bidding data to output a comprehensive risk score and warning level.

[0010] Optionally, the collaborative workbench is specifically used for: Based on the conflict-free copy data type algorithm, each user's editing operation is encapsulated into atomic operations with unique identifiers and broadcast. The reception status is tracked through state vectors, and conflicts are resolved according to a preset strategy. In the editing interface, the risk items marked by the risk identification engine are highlighted and risk details are displayed; A document status snapshot is generated after each approval node is completed. All snapshots form the directed acyclic graph structure, which supports previewing and backtracking of any historical version.

[0011] Optionally, the compliance dynamic adaptation module is specifically used for: By comparing the obtained policy clauses with the existing verification rule descriptions in the system using a semantic similarity model, the affected rule nodes can be located. The rule engine is invoked to match rule templates from the configuration center, automatically instantiate new rules and load them into the runtime cache, while marking old rules as invalid, thus completing hot updates; The updated verification rules will automatically take effect the next time the risk identification engine is started.

[0012] Optional, including: The feedback weighting scoring unit is based on the winning bid results of the adopted projects, which are provided by the strategy analysis module. The strategy analysis module is also used to: extract the price weight, technology weight and competitor influence factor of historical winning bids, and construct a winning bid probability prediction model; For the current bidding project, the risk items and risk levels output by the risk identification engine are used as one of the input features to generate a suggested price range and qualification configuration scheme.

[0013] This manual provides a smart cloud projection method, including: Obtain the requirements information for new bidding projects and construct project feature vectors; The project feature vector is input into the hybrid recommendation model to calculate the matching degree with each historical document in the bidding data database, generate a candidate data recommendation list and push it to the user interface; The system receives tender documents prepared by users based on the recommended list, activates the risk identification engine, performs multi-dimensional risk identification on the candidate materials and the tender documents, and outputs risk items and risk levels. In the collaborative workbench, the aforementioned risk items are displayed in association with the editing interface of the tender document, and the editing operations and risk information are synchronized in real time during the user's collaborative editing process; By monitoring external policy release sources, when changes in compliance requirements are detected, the verification rules referenced by the risk identification engine are dynamically updated so that the updated rules take effect in the next risk identification.

[0014] Optionally, the step of inputting the project feature vector into the hybrid recommendation model and calculating the matching degree with each historical document in the bidding database includes: Using an implicit semantic model decomposed by alternating least squares, a collaborative filtering score is calculated between the project feature vector and the historical winning project vector. A pre-trained semantic matching model is used to calculate the semantic similarity between the tender requirements text and the historical data summary text, which is then used as the content matching score. The feedback weight score is calculated based on the adoption rate of historical recommendation materials and subsequent bidding results; The collaborative filtering score, content matching score, and feedback weight score are merged with adjustable weights, sorted, and then a candidate material recommendation list is generated.

[0015] This specification also provides an electronic device, wherein the electronic device includes: A processor; and a memory storing computer-executable instructions, which, when executed, cause the processor to perform any of the methods described above.

[0016] This specification also provides a computer-readable storage medium that stores one or more programs that, when executed by a processor, implement any of the methods described above.

[0017] This invention achieves intelligent recommendation of bidding materials by integrating collaborative filtering scoring, content matching scoring, and feedback weight scoring through a hybrid recommendation model, reducing manual retrieval and repetitive compilation workload. The recommendation results are dynamically optimized based on adoption rate and bidding results, forming a continuously evolving intelligent reuse capability. Named entity recognition automatically detects the risk of expired qualifications, knowledge graph matching identifies parameter conflicts, and an ensemble learning model outputs a comprehensive risk level, significantly improving efficiency and reducing the error rate of bidding materials compared to manual review. The CRDT algorithm enables real-time collaborative editing by multiple users, automatically resolving conflicts; the approval process forms a directed acyclic graph version record, supporting one-click retrospection of any historical node, solving the problems of information fragmentation and traceability difficulties in traditional collaboration. A semantic similarity model automatically locates verification rules affected by policy changes, calling the rule engine for hot updates, controlling policy response time within 24 hours, and avoiding invalid bids due to compliance delays. A bidding probability prediction model is established through big data mining, using risk identification results as input features to output pricing suggestions and qualification configuration schemes, realizing a transformation from experience-driven to data-driven approaches. The adoption results, winning bid results, risk identification results, and other data form a closed-loop feedback between modules. The system continuously optimizes itself as business data accumulates, resulting in long-term value increase. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the structure of a smart cloud projection system provided in the embodiments of this specification; Figure 2 A schematic diagram illustrating the principle of a smart cloud projection method provided in the embodiments of this specification; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this specification; Figure 4 This is a schematic diagram of a computer-readable medium provided for embodiments of this specification. Detailed Implementation

[0020] The following description is intended to disclose the present invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art. The basic principles of the invention defined in the following description can be applied to other embodiments, modifications, improvements, equivalents, and other technical solutions that do not depart from the spirit and scope of the invention.

[0021] The following is in conjunction with the appendix Figures 1-4 Exemplary embodiments of the invention will be described more fully here. However, exemplary embodiments can be implemented in many forms and should not be construed as limiting the invention to the embodiments set forth herein. Rather, these exemplary embodiments are provided to make the invention more comprehensive and complete, and to facilitate a full communication of the inventive concept to those skilled in the art. The same reference numerals in the figures denote the same or similar elements, components, or parts, and therefore repeated descriptions of them are omitted.

[0022] Subject to the technical concept of this invention, the features, structures, characteristics or other details described in a particular embodiment may be combined in one or more other embodiments in a suitable manner.

[0023] In the description of specific embodiments, the features, structures, characteristics, or other details described in this invention are intended to enable those skilled in the art to fully understand the embodiments. However, it is not excluded that those skilled in the art can practice the technical solutions of this invention without one or more of the specific features, structures, characteristics, or other details.

[0024] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0025] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0026] The terms “and / or” or “and / or” include all combinations of any one or more of the listed items.

[0027] Figure 1 A schematic diagram of a smart cloud projection system provided in the embodiments of this specification includes: The bidding database 10 is used to store historical bidding documents and their feature annotation information, which includes at least project type, bidding party attributes, technical parameter tags and historical bidding results. The intelligent recommendation module 20 is used to obtain the demand information of new bidding projects and construct project feature vectors. Based on the hybrid recommendation model, it calculates the matching degree between the project feature vectors and each historical bidding document in the bidding data database, and generates a candidate data recommendation list. The risk identification engine 30 is used to automatically obtain the candidate material recommendation list generated by the intelligent recommendation module and the tender document prepared by the user based on the recommendation list, and to automatically identify the risks of the candidate materials and the tender document in multiple dimensions, and output the risk items and risk levels. The collaborative workbench 40 is used to provide real-time collaborative editing of the tender documents by multiple users, and displays the risk items output by the risk identification engine in the editing interface, and persists the approval process and version traceability in a directed acyclic graph structure. The compliance dynamic adaptation module 50 is used to monitor external policy release sources, capture compliance requirement change events, identify the affected verification rules and update them dynamically. The updated verification rules directly apply to the identification logic of the risk identification engine.

[0028] In the specific implementation described in this specification, the Smart Cloud Investment System adopts a three-layer design: "Front-end Interaction Layer - Business Logic Layer - Data Storage Layer". The front-end is developed based on the Vue.js framework, supporting adaptive layout across multiple terminals (PC and mobile) through component-based construction. The business logic layer is centered on a Spring Boot microservice architecture, divided into data management services, intelligent recommendation services, risk identification services, collaborative approval services, compliance adaptation services, and strategy analysis services. Each microservice is deployed independently and decoupled through message queues, supporting up to 500 simultaneous online users. The data storage layer stores core business data through a MySQL cluster, provides high-speed caching and publish / subscribe capabilities through a Redis cluster, supports full-text search and log analysis through Elasticsearch, and stores original tender documents through object storage services.

[0029] The bidding database stores and manages over 3,000 historical bidding documents and over 200 project case studies accumulated over more than ten years of bidding experience. Each document is tagged with features including: project type (e.g., engineering construction, goods procurement, services), bidding party attributes (including industry classification, company size, ownership type, etc.), technical parameter tags (key technical indicators extracted from bidding documents and bid proposals), and historical bidding results (won or not won, winning bid price and scoring details).

[0030] The bidding data database is built on the company's existing enterprise-level server cluster and data storage center, which can directly support system deployment. In terms of data security, AES-256 encryption algorithm is used to encrypt and store sensitive fields, and a three-level permission hierarchy system (system administrator, department head, and ordinary user) is used to control the scope of data access, realize the automatic recording of operation behavior audit logs and real-time monitoring of abnormal risk warnings.

[0031] Optionally, the hybrid recommendation model includes: The collaborative filtering scoring unit calculates the similarity between the new bidding project and the historical projects based on the latent semantic model of the historical winning bids. The content matching scoring unit uses a semantic matching model to calculate the semantic similarity between the tender requirements text and the historical data text; The feedback weighting scoring unit dynamically adjusts the weight of materials based on the adoption status of historical recommended materials and the bidding results of the adopted projects; The fusion unit integrates the collaborative filtering score, content matching score, and feedback weight score according to adjustable parameters to generate a final recommendation score and output the candidate material recommendation list in a sorted manner.

[0032] In the specific implementation of this specification, when a user creates a new bidding project, the system obtains project requirement information and constructs a project feature vector. The feature vector includes the following dimensions: project category, bid amount, technical parameter requirements, bidding party's industry attributes, and historical preferences, etc.

[0033] The hybrid recommendation model consists of four units working together: Collaborative filtering scoring unit: Employing an alternating least squares decomposition latent semantic model, this unit maps project feature vectors to the same latent space with historical winning project vectors and calculates cosine similarity. This unit is trained using data from over 200 project cases accumulated by the company, enabling it to capture implicit relationships between different bidding parties.

[0034] Content matching scoring unit: Employing a pre-trained semantic matching model, this unit performs semantic vector comparison between the technical requirements and commercial terms in the tender documents and the summary text of historical data in the tender database. The response time for a single data retrieval is less than 1 second, and for high-frequency retrievals, this can be further reduced to within 0.5 seconds using Redis caching preheating technology.

[0035] Feedback weighting scoring unit: Dynamically tracks the adoption status of recommended materials and the final bidding results of projects after adoption, forming a closed-loop feedback mechanism. If a certain type of material is adopted multiple times and has a high success rate, its weight is automatically increased; conversely, if the success rate is low after adoption, its weight gradually decreases.

[0036] Fusion Unit: The above three scores are weighted and fused according to adjustable parameters α, β, and γ. The final recommended score S = α·S1 + β·S2 + γ·S3 is then sorted and the Top-N candidate data recommendation list is output.

[0037] This module also features an intelligent recommendation function for bidding documents, which can automatically match historical high-quality cases and compliant documents based on project type, bidding party preferences, and other dimensions, reducing the workload of repetitive preparation.

[0038] Optionally, the risk identification engine 30 includes: The qualification validity detection unit uses a pre-trained named entity recognition model to perform sequence annotation on the OCR text of the qualification certificate images in the candidate material recommendation list, extracts the validity period entity and compares it with the current date to identify the risk of qualification expiration. The parameter conflict detection unit performs graph matching between the structured conditions in the tender document and the technical parameters in the bid document, and identifies parameter conflict risks based on a preset bid knowledge graph. The comprehensive scoring unit extracts multi-dimensional features, including qualification completeness and parameter conflict number, and inputs them into an ensemble learning scoring model trained with historical bidding data to output a comprehensive risk score and warning level.

[0039] In the specific implementation of this specification, the risk identification engine automatically obtains the candidate material recommendation list output by the intelligent recommendation module, as well as the tender document prepared by the user based on the list, and performs multi-dimensional risk identification.

[0040] The qualification validity detection unit first performs image OCR recognition on the qualification certificates in the candidate material recommendation list to extract the text content; then, it uses a pre-trained named entity recognition model (such as BERT-CRF) to perform sequence annotation on the OCR text and extract the "expiration date" entity; finally, it compares the result with the current date in the system. If the expiration date is less than a preset threshold, it is marked as "about to expire" risk; if it has exceeded the expiration date, it is marked as "expired" risk.

[0041] Parameter Conflict Detection Unit: Based on industry knowledge accumulated from bidding experience, a bidding knowledge graph is constructed. The structured conditions in the bidding documents (such as technical specifications, scope of supply, qualification thresholds, etc.) are matched with the technical parameters in the bidding proposal. Any discrepancies or contradictions are marked as parameter conflict risks.

[0042] The comprehensive scoring unit extracts 18 features from candidate materials and tender documents, including completeness of qualifications, number of parameter conflicts, price deviation, and deviation of historical projects. These features are then input into the LightGBM ensemble learning model trained with data from over 3,000 historical bidding cases, and output a comprehensive risk score of 0-1 and a warning level.

[0043] The engine can automatically identify eight core risks, such as expired qualifications and parameter conflicts, improving efficiency by 80% compared to traditional manual review. The identification results are pushed to the collaborative workbench interface in real time via WebSocket and highlighted.

[0044] Optionally, the collaborative workbench is specifically used for: Based on the conflict-free copy data type algorithm, each user's editing operation is encapsulated into atomic operations with unique identifiers and broadcast. The reception status is tracked through state vectors, and conflicts are resolved according to a preset strategy. In the editing interface, the risk items marked by the risk identification engine are highlighted and risk details are displayed; A document status snapshot is generated after each approval node is completed. All snapshots form the directed acyclic graph structure, which supports previewing and backtracking of any historical version.

[0045] In the specific implementation of this specification, the collaborative workbench provides an environment for cross-departmental collaborative preparation of tender documents, integrating functional modules such as requirement submission, approval process, and progress tracking.

[0046] Real-time collaborative editing: Enables multiple users to edit the same document simultaneously based on a conflict-free copy data type algorithm. Each user's insertion and deletion operations on document paragraphs are encapsulated as atomic operations with unique identifiers (including operation type, position, content, timestamp, and user ID), and broadcast in real-time to all online users via WebSocket. The receiving end tracks the reception status of each user's operations through a state vector, ensuring global consistency. Concurrent operations are automatically merged when they do not conflict; conflicts are resolved using a timestamp-first strategy, with a subtle prompt displayed on the interface.

[0047] Risk association display: In the editing interface, risk items marked by the risk identification engine are highlighted in red, and hovering the mouse over them displays risk details (risk type, severity level, and handling suggestions). Users can directly replace expired qualifications or adjust conflicting parameters during the editing process.

[0048] Version tracking: After each approval node is passed, the system generates a complete snapshot of the current document state. All snapshots form a directed acyclic graph structure according to the approval flow order, allowing users to preview and rewind the complete document content at any node. The approval process takes less than 1 minute. This mechanism replaces the traditional method of manually reviewing paper records, solving the problems of information loss and difficulty in tracing.

[0049] Optionally, the compliance dynamic adaptation module 50 is specifically used for: By comparing the obtained policy clauses with the existing verification rule descriptions in the system using a semantic similarity model, the affected rule nodes can be located. The rule engine is invoked to match rule templates from the configuration center, automatically instantiate new rules and load them into the runtime cache, while marking old rules as invalid, thus completing hot updates; The updated verification rules will automatically take effect the next time the risk identification engine is started.

[0050] In the specific implementation of this specification, the compliance dynamic adaptation module adopts the "compliance dynamic adaptation" mechanism to solve the problem that traditional systems need to be manually configured to adapt to policy changes.

[0051] Monitoring and Capture: Utilizing both scheduled tasks and webhooks, the system monitors policy releases from government procurement websites and industry regulatory APIs. Upon capturing new policies or revised clauses, the system retrieves the modified text.

[0052] Semantic comparison and localization: The BERT semantic similarity model is used to compare the text of the new policy clauses with the existing verification rule descriptions in the system one by one. When the similarity exceeds the threshold but there are differences, the rule node is marked as an "affected rule node".

[0053] Hot rule updates: The Drools rule engine is invoked to match the rule template corresponding to the new rule from the configuration center. The new rule is automatically instantiated and loaded into the runtime cache, while the old rule is marked as invalid. This process achieves hot rule updates without restarting system services or manually modifying code.

[0054] Activation Mechanism: After the update is completed, the risk identification engine will be notified via a message queue. The updated verification rules will automatically take effect the next time the risk identification engine starts a task, with the policy response time controlled within 24 hours.

[0055] Optional, including: The feedback weighting scoring unit is based on the winning bid results of the adopted projects, which are provided by the strategy analysis module. The strategy analysis module is also used to: extract the price weight, technology weight and competitor influence factor of historical winning bids, and construct a winning bid probability prediction model; For the current bidding project, the risk items and risk levels output by the risk identification engine are used as one of the input features to generate a suggested price range and qualification configuration scheme.

[0056] In the specific implementation of this specification, the strategy analysis module obtains risk item and risk level data from the risk identification engine and incorporates them as one of the input features into the analysis.

[0057] The module first conducts big data mining on historical winning bids, and uses feature engineering methods to extract key influencing factors from 200+ winning bid cases: price weight (the importance of the price in the bid evaluation), technology weight (the weight of the technical solution score), commercial weight (the weight of the qualification and performance score), and competitor influence factor (the pricing strategies and advantageous areas of common competitors).

[0058] A probability prediction model for winning bids is constructed using the aforementioned factors. For bidding projects currently being prepared, the number and level of risk items output by the risk identification engine are used as one of the input features (the fewer the risks and the lower the level, the higher the probability of winning the bid). The model outputs a suggested price range and qualification configuration scheme. This function realizes the transformation from "experience-driven" to "data-driven," making the formulation of bidding strategies more scientific.

[0059] The bidding results output by this module are simultaneously fed back to the feedback weight scoring unit of the intelligent recommendation module, forming a data closed loop and continuously optimizing the recommendation quality.

[0060] The system is designed with a dedicated data synchronization and caching unit to ensure data consistency among multiple modules.

[0061] The Canal component is used to monitor the MySQL binlog and capture CRUD (Create, Read, Update) events of business data. Change events are sent to the cache synchronization service via a message queue. This service is responsible for updating hot data in the Redis cluster and broadcasting change notifications to each microservice node through Redis publish / subscribe channels. Upon receiving the notification, each microservice node updates its local CaffeineL1 cache.

[0062] For high-frequency query scenarios (such as document retrieval), a cache preheating script was written to automatically load frequently accessed data into the cache during system startup or off-peak hours. The multi-level caching design reduces the response time for a single document retrieval to less than 1 second, and high-frequency retrievals can be compressed to within 0.5 seconds. System stability has been verified through stress testing with tens of millions of data points.

[0063] The system constructs a security system of "triple protection + dynamic adaptation": The first layer of transmission and storage encryption: Sensitive fields in the database (such as quotation amount and qualification certificate number) are encrypted and stored using the AES-256 encryption algorithm; the transmission layer uses the TLS protocol to ensure communication security.

[0064] The second-tier, three-level permission hierarchy: Based on the RBAC model, this extends the three-level permission system. System administrators have full permissions, department heads can manage their department's data, and ordinary users can only operate functions within their authorized scope. It also implements row-level data filtering to ensure controllable data access granularity for different roles within the same department.

[0065] The third layer of auditing and early warning: All user operations are recorded in real time through the ELK (Elasticsearch, Logstash, Kibana) log collection system, forming an immutable audit log of operational behavior. An isolated forest anomaly detection model is trained to perform real-time analysis of the log stream, automatically issuing alerts when violations such as data tampering and unauthorized access are detected.

[0066] This invention achieves intelligent recommendation of bidding materials by integrating collaborative filtering scoring, content matching scoring, and feedback weight scoring through a hybrid recommendation model, reducing manual retrieval and repetitive compilation workload. The recommendation results are dynamically optimized based on adoption rate and bidding results, forming a continuously evolving intelligent reuse capability. Named entity recognition automatically detects the risk of expired qualifications, knowledge graph matching identifies parameter conflicts, and an ensemble learning model outputs a comprehensive risk level, significantly improving efficiency and reducing the error rate of bidding materials compared to manual review. The CRDT algorithm enables real-time collaborative editing by multiple users, automatically resolving conflicts; the approval process forms a directed acyclic graph version record, supporting one-click retrospection of any historical node, solving the problems of information fragmentation and traceability difficulties in traditional collaboration. A semantic similarity model automatically locates verification rules affected by policy changes, calling the rule engine for hot updates, controlling policy response time within 24 hours, and avoiding invalid bids due to compliance delays. A bidding probability prediction model is established through big data mining, using risk identification results as input features to output pricing suggestions and qualification configuration schemes, realizing a transformation from experience-driven to data-driven approaches. The adoption results, winning bid results, risk identification results, and other data form a closed-loop feedback between modules. The system continuously optimizes itself as business data accumulates, resulting in long-term value increase.

[0067] Figure 2 A schematic diagram illustrating the principle of a smart cloud projection method provided in the embodiments of this specification includes: S110: Obtain the requirements information of new bidding projects and construct project feature vectors; S120: Input the project feature vector into the hybrid recommendation model, calculate the matching degree with each historical document in the bidding data database, generate a candidate data recommendation list and push it to the user interface; S130: Receive the tender document prepared by the user based on the recommended list, start the risk identification engine, perform multi-dimensional risk identification on the candidate data and the tender document, and output the risk items and risk levels; S140: In the collaborative workbench, the risk items are displayed in association on the editing interface of the tender document, and the editing operation and risk information are synchronized in real time during the user's collaborative editing process; S150: Monitor external policy release sources and dynamically update the verification rules referenced by the risk identification engine when a change in compliance requirements is detected, so that the updated rules take effect in the next risk identification.

[0068] Optionally, S120 includes: Using an implicit semantic model decomposed by alternating least squares, a collaborative filtering score is calculated between the project feature vector and the historical winning project vector. A pre-trained semantic matching model is used to calculate the semantic similarity between the tender requirements text and the historical data summary text, which is then used as the content matching score. The feedback weight score is calculated based on the adoption rate of historical recommendation materials and subsequent bidding results; The collaborative filtering score, content matching score, and feedback weight score are merged with adjustable weights, sorted, and then a candidate material recommendation list is generated.

[0069] The functionality of the method in this embodiment has been described in the above system embodiments. Therefore, for any details not covered in this embodiment, please refer to the relevant descriptions in the foregoing embodiments, which will not be repeated here.

[0070] Based on the same inventive concept, embodiments of this specification also provide an electronic device.

[0071] The following describes embodiments of the electronic device of the present invention, which can be considered as specific implementations of the methods and apparatus embodiments of the present invention described above. Details described in the embodiments of the electronic device of the present invention should be considered as supplements to the methods or apparatus embodiments described above; details not disclosed in the embodiments of the electronic device of the present invention can be implemented with reference to the methods or apparatus embodiments described above.

[0072] Figure 3 This is a schematic diagram of an electronic device provided as an embodiment of this specification. Refer to the following... Figure 3 The electronic device 300 according to this embodiment of the present invention will be described. Figure 3 The electronic device 300 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0073] like Figure 3 As shown, the electronic device 300 is presented in the form of a general-purpose computing device. The components of the electronic device 300 may include, but are not limited to: at least one processing unit 310, at least one storage unit 320, a bus 330 connecting different system components (including storage unit 320 and processing unit 310), a display unit 340, etc.

[0074] The storage unit stores program code that can be executed by the processing unit 310, causing the processing unit 310 to perform the steps described in the processing method section of this specification according to various exemplary embodiments of the present invention. For example, the processing unit 310 can perform, for example... Figure 1 The steps are shown.

[0075] The storage unit 320 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 3201 and / or a cache storage unit 3202, and may further include a read-only memory unit (ROM) 3203.

[0076] The storage unit 320 may also include a program / utility 3204 having a set (at least one) program module 3205, such program module 3205 including but not limited to: an operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0077] Bus 330 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.

[0078] Electronic device 300 can also communicate with one or more external devices 400 (e.g., keyboard, pointing device, Bluetooth device, etc.), and with one or more devices that enable viewers to interact with electronic device 300, and / or with any device that enables electronic device 300 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 350. Furthermore, electronic device 300 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 360. Network adapter 360 can communicate with other modules of electronic device 300 via bus 330. It should be understood that, although... Figure 3 As not shown, other hardware and / or software modules may be used in conjunction with electronic device 300, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0079] Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described in this invention can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this invention can be embodied in the form of a software product, which can be stored in a computer-readable storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, or network device, etc.) to execute the method described above according to this invention. When the computer program is executed by a data processing device, it enables the computer-readable medium to implement the method described above, i.e.: as... Figure 1 The method shown.

[0080] Figure 4 This is a schematic diagram of a computer-readable medium provided for embodiments of this specification.

[0081] accomplish Figure 1 The computer program of the method shown can be stored on one or more computer-readable media. A computer-readable medium can be a readable signal medium or a readable storage medium. A readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0082] The computer-readable storage medium may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The readable storage medium may also be any readable medium other than a readable storage medium, capable of transmitting, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0083] Program code for performing the operations of this invention can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the audience's computing device, partially on the audience's device, as a standalone software package, partially on the audience's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the audience's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0084] In summary, the present invention can be implemented in hardware, or as software modules running on one or more processors, or a combination thereof. Those skilled in the art will understand that in practice, general-purpose data processing devices such as microprocessors or digital signal processors (DSPs) can be used to implement some or all of the functions of some or all of the components according to the embodiments of the present invention. The present invention can also be implemented as a device or apparatus program (e.g., a computer program and computer program product) for performing part or all of the methods described herein. Such programs implementing the present invention can be stored on a computer-readable medium or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0085] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the present invention is not inherently related to any specific computer, virtual device, or electronic device, and various general-purpose devices can also implement the present invention. The above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0086] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0087] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A smart cloud projection system, characterized in that, include: A bidding database is used to store historical bidding documents and their feature annotation information, which includes at least project type, bidding party attributes, technical parameter tags, and historical bidding results. The intelligent recommendation module is used to obtain the demand information of new bidding projects and construct project feature vectors. Based on the hybrid recommendation model, it calculates the matching degree between the project feature vectors and each historical bidding document in the bidding data database, and generates a candidate data recommendation list. The risk identification engine is used to automatically obtain the candidate material recommendation list generated by the intelligent recommendation module and the tender document prepared by the user based on the recommendation list, and to automatically identify the risks of the candidate materials and the tender document in multiple dimensions, and output the risk items and risk levels. The collaborative workbench is used to provide real-time collaborative editing of the tender documents by multiple users. The risk items output by the risk identification engine are displayed in association on the editing interface, and the approval process and version traceability are persisted in a directed acyclic graph structure. The compliance dynamic adaptation module is used to monitor external policy release sources, capture compliance requirement change events, identify the affected verification rules and update them dynamically. The updated verification rules directly apply to the identification logic of the risk identification engine.

2. The intelligent cloud projection system as described in claim 1, characterized in that, The hybrid recommendation model includes: The collaborative filtering scoring unit calculates the similarity between the new bidding project and the historical projects based on the latent semantic model of the historical winning bids. The content matching scoring unit uses a semantic matching model to calculate the semantic similarity between the tender requirements text and the historical data text; The feedback weighting scoring unit dynamically adjusts the weight of materials based on the adoption status of historical recommended materials and the bidding results of the adopted projects; The fusion unit integrates the collaborative filtering score, content matching score, and feedback weight score according to adjustable parameters to generate a final recommendation score and output the candidate material recommendation list in a sorted manner.

3. The intelligent cloud projection system as described in claim 1, characterized in that, The risk identification engine includes: The qualification validity detection unit uses a pre-trained named entity recognition model to perform sequence annotation on the OCR text of the qualification certificate images in the candidate material recommendation list, extracts the validity period entity and compares it with the current date to identify the risk of qualification expiration. The parameter conflict detection unit performs graph matching between the structured conditions in the tender document and the technical parameters in the bid document, and identifies parameter conflict risks based on a preset bid knowledge graph. The comprehensive scoring unit extracts multi-dimensional features, including qualification completeness and parameter conflict number, and inputs them into an ensemble learning scoring model trained with historical bidding data to output a comprehensive risk score and warning level.

4. The intelligent cloud projection system as described in claim 1, characterized in that, The collaborative workbench is specifically used for: Based on the conflict-free copy data type algorithm, each user's editing operation is encapsulated into atomic operations with unique identifiers and broadcast. The reception status is tracked through state vectors, and conflicts are resolved according to a preset strategy. In the editing interface, the risk items marked by the risk identification engine are highlighted and risk details are displayed; A document status snapshot is generated after each approval node is completed. All snapshots form the directed acyclic graph structure, which supports previewing and backtracking of any historical version.

5. The intelligent cloud projection system as described in claim 1, characterized in that, The compliance dynamic adaptation module is specifically used for: By comparing the obtained policy clauses with the existing verification rule descriptions in the system using a semantic similarity model, the affected rule nodes can be located. The rule engine is invoked to match rule templates from the configuration center, automatically instantiate new rules and load them into the runtime cache, while marking old rules as invalid, thus completing hot updates; The updated verification rules will automatically take effect the next time the risk identification engine is started.

6. The intelligent cloud projection system as described in claim 2, characterized in that, include: The feedback weighting scoring unit is based on the winning bid results of the adopted projects, which are provided by the strategy analysis module. The strategy analysis module is also used to: extract the price weight, technology weight and competitor influence factor of historical winning bids, and construct a winning bid probability prediction model; For the current bidding project, the risk items and risk levels output by the risk identification engine are used as one of the input features to generate a suggested price range and qualification configuration scheme.

7. A smart cloud projection method, characterized in that, include: Obtain the requirements information for new bidding projects and construct project feature vectors; The project feature vector is input into the hybrid recommendation model to calculate the matching degree with each historical document in the bidding data database, generate a candidate data recommendation list and push it to the user interface; The system receives tender documents prepared by users based on the recommended list, activates the risk identification engine, performs multi-dimensional risk identification on the candidate materials and the tender documents, and outputs risk items and risk levels. In the collaborative workbench, the aforementioned risk items are displayed in association with the editing interface of the tender document, and the editing operations and risk information are synchronized in real time during the user's collaborative editing process; By monitoring external policy release sources, when changes in compliance requirements are detected, the verification rules referenced by the risk identification engine are dynamically updated so that the updated rules take effect in the next risk identification.

8. The intelligent cloud projection method as described in claim 7, characterized in that, The step of inputting the project feature vector into the hybrid recommendation model and calculating the matching degree with each historical document in the bidding database includes: Using an implicit semantic model decomposed by alternating least squares, a collaborative filtering score is calculated between the project feature vector and the historical winning project vector. A pre-trained semantic matching model is used to calculate the semantic similarity between the tender requirements text and the historical data summary text, which is then used as the content matching score. The feedback weight score is calculated based on the adoption rate of historical recommendation materials and subsequent bidding results; The collaborative filtering score, content matching score, and feedback weight score are merged with adjustable weights, sorted, and then a candidate material recommendation list is generated.

9. An electronic device, wherein, The electronic device includes: A processor; and a memory storing computer-executable instructions, which, when executed, cause the processor to perform the method according to any one of claims 7-8.

10. A computer-readable storage medium, wherein, The computer-readable storage medium stores one or more programs that, when executed by a processor, implement the method of any one of claims 7-8.

Images