Bid inviting and tendering risk early warning system and method based on large model

Abstract

The invention discloses a bidding and tendering risk early warning system and method based on a large model, and relates to the technical field of bidding and tendering, and the main points of the technical scheme comprise the following steps: obtaining basic information data and historical bidding and tendering data of a bidding and tendering project; generating a project feature map of the bidding project according to the basic information data, and obtaining a project risk feature library according to the project feature map; collecting a risk identification range of the early warning large model, and carrying out feature matching according to the risk identification range and a project risk feature library to obtain a to-be-early-warned risk point corresponding to the bidding project; according to historical bidding and tendering data of the bidding and tendering project, obtaining a risk association track of the risk point to be pre-warned and the bidding project, and according to the risk association track, obtaining a risk association coefficient corresponding to the risk point to be pre-warned; setting risk early warning parameters of the bidding and tendering project according to the basic information data; the method has the effect of building a risk prevention and control barrier for smooth development of bidding activities.

Description

Technical Field

[0001] The present invention relates to the technical field of bidding and tendering, and more specifically, to a bidding and tendering risk early warning system and method based on a large model. Background Art

[0002] In bidding activities, due to the involvement of multiple subjects and complex processes, traditional risk control models are gradually unable to adapt to the needs of refined management. Traditional methods have a low degree of risk quantification and are mostly vaguely graded as "high, medium, and low". They are unable to accurately distinguish the threat levels of different risk points, resulting in resource mismatch. It is easy to invest too much energy in risks with small actual threats, while key high-impact risks lead to problems such as inconsistent bidding documents and contract performance disputes due to insufficient attention.

[0003] At the same time, risk identification relies on manual experience to identify isolated risk points, lacking in-depth exploration of risk relevance. For example, focusing solely on the compliance of a current project bidder's qualification documents without linking their historical bid-rigging and bid-collusion records, or unusual collaborations with agencies, can make hidden risks difficult to detect. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide a bidding risk early warning system and method based on a large model.

[0005] To achieve the above object, the present invention provides the following technical solutions: A bidding risk early warning method based on a large model, the method comprising the following steps: Obtain basic information data and historical bidding data of bidding projects; generate project feature maps of bidding projects based on the basic information data, and obtain a project risk feature library based on the project feature maps; Collect the risk identification scope of the early warning model, and perform feature matching based on the risk identification scope and the project risk feature library to obtain the corresponding risk points to be warned for the bidding project; Based on the historical bidding data of the bidding projects, the risk correlation trajectory between the risk points to be warned and the bidding projects is obtained, and the risk correlation coefficient corresponding to the risk points to be warned is obtained based on the risk correlation trajectory; Set risk warning parameters for bidding projects based on basic information data; obtain historical risk parameters of the risk points to be warned, and obtain the risk warning coefficient corresponding to the risk points to be warned based on the risk warning parameters and the historical risk parameters of the risk points to be warned; Obtain the comprehensive risk coefficient corresponding to the risk point to be warned based on the risk correlation coefficient and risk impact coefficient; The target warning risk points corresponding to the bidding project are obtained based on the comprehensive risk coefficient, and the risk warning information of the bidding project is generated based on the target warning risk points.

[0006] Preferably, generating a project feature map of the bidding project based on the basic information data, and obtaining a project risk feature library based on the project feature map, specifically includes the following steps: Collect the first time point of the bidding project's start of bidding and the second time point of the bidding project's end of bid evaluation; Obtaining a project cycle of the bidding project based on the first time node and the second time node, and generating a project feature map of the bidding project based on basic information within the project cycle; The attributes of each feature point in the project feature map are obtained, including key feature attributes and secondary feature attributes. Feature classification standards are set based on the key feature attributes and secondary feature attributes, and a project risk feature library is constructed based on the feature classification standards.

[0007] Preferably, feature matching is performed based on the risk identification scope and the project risk feature library to obtain the corresponding risk points to be warned for the bidding project, specifically: If the risk identification scope of the early warning model includes the risk features in the project risk feature library, the risk points corresponding to the risk features will be marked as the risk points to be warned corresponding to the bidding project.

[0008] Preferably, obtaining the risk correlation trajectory between the risk points to be warned and the bidding projects based on the historical bidding data of the bidding projects specifically includes the following steps: Obtain the historical risk location of the risk point to be warned; obtain the event location of the historical risk event based on the historical bidding data of the bidding project; Obtain the risk correlation between historical risk events and risk points to be warned based on historical risk locations and event locations; Based on the risk correlation between historical risk events and risk points to be warned, a risk correlation trajectory between risk points to be warned and bidding projects is generated.

[0009] Preferably, obtaining the risk correlation coefficient corresponding to the risk point to be warned based on the risk correlation trajectory specifically includes the following steps: Setting risk association intervals, each of which corresponds to a risk weight; According to the inclusion relationship between the risk correlation degree and the risk correlation interval in the risk correlation trajectory, the risk correlation trajectory is divided to obtain at least one segmented risk correlation trajectory; Extract target risk weights of segmented risk association trajectories; The risk correlation coefficient corresponding to the risk point to be warned is obtained based on the segmented risk correlation trajectory and the target risk weight corresponding to the segmented risk correlation trajectory.

[0010] Preferably, the risk warning parameters of the bidding project are set based on the basic information data, specifically including the following steps: Obtain the project type and project scale of the bidding project based on basic information data; Setting risk warning parameters for bidding projects based on the project type and project scale; Among them, the risk warning parameters include the risk probability threshold of the bidding project, the risk impact threshold, the risk response difficulty threshold and the risk monitoring frequency.

[0011] Preferably, obtaining historical risk parameters of the risk point to be warned, and analyzing the risk warning parameters and the historical risk parameters of the risk point to be warned to obtain a risk warning coefficient corresponding to the risk point to be warned, specifically includes the following steps: The historical risk parameters include the historical probability of occurrence, historical impact, historical response difficulty and historical monitoring records of the risk point to be warned; Setting risk parameter weights; wherein the risk parameter weights include probability weight, impact weight, difficulty weight, and monitoring weight; Compare the risk warning parameters of the bidding project with the historical risk parameters of the risk point to be warned to obtain the parameter adaptation risk point of the risk point to be warned; The probability coefficient of the parameter adaptation risk point is obtained based on the probability weight, the risk occurrence probability threshold and the historical occurrence probability of the parameter adaptation risk point; The influence coefficient of the parameter adaptation risk point is obtained based on the influence weight, risk impact threshold and historical impact of the parameter adaptation risk point; The difficulty coefficient of the parameter adaptation risk point is obtained based on the difficulty weight, the risk response difficulty threshold, and the historical response difficulty of the parameter adaptation risk point; The monitoring coefficient of the parameter adaptation risk point is obtained based on the monitoring weight, risk monitoring frequency and historical monitoring records of the parameter adaptation risk point; The risk warning coefficient corresponding to the risk point to be warned is obtained based on the probability coefficient, impact coefficient, difficulty coefficient and monitoring coefficient.

[0012] Preferably, the target warning risk points corresponding to the bidding project are obtained based on the comprehensive risk coefficient, specifically: The risk point to be warned with the largest comprehensive risk coefficient is marked as the target warning risk point corresponding to the bidding project.

[0013] A bidding risk early warning system based on a large model, comprising: Acquisition module: acquires basic information data and historical bidding data of bidding projects; generates project feature maps of bidding projects based on basic information data, and obtains project risk feature libraries based on the project feature maps; Acquisition and matching module: collects the risk identification scope of the early warning model, performs feature matching based on the risk identification scope and the project risk feature library to obtain the corresponding risk points to be warned for the bidding project; The first processing module: obtains the risk correlation trajectory between the risk point to be warned and the bidding project based on the historical bidding data of the bidding project, and obtains the risk correlation coefficient corresponding to the risk point to be warned based on the risk correlation trajectory; Analysis module: Set risk warning parameters for bidding projects based on basic information data; obtain historical risk parameters of risk points to be warned, and obtain the risk warning coefficient corresponding to the risk points to be warned based on the risk warning parameters and the historical risk parameters of the risk points to be warned; The second processing module: obtains the comprehensive risk coefficient corresponding to the risk point to be warned based on the risk correlation coefficient and the risk impact coefficient; Risk warning module: obtains the target warning risk points corresponding to the bidding project based on the comprehensive risk coefficient, and generates risk warning information for the bidding project based on the target warning risk points.

[0014] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, a bidding risk warning method based on a large model is implemented.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present invention comprehensively measures the threat level of the risk points to be warned by calculating the comprehensive risk coefficient through comprehensive consideration of risk association trajectories, historical parameters, etc. After clarifying the target warning risk points in resource allocation optimization, project participants can concentrate manpower and material resources to prioritize the prevention and control of this risk, avoiding the dispersion of resources on secondary risks. For example, special plans can be formulated in advance for high-coefficient risk points, the monitoring frequency can be increased, the efficiency of resource utilization can be improved, and the cost of prevention and control can be reduced. From the perspective of project promotion and guarantee, focusing on the risk points with the largest comprehensive risk coefficient can curb the fermentation of major risks in advance, reduce their impact on the bidding process (such as bid opening and bid evaluation) and the subsequent implementation of the project (such as performance and cost), ensure that the project is promoted as planned, maintain the fairness and standardization of bidding activities, build a solid risk prevention and control barrier for the smooth development of bidding activities, and improve the overall risk response level of the industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of the steps of a bidding risk early warning method based on a large model is proposed in the present invention; Figure 2 The present invention proposes a module diagram of a bidding risk early warning system based on a large model; Figure 3 It is a structural diagram of an electronic device provided by an embodiment of the present invention.

[0017] 610 , processor; 620 , communication interface; 630 , memory; 640 , communication bus. DETAILED DESCRIPTION

[0018] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0019] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0020] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it constitute a separate or selective embodiment that is mutually exclusive with other embodiments.

[0021] Reference Figure 1-Figure 3 .

[0022] Example 1 further illustrates a bidding risk early warning system and method based on a large model proposed by the present invention.

[0023] A bidding risk early warning method based on a large model, the method comprising the following steps: Obtain basic information data and historical bidding data of bidding projects; generate project feature maps of bidding projects based on the basic information data, and obtain a project risk feature library based on the project feature maps; Collect the risk identification scope of the early warning model, and perform feature matching based on the risk identification scope and the project risk feature library to obtain the corresponding risk points to be warned for the bidding project; Based on the historical bidding data of the bidding projects, the risk correlation trajectory between the risk points to be warned and the bidding projects is obtained, and the risk correlation coefficient corresponding to the risk points to be warned is obtained based on the risk correlation trajectory; Set risk warning parameters for bidding projects based on basic information data; obtain historical risk parameters of the risk points to be warned, and obtain the risk warning coefficient corresponding to the risk points to be warned based on the risk warning parameters and the historical risk parameters of the risk points to be warned; Obtain the comprehensive risk coefficient corresponding to the risk point to be warned based on the risk correlation coefficient and risk impact coefficient; The target warning risk points corresponding to the bidding project are obtained based on the comprehensive risk coefficient, and the risk warning information of the bidding project is generated based on the target warning risk points.

[0024] This application builds a complete risk warning process based on data and supported by large models. First, by obtaining basic information data of bidding projects (covering basic project information, participant information, etc.) and historical bidding data (various records of past projects), we use this basic information data to generate project feature maps, sort out key project characteristics, and then extract a project risk feature library from this, classifying and storing potential project risks in the form of characteristics.

[0025] This application captures comprehensive data details across the entire bidding process. For basic data, it focuses on the current bidding project being analyzed, collecting essential attributes such as project name, industry (whether it's construction, goods procurement, or services), budget, scope (list of equipment to be procured, service scope), and bidding method (open or invited). It also collects information on participating entities, such as the tenderer's corporate qualifications and credit rating, potential bidders' business scope, past participation in similar projects, and key milestones in the bidding process (such as application deadlines, bid opening locations, and composition of bid evaluation experts). Historical bidding data, on similar or related projects, is retrieved from enterprise or industry project databases. This data includes historical risk events (such as bid rigging and contract performance disputes), historical behavior of participants (a bidder's past bidding strategies and the number of defaults after winning a bid), historical project outcomes (successful delivery, acceptance issues), and industry cyclical data (price fluctuations and high-risk periods for similar projects over different years). This data is obtained through automatic capture via public resource trading platforms, export from internal enterprise project management systems, and manual verification of historical archives.

[0026] First, core characteristic entities are extracted from the basic data, such as the project entity (including attributes such as project budget and bidding type), participating entities (tenderer, bidder, and agency, each with its own qualifications and credit labels), and process nodes (stages such as registration, bid opening, and bid evaluation, along with time attributes). The relationships between these entities are then analyzed, such as the entrustment relationship between "tenderer-entruster-agency," the participation relationship between "bidder-participation-project section," and the potential risk associations between "process node-association-risk history." Knowledge graph technology (such as graph database tools) is then used to construct a visual project feature map, using these entities as "nodes" and relationships as "edges." For example, the "Project A" node is connected to "Tenderer A," "Bidder B," and "Bidder B" is, through the "Qualification" relationship, associated with the "Grade B Qualification" label. The "Bid Opening Stage" node is associated with the risk alert "Anomalies in Historical Project Bid Openings."

[0027] Building a project risk signature database based on project feature maps is a process of uncovering risk patterns. Analyzing abnormal signals within node attributes, for example, for example, "Bidder B's qualifications are Class B," but the project requires Class A, identifies a "qualification failure risk signature." For example, if "the project budget is 30% lower than similar projects," this corresponds to a "quality risk signature due to insufficient costs." Risks are identified within the relationship network, using entity associations to identify potential risks. For example, if "Agency C and Bidder B have three joint bid-winning records," this can be combined with industry examples of bid-rigging to identify a "related-party bid-rigging risk signature." If "the bid opening is scheduled outside business hours," this can be linked to "process non-compliance risks" in historical projects to identify a "time compliance risk signature." These risk signatures are then standardized and stored, categorized by risk types such as compliance, economics, and contract performance. Each signature is defined with its "name, trigger condition, and associated entity" (for example, the "bid-rigging risk" signature is triggered by "the agency and bidder have collaborated at least twice with a bid price difference of ≤5%" and is associated with the "agency, bidder" entities). This forms the project risk signature database.

[0028] The risk identification scope of the early warning model is collected, and the features in the project risk feature library are matched with this identification scope. If the risk features identifiable by the model have corresponding risk points in the library, they are marked as risk points to be warned. The recognition capabilities of the large model are then used to preliminarily screen out risks that require attention. Based on historical bidding data, the association between the risk points to be warned and the project is mined. The historical risk locations of the risk points to be warned are obtained. The risk correlation degree is calculated based on the event locations of historical risk events to generate a risk correlation trajectory. The risk correlation coefficient of the risk point to be warned is then obtained by dividing the correlation trajectory interval and extracting the corresponding risk weight, quantifying the degree of association between the risk point and the project.

[0029] Risk warning parameters (including risk probability thresholds and impact thresholds) are set based on the project type and scale in the basic information data. Historical risk parameters (such as historical probability of occurrence and impact) for the risk point to be warned are obtained. The corresponding parameters are compared and combined with pre-set risk parameter weights such as probability weights and impact weights to calculate sub-item coefficients such as probability coefficients and impact coefficients. The final summary results in a risk warning coefficient, which measures risk based on the project's own requirements and historical risk performance. The risk correlation coefficient and risk impact coefficient are combined to obtain a comprehensive risk coefficient for the risk point to be warned. The risk is assessed by comprehensively considering the correlation and impact.

[0030] From all the risk points to be warned, the one with the largest comprehensive risk coefficient is marked as the target warning risk point. Specific risk warning information is generated for this risk point to inform the bidding participants of the risks that they need to pay attention to and prevent, helping them to take measures in advance to avoid or reduce risks and ensure the smooth progress of the bidding project.

[0031] Generate a project feature map of the bidding project based on basic information data, and obtain a project risk feature library based on the project feature map, specifically including the following steps: Collect the first time point of the bidding project's start of bidding and the second time point of the bidding project's end of bid evaluation; Obtaining a project cycle of the bidding project based on the first time node and the second time node, and generating a project feature map of the bidding project based on basic information within the project cycle; Obtain the attributes of each feature point in the project feature map, including key feature attributes and secondary feature attributes, set feature classification standards based on the key feature attributes and secondary feature attributes, and build a project risk feature library based on the feature classification standards.

[0032] After calculating the project cycle based on the first and second time nodes, a feature graph is generated. Based on the two determined time nodes, the duration and phase distribution of the project from the start of bidding to the end of bid evaluation are calculated. All basic information within the cycle (project budget, participant information, process requirements for each phase, etc.) is converted into feature points and relationships in the graph along the timeline and business node dimensions. For example, in the graph, the registration phase node is associated with attributes such as the registration deadline and the number of potential bidders, while the bidder A node is associated with features such as qualification level and past bidding records.

[0033] From the characteristic points in the graph, we distinguish between key attribute features (core factors that directly impact project risk, such as whether the bidder's qualifications meet the standards and the deviation rate between the project budget and the market price) and secondary attribute features (information that has a lesser impact on risk or serves as auxiliary explanation, such as the bidder's office address and project promotional slogan). The information in these secondary attribute features serves as supplementary explanations or related clues to the risk characteristics. Based on this standard, various risk-related features mined from the graph (such as "delay at a certain stage time node" corresponding to "process risk characteristics") are standardized, organized, and stored to form a project risk feature library.

[0034] Based on the risk identification scope and the project risk feature database, feature matching is performed to obtain the corresponding risk points to be warned for the bidding project, specifically: If the risk identification scope of the early warning model includes the risk features in the project risk feature library, the risk points corresponding to the risk features will be marked as the risk points to be warned corresponding to the bidding project.

[0035] This application must clearly state that the early warning model has pre-set the risk identification scope, which covers all types of risk characteristics that may appear in bidding projects. At the same time, a project risk feature library is constructed through the early processing of the basic information of the bidding projects. The project risk feature library stores feature content related to the potential risks of the project extracted from the project feature map.

[0036] When identifying risk points requiring early warning, the risk identification scope of the early warning model is compared with the project risk signature database. If a risk signature identified by the early warning model matches a corresponding entry in the project risk signature database, this indicates that the risk point identified by the risk signature is related to the current bidding project. At this point, the risk points corresponding to these risk signatures are marked as requiring special attention and requiring early warning for the current bidding project.

[0037] Based on the historical bidding data of bidding projects, the risk correlation trajectory between the risk points to be warned and the bidding projects is obtained, which specifically includes the following steps: Obtain the historical risk location of the risk point to be warned; obtain the event location of the historical risk event based on the historical bidding data of the bidding project; Obtain the risk correlation between historical risk events and risk points to be warned based on historical risk locations and event locations; Based on the risk correlation between historical risk events and risk points to be warned, a risk correlation trajectory between risk points to be warned and bidding projects is generated.

[0038] This application explores the logical correlation between potential risk points and bidding projects, conducting analysis based on historical data. The historical bidding data identifies the historical risk locations of past occurrences of potential risk points. This could be project process stages (such as tender registration and bid evaluation), time points (such as a quarter or month), or participant-related scenarios (collaboration with specific bidders or agencies). The application also searches for the locations of past real risk events within historical data, also defining them based on process, time, and participant dimensions.

[0039] Based on these two types of location information, the "risk correlation" between historical risk events and the risk points to be warned is determined through spatial or logical association. For example, if the historical risk location of the risk point to be warned frequently overlaps with the location of a certain type of historical risk event, or if they are highly connected in time or process, then the correlation is high, quantifying the closeness of the connection between the two.

[0040] Based on this risk correlation data, we analyze the correlation patterns and transmission paths between the potential risk points and various historical risk events throughout the bidding process, generating a risk correlation trajectory. This clearly illustrates the potential for risk events to be triggered or associated with the potential risk points at different project stages, providing an intuitive correlation logic for subsequent risk impact assessment and prevention and control strategy formulation.

[0041] And according to the risk correlation trajectory, the risk correlation coefficient corresponding to the risk point to be warned is obtained, which specifically includes the following steps: Set risk association intervals, each of which corresponds to a risk weight; According to the inclusion relationship between the risk correlation degree and the risk correlation interval in the risk correlation trajectory, the risk correlation trajectory is divided to obtain at least one segmented risk correlation trajectory; Extract target risk weights of segmented risk association trajectories; The risk correlation coefficient corresponding to the risk point to be warned is obtained based on the segmented risk correlation trajectory and the target risk weight corresponding to the segmented risk correlation trajectory.

[0042] This application obtains the risk correlation coefficient of the risk point to be warned by quantifying the risk correlation trajectory to measure the degree of risk correlation. First, the risk correlation interval is set in advance, and a corresponding risk weight is assigned to each interval. These intervals and weights are determined based on historical experience, industry standards or data analysis, and are used for subsequent classification and quantification of the correlation degree. Next, the risk correlation degree in the risk correlation trajectory is determined, and it is judged which preset interval it falls into. The complete risk correlation trajectory is divided into multiple segmented trajectories according to the inclusion relationship, and each segment corresponds to a specific interval. Then, the corresponding target risk weight is extracted from the divided segmented risk correlation trajectory to clarify the weight of each segment trajectory. Finally, combined with the specific conditions of the segmented risk correlation trajectory (such as length, coverage stage, etc.) and the corresponding target risk weight, the risk correlation coefficient of the risk point to be warned is comprehensively obtained through weighted calculation and other methods, and the complex trajectory of risk correlation is converted into a quantitative indicator that can be used for risk assessment, providing data support for subsequent risk analysis and warning.

[0043] And set the risk warning parameters of the bidding project based on the basic information data, which specifically includes the following steps: Obtain the project type and project scale of the bidding project based on basic information data; Set risk warning parameters for bidding projects based on project type and scale; Among them, the risk warning parameters include the risk probability threshold of the bidding project, the risk impact threshold, the risk response difficulty threshold and the risk monitoring frequency.

[0044] This application configures risk warning parameters based on basic project information, laying a solid foundation for bidding and tendering risk management. First, we mine the basic information data of the bidding and tendering projects to clearly define the project type (such as engineering construction, goods procurement, service outsourcing, and other categories) and the project size (measured by indicators such as investment amount and business volume). This forms the core basis for subsequent parameter setting, as projects of different types and sizes have significant differences in risk performance and management requirements.

[0045] Risk warning parameters are set based on the clear project type and scale. Due to different project characteristics, the threshold standards for risk probability, impact, and difficulty of response will be different. For example, for large-scale engineering construction projects, the risk probability threshold may be relatively low because there are many links and complex participants. The risk impact threshold will also be set according to the impact of the risk on the entire project once it occurs; the risk response difficulty threshold is determined based on the professional characteristics of the project type, and the difficulty standards for handling various risks are clarified; the risk monitoring frequency is reasonably planned based on the project scale and potential risk threats. The monitoring frequency will naturally be more intensive for large-scale and high-risk projects. In this way, quantitative standards and monitoring rhythms for risk warnings are constructed for bidding projects in an all-round and adaptive manner, so that subsequent risk identification, assessment, and response have clear parameter guidance.

[0046] Obtain historical risk parameters of the risk point to be warned, and analyze the risk warning parameters and the historical risk parameters of the risk point to be warned to obtain the risk warning coefficient corresponding to the risk point to be warned, specifically including the following steps: Historical risk parameters include the historical probability of occurrence, historical impact, historical response difficulty, and historical monitoring records of the risk point to be warned; Set risk parameter weights; risk parameter weights include probability weight, impact weight, difficulty weight, and monitoring weight; Compare the risk warning parameters of the bidding project with the historical risk parameters of the risk point to be warned to obtain the parameter adaptation risk point of the risk point to be warned; The probability coefficient of the parameter adaptation risk point is obtained based on the probability weight, the risk occurrence probability threshold and the historical occurrence probability of the parameter adaptation risk point; The influence coefficient of the parameter adaptation risk point is obtained based on the influence weight, risk impact threshold and historical impact of the parameter adaptation risk point; The difficulty coefficient of the parameter adaptation risk point is obtained based on the difficulty weight, the risk response difficulty threshold, and the historical response difficulty of the parameter adaptation risk point; The monitoring coefficient of the parameter adaptation risk point is obtained based on the monitoring weight, risk monitoring frequency and historical monitoring records of the parameter adaptation risk point; The risk warning coefficient corresponding to the risk point to be warned is obtained based on the probability coefficient, impact coefficient, difficulty coefficient and monitoring coefficient.

[0047] This application extracts key historical occurrence probabilities, historical impacts, historical response difficulties, and historical monitoring records from historical information about the risk points to be warned. To reasonably measure the impact of historical data from different dimensions on the current risk warning, it is necessary to set risk parameter weights. This means assigning weights to the probability, impact, difficulty, and monitoring dimensions. The weightings are determined based on factors such as the actual project needs and industry characteristics, highlighting the importance of different dimensions in risk assessment.

[0048] The pre-set risk warning parameters of the current bidding project (such as risk probability threshold, risk impact threshold, etc.) are compared one by one with the historical risk parameters of the risk points to be warned to find the risk points where the parameters are adapted, clarify the matching status of historical risk data under the current project risk management standards, and determine which historical risk characteristics are related to the current project warning requirements.

[0049] After obtaining the parameter-adapted risk point, coefficient calculation is carried out for each dimension. The probability coefficient is calculated using the probability weight, the project's risk probability threshold and the historical probability of occurrence of the adapted risk point to measure the warning level of the risk point's occurrence probability. Similarly, the impact coefficient is obtained through the impact weight, the risk impact threshold and the historical impact level to reflect the impact of the risk on the project. The difficulty coefficient is calculated with the help of the difficulty weight, the risk response difficulty threshold and the historical response difficulty to reflect the difficulty of handling the risk. The monitoring coefficient is determined based on the monitoring weight, the risk monitoring frequency and the historical monitoring records to show the perfection of the historical monitoring.

[0050] The coefficients of the four dimensions of probability coefficient, impact coefficient, difficulty coefficient and monitoring coefficient are summarized and the risk warning coefficient of the risk point to be warned is obtained through weighted summation and other operations. The complex characteristics of the risk point are converted into a quantitative value, which intuitively presents the warning level of the risk point in the current project.

[0051] The target warning risk points corresponding to the bidding project are obtained based on the comprehensive risk coefficient, specifically: The risk point to be warned with the largest comprehensive risk coefficient is marked as the target warning risk point corresponding to the bidding project.

[0052] This application obtains a comprehensive risk coefficient for each risk point to be warned through data processing and calculation in the bidding risk warning process. This coefficient integrates multi-dimensional factors such as risk association and impact, and can reflect the degree of threat posed by the risk point to the project. When determining the target warning risk point, by comparing the comprehensive risk coefficients of all risk points to be warned, the risk point with the largest coefficient is selected, because the larger the comprehensive risk coefficient, the more prominent the potential threat of the risk point to the bidding project in terms of probability of occurrence, degree of impact, and closeness of correlation. Marking it as a target warning risk point allows project participants to focus on the most critical risks that require the most priority attention and disposal, accurately allocate resources for prevention and control, improve the pertinence and efficiency of risk warning and management, and ensure that the bidding project proceeds more smoothly.

[0053] Example 2 further illustrates a bidding risk early warning system based on a large model proposed by the present invention.

[0054] A bidding risk early warning system based on a large model, comprising: Acquisition module: acquires basic information data and historical bidding data of bidding projects; generates project feature maps of bidding projects based on basic information data, and obtains project risk feature libraries based on the project feature maps; Acquisition and matching module: collects the risk identification scope of the early warning model, performs feature matching based on the risk identification scope and the project risk feature library to obtain the corresponding risk points to be warned for the bidding project; The first processing module: obtains the risk correlation trajectory between the risk point to be warned and the bidding project based on the historical bidding data of the bidding project, and obtains the risk correlation coefficient corresponding to the risk point to be warned based on the risk correlation trajectory; Analysis module: Set risk warning parameters for bidding projects based on basic information data; obtain historical risk parameters of risk points to be warned, and obtain the risk warning coefficient corresponding to the risk points to be warned based on the risk warning parameters and the historical risk parameters of the risk points to be warned; The second processing module: obtains the comprehensive risk coefficient corresponding to the risk point to be warned based on the risk correlation coefficient and the risk impact coefficient; Risk warning module: obtains the target warning risk points corresponding to the bidding project based on the comprehensive risk coefficient, and generates risk warning information for the bidding project based on the target warning risk points.

[0055] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, a bidding risk warning method based on a large model is implemented.

[0056] like Figure 3 As shown, the electronic device may include: a processor 610, a communications interface 620, a memory 630, and a communication bus 640, wherein the processor 610, the communications interface 620, and the memory 630 communicate with each other via the communication bus 640. The processor 610 may call logic instructions in the memory 630 to execute a large model-based bidding risk early warning method.

[0057] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0058] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0059] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A bidding risk early warning method based on a large model, characterized by: The method comprises the following steps: Obtain basic information data and historical bidding data of bidding projects; generate project feature maps of bidding projects based on the basic information data, and obtain a project risk feature library based on the project feature maps; Collect the risk identification scope of the early warning model, and perform feature matching based on the risk identification scope and the project risk feature library to obtain the corresponding risk points to be warned for the bidding project; Based on the historical bidding data of the bidding projects, the risk correlation trajectory between the risk points to be warned and the bidding projects is obtained, and the risk correlation coefficient corresponding to the risk points to be warned is obtained based on the risk correlation trajectory; Set risk warning parameters for bidding projects based on basic information data; obtain historical risk parameters of the risk points to be warned, and obtain the risk warning coefficient corresponding to the risk points to be warned based on the risk warning parameters and the historical risk parameters of the risk points to be warned; Obtain the comprehensive risk coefficient corresponding to the risk point to be warned based on the risk correlation coefficient and risk impact coefficient; The target warning risk points corresponding to the bidding project are obtained based on the comprehensive risk coefficient, and the risk warning information of the bidding project is generated based on the target warning risk points.

2. A bidding risk early warning method based on a large model according to claim 1, characterized in that: Generate a project feature map of the bidding project based on the basic information data, and obtain a project risk feature library based on the project feature map, specifically including the following steps: Collect the first time point of the bidding project's start of bidding and the second time point of the bidding project's end of bid evaluation; Obtaining a project cycle of the bidding project based on the first time node and the second time node, and generating a project feature map of the bidding project based on basic information within the project cycle; The attributes of each feature point in the project feature map are obtained, including key feature attributes and secondary feature attributes. Feature classification standards are set based on the key feature attributes and secondary feature attributes, and a project risk feature library is constructed based on the feature classification standards.

3. A bidding risk early warning method based on a large model according to claim 2, characterized in that: Based on the risk identification scope and the project risk feature database, feature matching is performed to obtain the corresponding risk points to be warned for the bidding project, specifically: If the risk identification scope of the early warning model includes the risk features in the project risk feature library, the risk points corresponding to the risk features will be marked as the risk points to be warned corresponding to the bidding project.

4. A bidding risk early warning method based on a large model according to claim 3, characterized in that: Based on the historical bidding data of bidding projects, the risk correlation trajectory between the risk points to be warned and the bidding projects is obtained, which specifically includes the following steps: Obtain the historical risk location of the risk point to be warned; obtain the event location of the historical risk event based on the historical bidding data of the bidding project; Obtain the risk correlation between historical risk events and risk points to be warned based on historical risk locations and event locations; Based on the risk correlation between historical risk events and risk points to be warned, a risk correlation trajectory between risk points to be warned and bidding projects is generated.

5. The bidding risk early warning method based on a large model according to claim 4 is characterized in that: And according to the risk correlation trajectory, the risk correlation coefficient corresponding to the risk point to be warned is obtained, which specifically includes the following steps: Setting risk association intervals, each of which corresponds to a risk weight; According to the inclusion relationship between the risk correlation degree and the risk correlation interval in the risk correlation trajectory, the risk correlation trajectory is divided to obtain at least one segmented risk correlation trajectory; Extract target risk weights of segmented risk association trajectories; The risk correlation coefficient corresponding to the risk point to be warned is obtained based on the segmented risk correlation trajectory and the target risk weight corresponding to the segmented risk correlation trajectory.

6. The bidding risk early warning method based on a large model according to claim 5 is characterized in that: And set the risk warning parameters of the bidding project based on the basic information data, which specifically includes the following steps: Obtain the project type and project scale of the bidding project based on basic information data; Setting risk warning parameters for bidding projects based on the project type and project scale; Among them, the risk warning parameters include the risk probability threshold of the bidding project, the risk impact threshold, the risk response difficulty threshold and the risk monitoring frequency.

7. The bidding risk early warning method based on a large model according to claim 6 is characterized in that: Obtain historical risk parameters of the risk point to be warned, and analyze the risk warning parameters and the historical risk parameters of the risk point to be warned to obtain the risk warning coefficient corresponding to the risk point to be warned, specifically including the following steps: The historical risk parameters include the historical probability of occurrence, historical impact, historical response difficulty and historical monitoring records of the risk point to be warned; Setting risk parameter weights; wherein the risk parameter weights include probability weight, impact weight, difficulty weight, and monitoring weight; Compare the risk warning parameters of the bidding project with the historical risk parameters of the risk point to be warned to obtain the parameter adaptation risk point of the risk point to be warned; The probability coefficient of the parameter adaptation risk point is obtained based on the probability weight, the risk occurrence probability threshold and the historical occurrence probability of the parameter adaptation risk point; The influence coefficient of the parameter adaptation risk point is obtained based on the influence weight, risk impact threshold and historical impact of the parameter adaptation risk point; The difficulty coefficient of the parameter adaptation risk point is obtained based on the difficulty weight, the risk response difficulty threshold, and the historical response difficulty of the parameter adaptation risk point; The monitoring coefficient of the parameter adaptation risk point is obtained based on the monitoring weight, risk monitoring frequency and historical monitoring records of the parameter adaptation risk point; The risk warning coefficient corresponding to the risk point to be warned is obtained based on the probability coefficient, impact coefficient, difficulty coefficient and monitoring coefficient.

8. The bidding risk early warning method based on a large model according to claim 7 is characterized in that: The target warning risk points corresponding to the bidding project are obtained based on the comprehensive risk coefficient, specifically: The risk point to be warned with the largest comprehensive risk coefficient is marked as the target warning risk point corresponding to the bidding project.

9. A bidding risk warning system based on a large model, applied to a bidding risk warning method based on a large model according to any one of claims 1 to 8, characterized in that: include: Acquisition module: obtains basic information data and historical bidding data of bidding projects; Generate a project feature map of the bidding project based on basic information data, and obtain a project risk feature library based on the project feature map; Acquisition and matching module: collects the risk identification scope of the early warning model, performs feature matching based on the risk identification scope and the project risk feature library to obtain the corresponding risk points to be warned for the bidding project; The first processing module: obtains the risk correlation trajectory between the risk point to be warned and the bidding project based on the historical bidding data of the bidding project, and obtains the risk correlation coefficient corresponding to the risk point to be warned based on the risk correlation trajectory; Analysis module: Set risk warning parameters for bidding projects based on basic information data; Obtain historical risk parameters of the risk point to be warned, and obtain the risk warning coefficient corresponding to the risk point to be warned based on the risk warning parameters and the historical risk parameters of the risk point to be warned; The second processing module: obtains the comprehensive risk coefficient corresponding to the risk point to be warned based on the risk correlation coefficient and the risk impact coefficient; Risk warning module: obtains the target warning risk points corresponding to the bidding project based on the comprehensive risk coefficient, and generates risk warning information for the bidding project based on the target warning risk points.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, a bidding risk warning method based on a large model as claimed in any one of claims 1 to 8 is implemented.

Images