Supplier risk automatic evaluation control method and system in bidding and tendering process
By conducting multi-dimensional analysis of bidder back-tuning information, historical behavior data and bid quotations of bidders, a supplier risk assessment report is generated, which solves the problem of inability to effectively identify potential risks in the existing technology, and improves the fairness of the bidding process and Transparency enables real-time risk monitoring and early warning.
Patent Information
- Application Number
- CN202411016254.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-27
- Publication Date
- 2024-11-15
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing bidding technology fails to make full use of big data and artificial intelligence, and cannot effectively identify and control the potential risks of suppliers, resulting in low impartiality, transparency and efficiency in the bidding process.
By conducting multi-dimensional analysis of bidder back-tuning information, historical behavior data and bid quotations, algorithm models are used to evaluate credit risks, associated risks, bid behavior risks and bid document risks, generate supplier risk assessment reports, and realize automated risk monitoring and assessment.
It realizes rapid processing and efficient use of data, improves the accuracy and efficiency of risk assessment, reduces interference from human factors, can monitor and warn of potential risks in real time, and improves the fairness and transparency of the bidding process.
Smart Images

Figure CN118966763A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of bidding, and in particular to a method and system for automatically evaluating and controlling supplier risks in a bidding process. Background Art
[0002] In recent years, with the rapid development of big data and artificial intelligence technology, new solutions have been provided for supplier risk assessment in the bidding process. Through big data analysis and mining technology and artificial intelligence algorithms, supplier risks can be more effectively identified, evaluated and controlled, and the fairness, transparency and efficiency of bidding can be improved.
[0003] In the current bidding process, traditional methods fail to fully utilize big data and advanced technologies for data analysis and mining, and thus fail to discover the deep-seated risks hidden behind the data. Summary of the invention
[0004] The present invention provides a method and system for automatically evaluating and controlling supplier risks during a bidding process, which can timely discover potential risks through continuous monitoring and evaluation of various risks of bidders.
[0005] In order to solve the above technical problems, the technical solutions provided by the present invention are as follows:
[0006] In a first aspect, a method for automatically evaluating and controlling supplier risks during a bidding process comprises the following steps:
[0007] Step 1: Based on the basic information of the bidder, obtain the credit risk and business operation risk assessment results of the bidder;
[0008] Step 2: Based on the background check information penetration analysis between multiple bidders, the association relationship between bidders is evaluated to obtain the bidder association risk assessment results;
[0009] Step 3: Based on the big data historical bidding data disclosure platform, obtain the historical behavior information of multiple bidders, use the algorithm model to evaluate the bidders' multi-faceted bidding behaviors, and obtain the bidding behavior correlation evaluation results;
[0010] Step 4: Testing the bidding documents of multiple bidders to obtain the risk assessment results associated with the bidding documents;
[0011] Step 5: Detect the bids of multiple bidders, evaluate the regularity of bids and the possibility of several bidders bidding at low prices; and construct model indicators through historical big data to obtain the bid curve of the subject matter, evaluate the abnormality of bids, and thus obtain the bid risk assessment results;
[0012] Step 6: Generate a supplier risk assessment report based on the bidder's credit risk and business operation risk assessment results, the bidder's associated risk assessment results, the bidding behavior associated risk assessment results, the bidding document associated risk assessment results and the bid quotation risk assessment results.
[0013] Furthermore, based on the basic information of the bidder, the credit risk and business operation risk assessment results of the bidder are obtained, including:
[0014] Obtain the bidder’s credit information from the website;
[0015] Based on the credit information of the bidder, determine whether the bidder has a serious record of dishonesty. If so, it is considered a negative item; if not, obtain the basic background check information of the bidder from the website; the basic background check information includes the legal representative, senior management, shareholder structure, industrial and commercial registration information, historical litigation status and public opinion information;
[0016] Based on the background check information of the bidder, it is determined whether the bidder has legal litigation risks and public opinion risks. If there is asset mortgage or pledge information, as well as public opinion risks, the corresponding bidder will be listed as medium risk.
[0017] Furthermore, based on the penetration analysis of the background check information between multiple bidders, the association relationships of the bidders are evaluated to obtain the bidder association risk assessment results, including:
[0018] Obtain basic background information of bidders through the website;
[0019] Based on the basic background information of the bidders, conduct risk assessment on the relationship between senior management, actual holding relationship, and contact information relationship between the bidders;
[0020] Whether the legal representative of each bidder is the same person, or whether the management personnel of each bidder are the same person or they hold cross-positions. If so, the corresponding bidder will be listed as high risk;
[0021] Based on the basic information of each bidder, deep legal person relationships, historical equity changes and historical litigation situations, determine whether there is an actual controlling relationship between the bidders. If so, the corresponding bidder will be listed as high risk;
[0022] Determine whether the business registration telephone number, business registration email address, and registered residence of each bidder are the same; if so, the corresponding bidder will be listed as medium risk.
[0023] Furthermore, based on the penetration analysis of the background check information between multiple bidders, the association relationships of the bidders are evaluated to obtain the bidder association risk assessment results, including:
[0024] Through the big data historical bidding data disclosure platform, the historical behavior information of bidders is obtained, and the historical bidding behavior of bidders is evaluated to obtain the historical bidding behavior evaluation results;
[0025] Obtain data on the bidder's bid times, bid items, and winning bid times and winning items in the past three years;
[0026] The bidder's winning rate is calculated based on the bid number, bid items, winning number and winning items of the bidder in the past three years. If the winning rate is less than the threshold, the corresponding bidder will be marked as high risk;
[0027] Calculate the industry distribution of the projects that the bidder participates in. If the distribution rate is less than the threshold, the corresponding bidder will be marked as high risk;
[0028] Intelligently identify the units that have jointly bid with a specific bidder in the past three years, and analyze the winning patterns of the jointly bidders. Based on the winning patterns, determine whether there are abnormal encounters. If there are abnormal encounters > the threshold, the corresponding bidder will be marked as high risk.
[0029] Furthermore, through the big data historical bidding data disclosure platform, the historical behavior information of bidders is obtained, and the historical bidding behavior of bidders is evaluated to obtain the historical bidding behavior evaluation results, including
[0030] Obtain the remittance account information of the bid bond from the bidding transaction platform interface, and check whether the bid bond is transferred or returned to multiple suppliers by the same unit or individual account to determine whether there is improper bidding collaboration; if so, treat it as a negative item;
[0031] Obtain the software and hardware attribute information of suppliers through the bidding transaction platform interface, compare the software and hardware attributes of different suppliers, and determine whether there are identical or similar situations, so as to determine whether there is collusion or sharing of bidding documents; if so, mark the corresponding bidder as high risk;
[0032] The physical characteristics of the bidding documents are judged. If it is found that multiple bidding documents are printed with the same ink or the USB flash drives are from the same store, it is determined whether there is collusion or sharing of the bidding documents; if so, the corresponding bidder will be marked as high risk.
[0033] Furthermore, the bidding documents of multiple bidders are tested to obtain the associated risk assessment results of the bidding documents, including:
[0034] Verify whether the supplier’s authorized representative is duplicated in the document; if so, treat it as a negative item;
[0035] Verify whether the bidding documents of different bidders are mixed; if so, treat it as a negative item;
[0036] Analyze the packaging material, size, seal color and style of paper tender documents, and check whether the binding form, paper condition, directory layout and text style of the tender documents are similar or consistent; if so, mark the corresponding bidder as high risk;
[0037] Verify the name of the author of the electronic bidding document. If it is the same person, mark the corresponding bidder as high risk;
[0038] Detect the repetition rate of the text content of technical documents in the electronic bidding documents of multiple bidders. If the repetition rate is greater than the threshold, the corresponding bidder will be marked as high risk; detect and analyze the repeated insertion of special pictures, the position of stamps in the text, and the abnormal consistency of simple writing errors; if so, the corresponding bidder will be marked as high risk, and compare the similarity of the technical document formats in the electronic bidding documents. If the similarity is greater than the threshold, the corresponding bidder will be marked as medium risk; compare the last save time of the electronic bidding documents. If the save time interval is less than the threshold, the corresponding bidder will be marked as medium risk.
[0039] Furthermore, the bids of multiple bidders are tested to assess the regularity of the bids and the possibility of several bidders bidding at low prices, including:
[0040] Determine whether the bids of multiple bidders show regular differences; if so, mark the corresponding bidders as high risk;
[0041] Determine whether the bidder's bid price shows the situation of narrowing the price gap between bidders by means of ultra-low bids and ultra-low scores. If so, the corresponding bidder will be marked as high risk.
[0042] Furthermore, the bidding price curve of the subject matter is obtained by constructing model indicators through historical big data, and the abnormality of the bidding price is evaluated, so as to obtain the bidding price risk assessment results, including:
[0043] Obtain historical big data and construct a price curve chart for similar product services or similar equipment services in the past three years; compare the bid price with the price curve, evaluate the abnormality of the bid price, and obtain the bid price risk assessment result.
[0044] In the second aspect, a supplier risk automatic evaluation control system in the bidding process includes:
[0045] The supplier credit analysis module is used to retrieve the supplier's credit information, legal proceedings and public opinion risk information from the preset website, and analyze the supplier's credit risk based on the search results;
[0046] The supplier association analysis module is used to obtain basic information of suppliers during background investigation and evaluate the association risks between suppliers based on the basic information;
[0047] The bidding behavior correlation analysis module is used to obtain the supplier's behavior data through historical bidding data, and evaluate the correlation risk of the supplier's bidding behavior based on the behavior data;
[0048] The bidding document analysis module is used to conduct risk assessment on suppliers’ abnormal bidding documents by combining paper bidding documents and electronic bidding documents;
[0049] The bidding and quotation analysis module is used to conduct risk assessment on abnormal bidding and quotation of suppliers based on the bidding and quotation situations of different suppliers;
[0050] The risk report generation module is used to generate a bidding risk report by combining the supplier's credit risk assessment results, associated risk assessment results, bidding behavior risk assessment results, bidding document risk assessment results and bidding quotation risk assessment results.
[0051] Compared with the prior art, the invention has the beneficial effects.
[0052] The above-mentioned scheme of the present invention can automatically collect and integrate data from different channels, such as bidder background information, historical bidding data, etc., to achieve rapid processing and efficient use of data, and greatly improve the efficiency and accuracy of data processing. Through the penetrating analysis of bidder background information, algorithm model evaluation of historical behavior information, and regularity and abnormality detection of bid quotations, the present invention can deeply explore potential risk points in the data. The present invention is data-driven, and through multi-dimensional and all-round data analysis, it objectively and comprehensively evaluates supplier risks, reducing the interference of human factors in data interpretation and risk assessment. The present invention can obtain and process the latest data in real time, promptly reflect the risk changes of bidders, and realize dynamic monitoring and early warning of supplier risks. BRIEF DESCRIPTION OF THE DRAWINGS
[0053] The accompanying drawings are used for further understanding of the present invention and are used to explain the present invention together with the embodiments of the present invention, but do not constitute a limitation of the present invention.
[0054] Figure 1 It is a flow chart of a method for automatically evaluating and controlling supplier risks in a bidding process provided by an embodiment of the present invention.
[0055] Figure 2 It is a schematic diagram of an automatic supplier risk assessment control system in a bidding process provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0056] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.
[0057] In order to better understand the above technical solution, the technical solution of the present invention will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.
[0058] like Figure 1 As shown, the present invention provides a method for automatically evaluating and controlling supplier risks in a bidding process, the method comprising the following steps:
[0059] Step 1: Based on the basic information of the bidder, obtain the credit risk and business operation risk assessment results of the bidder;
[0060] Step 2: Based on the background check information penetration analysis between multiple bidders, the association relationship between bidders is evaluated to obtain the bidder association risk assessment results;
[0061] Step 3: Based on the big data historical bidding data disclosure platform, obtain the historical behavior information of multiple bidders, use the algorithm model to evaluate the bidders' multi-faceted bidding behaviors, and obtain the bidding behavior correlation evaluation results;
[0062] Step 4: Testing the bidding documents of multiple bidders to obtain the risk assessment results associated with the bidding documents;
[0063] Step 5: Detect the bids of multiple bidders, evaluate the regularity of bids and the possibility of several bidders bidding at low prices; and construct model indicators through historical big data to obtain the bid curve of the subject matter, evaluate the abnormality of bids, and thus obtain the bid risk assessment results;
[0064] Step 6: Generate a supplier risk assessment report based on the bidder's credit risk and business operation risk assessment results, the bidder's associated risk assessment results, the bidding behavior associated risk assessment results, the bidding document associated risk assessment results and the bid quotation risk assessment results.
[0065] In an embodiment of the present invention, the present invention can automatically collect and integrate data from different channels, such as bidder background information, historical bidding data, etc., to achieve rapid processing and efficient use of data, and greatly improve the efficiency and accuracy of data processing. Through the penetrating analysis of bidder background information, algorithm model evaluation of historical behavior information, and regularity and abnormality detection of bid quotations, the present invention can deeply explore potential risk points in the data. The present invention is data-driven, and through multi-dimensional and comprehensive data analysis, it objectively and comprehensively evaluates supplier risks, reducing the interference of human factors in data interpretation and risk assessment. The present invention can obtain and process the latest data in real time, promptly reflect the risk changes of bidders, and realize dynamic monitoring and early warning of supplier risks.
[0066] In a preferred embodiment of the present invention, based on the basic information of the bidder, the credit risk and business operation risk assessment results of the bidder are obtained, including:
[0067] Obtain the bidder’s credit information from the website;
[0068] Based on the credit information of the bidder, determine whether the bidder has a serious record of dishonesty. If so, it is considered a negative item; if not, obtain the basic background check information of the bidder from the website; the basic background check information includes the legal representative, senior management, shareholder structure, industrial and commercial registration information, historical litigation status and public opinion information;
[0069] Based on the background check information of the bidder, it is determined whether the bidder has legal litigation risks and public opinion risks. If there is asset mortgage or pledge information, as well as public opinion risks, the corresponding bidder will be listed as medium risk.
[0070] In an embodiment of the present invention, a relevant credit information query website or government public service platform is accessed, and relevant information of the bidder, such as the company name, unified social credit code, etc., is input to retrieve its credit record, obtain and download the bidder's credit report or relevant credit information, and judge whether the bidder has a serious record of dishonesty based on the bidder's credit information. The bidder's credit report is carefully reviewed to check whether there is a record of serious dishonesty, such as a record of illegal and irregular behavior, an unfulfilled court judgment, administrative penalties, etc. If it is found that the bidder has a serious record of dishonesty, such as being included in the list of dishonest persons subject to execution or having multiple illegal records, this is regarded as a negative item, that is, the bidder does not meet the bidding qualifications. If the bidder does not have a serious record of dishonesty, the bidder's basic background information is obtained from the website: the relevant business information query website or government service platform is accessed, and relevant information of the bidder is input to retrieve and obtain its basic background information, which includes legal representative, senior management information, shareholder structure, industrial and commercial registration information, historical litigation status, and public opinion information.
[0071] Determine whether the bidder has legal litigation risk and public opinion risk based on the bidder's background information: Analyze the bidder's historical litigation situation, check whether there are major legal disputes or pending lawsuits, which may bring potential risks and costs, evaluate the bidder's public opinion information, check whether there are negative news reports, consumer complaints or social opinions that are unfavorable to it, if the bidder has asset mortgage or pledge information, further evaluate its asset status and operating risks, if the above situation exists, the corresponding bidder will be classified as medium risk:
[0072] Based on the above analysis, if the bidder faces legal litigation risks, public opinion risks, or operational risks brought about by asset mortgage or pledge, the bidder will be classified as a medium-risk level and corresponding risk management measures will be formulated, such as requiring the bidder to provide more explanations or supporting documents, or setting corresponding risk prevention clauses in the contract.
[0073] Through detailed background check information retrieval and analysis, the credit risk and business operation risk of bidders can be more effectively identified, and bidders with serious dishonesty records or high risks can be screened out at the beginning of the bidding process, which will help reduce the risk of default, legal risk and public opinion risk that may arise in subsequent cooperation. Through preliminary risk assessment of bidders, unnecessary waste of resources can be reduced. Assessments based on publicly available credit information and background check data make the bidding process more fair and transparent, reducing the space for human manipulation.
[0074] In a preferred embodiment of the present invention, based on the background information penetration analysis between multiple bidders, the association relationship of the bidders is evaluated to obtain the bidder association risk assessment results, including:
[0075] Obtain basic background information of bidders through the website;
[0076] Based on the basic background information of the bidders, conduct risk assessment on the relationship between senior management, actual holding relationship, and contact information relationship between the bidders;
[0077] Whether the legal representative of each bidder is the same person, or whether the management personnel of each bidder are the same person or they hold cross-positions. If so, the corresponding bidder will be listed as high risk;
[0078] Based on the basic information of each bidder, deep legal person relationships, historical equity changes and historical litigation situations, determine whether there is an actual controlling relationship between the bidders. If so, the corresponding bidder will be listed as high risk;
[0079] Determine whether the business registration telephone number, business registration email address, and registered residence of each bidder are the same; if so, the corresponding bidder will be listed as medium risk.
[0080] In an embodiment of the present invention, a relevant business information query website or government service platform is accessed to retrieve and download detailed background information of each bidder, including key data such as the list of senior management personnel, shareholder structure, actual controller, contact information, and registered address. Compare whether there are managers with the same name appearing in the lists of different bidders, analyze the management personnel's employment history, and check whether there are cross-appointments between different bidders; deeply analyze the shareholder structure of each bidder, track down the ultimate actual controller, and compare the actual controllers or major shareholders of each bidder to determine whether there are the same actual controllers or shareholder groups. Check the legal representative information provided by each bidder to confirm whether the same person serves as the legal representative of multiple bidders, collect and analyze the historical equity change records of each bidder, observe whether there are frequent equity transfers or common equity transaction counterparties, review historical litigation situations, and pay special attention to cases involving multiple bidders, which may reveal potential associations. Check the industrial and commercial registration telephone number, email address, and registered residence of each bidder to check whether there are duplications. By integrating background information from different bidders through automated tools and performing efficient data comparison, potential relationships between bidders can be quickly identified. By deeply mining historical data, such as equity changes and litigation records, potential correlation risks can be discovered and warned of earlier.
[0081] In a preferred embodiment of the present invention, based on the background information penetration analysis between multiple bidders, the association relationship of the bidders is evaluated to obtain the bidder association risk assessment results, including:
[0082] Through the big data historical bidding data disclosure platform, the historical behavior information of bidders is obtained, and the historical bidding behavior of bidders is evaluated to obtain the historical bidding behavior evaluation results;
[0083] Obtain data on the bidder's bid times, bid items, and winning bid times and winning items in the past three years;
[0084] The bidder's winning rate is calculated based on the bid number, bid items, winning number and winning items of the bidder in the past three years. If the winning rate is less than the threshold, the corresponding bidder will be marked as high risk;
[0085] Calculate the industry distribution of the projects that the bidder participates in. If the distribution rate is less than the threshold, the corresponding bidder will be marked as high risk;
[0086] Intelligently identify the units that have jointly bid with a specific bidder in the past three years, and analyze the winning patterns of the jointly bidders. Based on the winning patterns, determine whether there are abnormal encounters. If there are abnormal encounters > the threshold, the corresponding bidder will be marked as high risk.
[0087] In an embodiment of the present invention, a big data historical bidding data disclosure platform, a bidding agency or an electronic bidding data platform of a tenderer is accessed. Such platforms have accumulated a large amount of historical bidding data, and the bidder's historical bidding record data is obtained, including bidding time, bidding project, bidding amount, winning bid situation and other key information. The bid records of the bidder in the past three years are screened out from the historical bidding data, and the number of bids, bidding projects participated in, and the number of winning bids and winning projects of the bidder in these three years are counted. The winning bid rate calculation formula is: number of winning bids ÷ number of bids × 100%. Set a threshold for the winning bid rate (such as the industry average level or the minimum acceptance standard set by the tenderer), compare the calculated winning bid rate with the threshold, and if the winning bid rate is lower than the threshold, the bidder is marked as high risk.
[0088] The projects in which the bidders participate are classified by industry, and the bidders' project participation or proportion in each industry is calculated. A threshold for the industry distribution rate is set (which can be based on industry characteristics or the needs of the tenderer). If the bidder's distribution rate in a certain industry is lower than the threshold, it means that the bidder may lack experience in the industry, so it will be marked as high risk.
[0089] Identify units that frequently bid together with a specific bidder from historical data, analyze the historical winning patterns of these co-bidding units, such as the winning frequency, the types of winning projects, etc., set a threshold for abnormal encounters to determine whether there is an improper cooperative relationship or collusion between the co-bidding units. If the number of abnormal encounters exceeds the threshold, the bidder will be marked as high risk.
[0090] Through the big data historical bidding data disclosure platform and related electronic bidding transaction data platform, the bidders' historical bidding data can be efficiently integrated and analyzed, so as to quickly obtain risk assessment results. Based on the analysis of a large amount of historical data, the bidders' abnormal behaviors and high-risk patterns can be more accurately identified. By setting thresholds and real-time monitoring data, potential risk points can be discovered in a timely manner, providing early warning information for the tenderer.
[0091] In a preferred embodiment of the present invention, the historical behavior information of the bidder is obtained through the relevant electronic bidding transaction data platform, and the historical bidding behavior of the bidder is evaluated to obtain the historical bidding behavior evaluation result, including
[0092] Obtain the remittance account information of the bid bond from the bidding transaction platform interface, and check whether the bid bond is transferred or returned to multiple suppliers by the same unit or individual account to determine whether there is improper bidding collaboration; if so, treat it as a negative item;
[0093] Obtain the software and hardware attribute information of suppliers through the bidding transaction platform interface, compare the software and hardware attributes of different suppliers, and determine whether there are identical or similar situations, so as to determine whether there is collusion or sharing of bidding documents; if so, mark the corresponding bidder as high risk;
[0094] The physical characteristics of the bidding documents are judged. If it is found that multiple bidding documents are printed with the same ink or the USB flash drives are from the same store, it is determined whether there is collusion or sharing of the bidding documents; if so, the corresponding bidder will be marked as high risk.
[0095] In an embodiment of the present invention, the bid bond remittance account information of all bidders is obtained from the bidding transaction platform through an API interface or a data exchange protocol. This information may include key data such as account name, account number, bank of account opening, remittance time, etc. The obtained account information is cleaned and sorted to ensure that the data format is unified and accurate. Data analysis tools or scripts are used to check whether the same account has transferred bid bonds for multiple different suppliers, or whether the bid bonds of multiple suppliers have been returned to the same account. If this is found, it may indicate the existence of improper bidding collaboration, such as bid rigging and bid collusion.
[0096] Through the API interface or data exchange protocol, the software and hardware attribute information of all suppliers is obtained from the bidding transaction platform. This information may include computer configuration, operating system version, office software version, printer model, etc. The obtained software and hardware attribute information is cleaned and standardized, and data analysis tools are used to compare whether the software and hardware attributes of different suppliers are abnormally similar or exactly the same. If the software and hardware attributes of multiple suppliers are highly consistent, it may mean that there is collusion or sharing of bidding documents among them.
[0097] Through the scanned version of the bidding documents, image processing technology is used to extract the physical features of the documents. These features may include the depth and distribution of the printed ink, the texture number of the paper, etc. By comparing the physical features of different bidding documents, it is determined whether there are abnormal similarities and the bidder is marked as high risk. The automatic acquisition of data through the API interface reduces the workload of manual data collection and collation, improves the efficiency and accuracy of data processing, and uses big data analysis and comparison technology to more accurately identify improper behavior and high-risk patterns in the bidding process. The analysis and processing results based on public data can provide a more fair and transparent bidding environment for all participants. Through real-time monitoring and data analysis, potential risk points can be discovered and prevented in a timely manner during the bidding process.
[0098] Among them, the physical features of the file can be extracted using image processing technology, which can be achieved through the following specific steps:
[0099] First, paper documents are converted into digital images using a scanner or high-resolution camera, and the acquired images are preprocessed, including denoising, binarization, contrast enhancement, etc., to highlight the physical features of the documents. The image is processed to obtain the pixel value of the filtered image at the coordinate (x, y). I′(x, y) is the pixel value of the filtered image at the coordinate (x, y); A(i, j) is an attenuation function used to reduce the influence of pixels (x+i, y+j) far from the center pixel (x, y). When (i, j) is far from (0, 0), the value of A(i, j) will gradually decrease to ensure that the distant pixels contribute less to the center pixel; W(x+i, y+j) is a weight function that weights different pixel positions based on the local features of the image (such as edge strength, texture, etc.). For example, pixels near the edge may be given higher weights to retain edge information; K(i, j) is a filter kernel. kernel), is a two-dimensional function used to determine the contribution of surrounding pixels to the central pixel (x, y) during the filtering process; I(x+i, y+j) is the pixel value of the original image at the coordinate (x+i, y+j); di and dj are integral variables, representing infinitesimal increments to i and j during the integration process; the integral range It means that the integration of i and j is performed on the entire image plane.
[0100] Convert the file image into a grayscale image, calculate the gradient magnitude and direction of each pixel in the image, apply non-maximum suppression, retain only the points with the local maximum gradient magnitude, and use a double threshold algorithm to detect and connect edges; determine a pixel pair direction (such as 0 degrees, 45 degrees, 90 degrees, 135 degrees) and distance d, traverse each pixel in the image, record the grayscale pair of the pixel and the target pixel (the pixel at a distance d along the selected direction), and construct a grayscale co-occurrence matrix, in which each element (i, j) represents the number of transitions from grayscale level i to j, and calculate texture features such as contrast, homogeneity, energy, and entropy from the GLCM.
[0101] Through Hough transform, the shape in the image space can be converted into a peak in the parameter space, so as to identify a straight line or other regular shape. For line detection, Hough transform usually uses polar coordinates to represent the line, and uses an accumulator to count the number of times each point in the parameter space is passed, so as to identify the line. The contour is tracked and extracted by analyzing the connectivity and brightness changes between pixels; the arrangement of the text, such as horizontal or vertical, can be determined by analyzing the arrangement and spacing of the contour; seals usually have specific shapes and patterns, and the shape and type of the seal can be identified by analyzing the shape and feature points of the contour.
[0102] The extracted physical features are quantified and encoded and converted into a format that can be understood and compared by computers. For example, edge information can be converted into a series of coordinate points, and texture features can be converted into numerical vectors. The extracted physical features of the file are compared with the features of known samples or databases to identify the authenticity, source or version of the file. The Euclidean distance is used to evaluate the similarity between the two features. Based on the comparison results, the identification conclusion of the file is output, such as authenticity, forgery, tampering, etc. If necessary, the extracted features and comparison results can be displayed in a visual form to facilitate user understanding and analysis.
[0103] In a preferred embodiment of the present invention, intelligent comparison is performed on the bidding documents of multiple bidders to obtain the associated risk assessment results of the bidding documents, including:
[0104] Verify whether the authorized representatives of different bidders in the document are duplicated; if so, treat it as a negative item;
[0105] Verify whether the bidding documents of different bidders are mixed; if so, treat it as a negative item;
[0106] Analyze the packaging material, size, seal color and style of paper tender documents, and check whether the binding form, paper condition, directory layout and text style of the tender documents are similar or consistent; if so, mark the corresponding bidder as high risk;
[0107] Verify the name of the author of the electronic bidding document. If it is the same person, mark the corresponding bidder as high risk;
[0108] Detect the repetition rate of the text content of technical documents in the electronic bidding documents of multiple bidders. If the repetition rate is greater than the threshold, the corresponding bidder will be marked as high risk; detect and analyze the repeated insertion of special pictures, the position of stamps in the text, and the abnormal consistency of simple writing errors; if so, the corresponding bidder will be marked as high risk, and compare the similarity of the technical document formats in the electronic bidding documents. If the similarity is greater than the threshold, the corresponding bidder will be marked as medium risk; compare the last save time of the electronic bidding documents. If the save time interval is less than the threshold, the corresponding bidder will be marked as medium risk.
[0109] In an embodiment of the present invention, a database or list is created to record the authorized representative information of the supplier in each bidding document. For each new bidding document, check whether its authorized representative already exists in the database or list. If a duplicate is found, the bidding document is marked as a negative item. Perform a detailed unpacking inspection on each bidding document to ensure that the document only contains the information of the bidder. If the information of other bidders is found, the bidding document is marked as a negative item. The packaging material, size, seal color and style of each bidding document are recorded in detail, and the binding form, paper condition, directory layout and text style of the bidding document are checked. Image processing technology (such as image comparison after scanning) is used to assist in judging the similarity of these physical features. If highly similar or consistent features are found, the bidder is marked as high risk. Extract metadata from electronic bidding documents, including information such as author names, and compare whether the author names of different bidding documents are the same. If the same author is found, the bidder is marked as high risk. Using the text similarity algorithm, the repetition rate of the technical document content in the electronic bidding documents of multiple bidders is detected, and a reasonable repetition rate threshold is set (such as 20% or set according to industry standards). If the repetition rate exceeds the threshold, the bidder is marked as high risk.
[0110] Perform image recognition on special images in electronic bidding documents, check whether there are repeated images, analyze the seal position in the text, check whether there are abnormal seal patterns, and use natural language processing technology to detect whether simple writing errors are abnormally consistent. If any of the above abnormalities are found, the bidder will be marked as high risk. Use file format comparison tools or algorithms to check whether the technical file formats of different bidders are similar, set a similarity threshold to judge the consistency of the format, and if the similarity exceeds the threshold, the bidder will be marked as medium risk. Extract the metadata of electronic bidding documents, including the last save time, compare the save time of different bidding documents, calculate the time interval, and set a reasonable time interval threshold (such as a few minutes or seconds). If the save time interval is less than the threshold, the bidder will be marked as medium risk. By processing large amounts of data through algorithms and automated tools, it is possible to quickly and accurately identify the associated risks in bidding documents, reduce human errors and subjective judgments, and timely mark high-risk and medium-risk bidding documents, provide risk warnings for tenderers, and help prevent potential improper behaviors such as collusion and bidding.
[0111] In a preferred embodiment of the present invention, the bids of multiple bidders are tested, and the regularity of the bids and the possibility of several bidders bidding at low prices are evaluated, including:
[0112] Determine whether the bids of multiple bidders show regular differences; if so, mark the corresponding bidders as high risk;
[0113] Determine whether the bidder's bid price shows the situation of narrowing the price gap between bidders by means of ultra-low bids and ultra-low scores. If so, the corresponding bidder will be marked as high risk.
[0114] In an embodiment of the present invention, bid quotation data is obtained from all bidders to ensure the integrity and accuracy of the data, remove duplicate, erroneous or incomplete data entries, and ensure data quality. If the format or unit of the bid quotation data is not uniform, it needs to be converted to ensure data consistency. According to actual needs or through some statistical methods (such as the elbow rule), the optimal number of clusters K is determined, and K data points are randomly selected as the initial centroid; data points are assigned to the nearest centroid: the Euclidean distance between each data point and each centroid is calculated, and it is assigned to the nearest centroid to form K clusters. For each cluster, the average value of all data points is calculated, and the average value is set as the new centroid. Repeat the above two steps until the position of the centroid no longer changes significantly or reaches a preset number of iterations. If the data is two-dimensional or three-dimensional, the clustering results can be visualized by a scatter plot or a three-dimensional plot. For high-dimensional data, dimensionality reduction techniques (such as PCA) can be used and then visualized to observe the clustering results. If the quotations of some bidders are obviously clustered together to form clusters that are significantly different from other quotations, this may indicate that there is a regularity in the quotations of these bidders. This regularity may manifest itself in similar price levels, similar price fluctuation patterns, etc. If obvious regularity is detected, the corresponding bidder will be marked as high risk.
[0115] Based on historical data or industry standards, set a threshold for ultra-low bids, and filter out bids below the set threshold from the bid quotation data; based on historical data, market conditions or expert advice, set a threshold for ultra-low bids. This threshold should be a reasonable limit below the normal market price or historical average bid; filter out ultra-low bid data below the set threshold from all bid quotation data to form a new data set. For the filtered ultra-low bid data, further check the integrity and accuracy of the data, remove any abnormal or erroneous data points, and standardize the data to eliminate the influence of different quotation units or orders of magnitude so that all data are compared on the same scale. Calculate the distance between each pair of data points in the ultra-low bid data set and construct a distance matrix. This distance can be Euclidean distance, etc. Starting with each data point as a cluster, gradually merge the nearest clusters until the preset number of clusters is reached or a certain stop condition is met. In the merging process, different link criteria can be used, such as single link, full link or group average link. The result of hierarchical clustering is represented by a dendrogram. By observing the dendrogram, it can be identified which ultra-low bid data points are first merged together to form stable clusters. If the ultra-low bids of some bidders are merged together early in the dendrogram to form an obvious cluster, this may indicate that these bidders narrow the price gap with other bidders through ultra-low bids. This situation may involve unfair competition strategies or price manipulation. Based on the results of hierarchical clustering, the risk of bidders narrowing the price gap through ultra-low bids is assessed and marked as high risk.
[0116] Through data cleaning and conversion, the accuracy and consistency of bidding and quotation data are ensured. Cluster analysis can reveal the hidden rules that may exist in bidding and quotation. Through clustering results, the bidders are risk assessed and marked, so that the tenderer can more accurately identify high-risk bidders and take corresponding preventive measures. The clustering analysis results based on data provide strong data support for the tenderer's decision-making, improving the scientificity and accuracy of the decision-making. Through automated data processing and analysis processes, the supervision efficiency of the bidding process is improved, and the cost and time consumption of manual review are reduced.
[0117] In a preferred embodiment of the present invention, the bidding price curve of the subject matter is obtained by constructing a model indicator through historical big data, and the abnormality of the bidding price is evaluated, thereby obtaining the bidding price risk assessment result, including:
[0118] Obtain historical big data and construct a price curve chart for similar product services or similar equipment services in the past three years; compare the bid price with the price curve, evaluate the abnormality of the bid price, and obtain the bid price risk assessment result.
[0119] In an embodiment of the present invention, all transaction data of similar products or similar equipment services in the past three years are systematically collected from the enterprise's database or external data source, including but not limited to key information such as bid price, transaction price, transaction time, etc., and duplicate, erroneous or invalid data records are cleared to ensure the accuracy and reliability of the data. The data from different sources are unified and standardized in format to facilitate subsequent data analysis and modeling.
[0120] Sort the historical data in chronological order, and group them by month, quarter or year as needed. Calculate the average bid price or transaction price for the data in each time period, and use statistical software or data visualization tools to draw a price curve chart with time as the horizontal axis and average price as the vertical axis. Smoothing can be used to better display the price trend; input the bid price data of the subject matter into the analysis system; use the algorithm Z-score to identify bid prices that deviate significantly from the price curve, which may be considered abnormal, and mark the position of the bid price on the price curve to intuitively show whether it is within the normal price fluctuation range.
[0121] Based on historical data, a risk assessment model is constructed, which can quantify the degree of abnormality of bid quotations. According to the degree to which the bid quotations deviate from the price curve, they are divided into different risk levels, such as low risk, medium risk and high risk. The analysis results are collated and a bid quotation risk assessment report is generated for reference by decision makers.
[0122] By systematically collecting, cleaning and integrating historical big data, enterprises can use these data resources more effectively, dig out valuable information, and construct price curve charts to help enterprises more intuitively understand the price trends of products or services, and provide data support for the rationality of bid quotations. By comparing with price curves, enterprises can accurately identify abnormal bid quotations and avoid potential cost risks.
[0123] An automatic supplier risk assessment control system in the bidding process, comprising:
[0124] Supplier credit analysis module 1, used to retrieve the supplier's credit information, legal proceedings and public opinion risk information according to the preset website, and analyze the supplier's credit risk based on the search results;
[0125] The supplier association analysis module 2 is used to obtain basic information of suppliers during background investigation and evaluate the association risks between suppliers based on the basic information;
[0126] The bidding behavior correlation analysis module 3 is used to obtain the supplier's behavior data through historical bidding data, and evaluate the correlation risk of the supplier's bidding behavior based on the behavior data;
[0127] The bidding document analysis module 4 is used to conduct risk assessment on abnormal bidding documents of suppliers by combining paper bidding documents and electronic bidding documents;
[0128] The bidding and quotation analysis module 5 is used to conduct risk assessment on abnormal bidding and quotation of suppliers according to the bidding and quotation situations of different suppliers;
[0129] The risk report generation module 6 is used to generate a bidding risk report by combining the supplier's credit risk assessment results, associated risk assessment results, bidding behavior risk assessment results, bidding document risk assessment results and bidding quotation risk assessment results.
[0130] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A method for automatically evaluating and controlling supplier risks during a bidding process, characterized in that: The following steps are involved: Step 1: Based on the basic information of the bidder, obtain the credit risk and business operation risk assessment results of the bidder; Step 2: Based on the background check information penetration analysis between multiple bidders, the association relationship between bidders is evaluated to obtain the bidder association risk assessment results; Step 3: Based on the big data historical bidding data disclosure platform, obtain the historical behavior information of multiple bidders, use the algorithm model to evaluate the bidders' multi-faceted bidding behaviors, and obtain the bidding behavior correlation evaluation results; Step 4: Testing the bidding documents of multiple bidders to obtain the risk assessment results associated with the bidding documents; Step 5: Check the bids of multiple bidders and evaluate the regularity of the bids and the possibility of several bidders bidding at low prices; The model indicators are constructed through historical big data to obtain the bid price curve of the subject matter, and the abnormality of the bid price is evaluated to obtain the bid price risk assessment result; Step 6: Generate a supplier risk assessment report based on the bidder's credit risk and business operation risk assessment results, the bidder's associated risk assessment results, the bidding behavior associated risk assessment results, the bidding document associated risk assessment results and the bid quotation risk assessment results.
2. The method for automatically evaluating supplier risks in the bidding process according to claim 1 is characterized by: Based on the basic information of the bidder, the credit risk and business risk assessment results of the bidder are obtained, including: Obtain the bidder’s credit information from the website; Based on the credit information of the bidder, determine whether the bidder has a serious record of dishonesty. If so, it is considered a negative item; if not, obtain the basic background check information of the bidder from the website; the basic background check information includes the legal representative, senior management, shareholder structure, industrial and commercial registration information, historical litigation status and public opinion information; Based on the background check information of the bidder, it is determined whether the bidder has legal litigation risks and public opinion risks. If there is asset mortgage or pledge information, as well as public opinion risks, the corresponding bidder will be listed as medium risk.
3. The method for automatically evaluating supplier risks in the bidding process according to claim 2 is characterized by: Based on the penetration analysis of background information between multiple bidders, the association relationships of bidders are evaluated to obtain the bidder association risk assessment results, including: Obtain basic background information of bidders through the website; Based on the basic background information of the bidders, conduct risk assessment on the relationship between senior management, actual holding relationship, and contact information relationship between the bidders; Whether the legal representative of each bidder is the same person, or whether the management personnel of each bidder are the same person or they hold cross-positions. If so, the corresponding bidder will be listed as high risk; Based on the basic information of each bidder, deep legal person relationships, historical equity changes and historical litigation situations, determine whether there is an actual controlling relationship between the bidders. If so, the corresponding bidder will be listed as high risk; Determine whether the business registration telephone number, business registration email address, and registered residence of each bidder are the same; if so, the corresponding bidder will be listed as medium risk.
4. The method for automatically evaluating supplier risks in the bidding process according to claim 3 is characterized by: Based on the penetration analysis of background information between multiple bidders, the association relationships of bidders are evaluated to obtain the bidder association risk assessment results, including: Through the big data historical bidding data disclosure platform, the historical behavior information of bidders is obtained, and the historical bidding behavior of bidders is evaluated to obtain the historical bidding behavior evaluation results; Obtain data on the bidder's bid times, bid items, and winning bid times and winning items in the past three years; The bidder's winning rate is calculated based on the bid number, bid items, winning number and winning items of the bidder in the past three years. If the winning rate is less than the threshold, the corresponding bidder will be marked as high risk; Calculate the industry distribution of the projects that the bidder participates in. If the distribution rate is less than the threshold, the corresponding bidder will be marked as high risk; Intelligently identify the units that have jointly bid with a specific bidder in the past three years, and analyze the winning patterns of the jointly bidding units. Based on the winning patterns, determine whether there are abnormal encounters; if there are abnormal encounters > the threshold, the corresponding bidder will be marked as high risk.
5. The method for automatically evaluating and controlling supplier risks during bidding according to claim 4 is characterized in that: Through the big data historical bidding data disclosure platform, the historical behavior information of bidders is obtained, and the historical bidding behavior of bidders is evaluated to obtain the historical bidding behavior evaluation results, including: Obtain the remittance account information of the bid bond from the bidding transaction platform interface, and check whether the bid bond is transferred or returned to multiple suppliers by the same unit or individual account to determine whether there is improper bidding collaboration; if so, treat it as a negative item; Obtain the software and hardware attribute information of suppliers through the bidding transaction platform interface, compare the software and hardware attributes of different suppliers, and determine whether there are identical or similar situations, so as to determine whether there is collusion or sharing of bidding documents; if so, mark the corresponding bidder as high risk; The physical characteristics of the bidding documents are judged. If it is found that multiple bidding documents are printed with the same ink or the USB flash drives are from the same store, it is determined whether there is collusion or sharing of the bidding documents; if so, the corresponding bidder will be marked as high risk.
6. The method for automatically evaluating and controlling supplier risks during bidding according to claim 5 is characterized in that: Test the bidding documents of multiple bidders and obtain the associated risk assessment results of the bidding documents, including: Verify whether the supplier’s authorized representative is duplicated in the document; if so, treat it as a negative item; Verify whether the bidding documents of different bidders are mixed; if so, treat it as a negative item; Analyze the packaging material, size, seal color and style of paper tender documents, and check whether the binding form, paper condition, directory layout and text style of the tender documents are similar or consistent; if so, mark the corresponding bidder as high risk; Verify the name of the author of the electronic bidding document. If it is the same person, mark the corresponding bidder as high risk; Detect the repetition rate of the text content of technical documents in the electronic bidding documents of multiple bidders. If the repetition rate is greater than the threshold, the corresponding bidder will be marked as high risk; detect and analyze the repeated insertion of special pictures, the position of stamps in the text, and the abnormal consistency of simple writing errors; if so, the corresponding bidder will be marked as high risk, and compare the similarity of the technical document formats in the electronic bidding documents. If the similarity is greater than the threshold, the corresponding bidder will be marked as medium risk; compare the last save time of the electronic bidding documents. If the save time interval is less than the threshold, the corresponding bidder will be marked as medium risk.
7. The method for automatically evaluating and controlling supplier risks during bidding according to claim 6 is characterized in that: Test the bids of multiple bidders and evaluate the regularity of the bids and the possibility of several bidders bidding at low prices, including: Determine whether the bids of multiple bidders show regular differences; if so, mark the corresponding bidders as high risk; Determine whether the bidder's bid price shows the situation of narrowing the price gap between bidders by means of ultra-low bids and ultra-low scores. If so, the corresponding bidder will be marked as high risk.
8. The method for automatically evaluating and controlling supplier risks during bidding according to claim 7 is characterized in that: The model indicators are constructed through historical big data to obtain the bid price curve of the subject matter, and the abnormality of the bid price is evaluated to obtain the bid price risk assessment results, including: Obtain historical big data and build price curve charts for similar product services or similar equipment services in the past three years; The bid price is compared with the price curve, and the abnormality of the bid price is evaluated to obtain the bid price risk assessment result.
9. An automatic risk assessment control system for suppliers during bidding, characterized in that: include: The supplier credit analysis module (1) is used to retrieve the supplier's credit information, legal proceedings and public opinion risk information from a preset website, and analyze the supplier's credit risk based on the search results; The supplier association analysis module (2) is used to obtain basic information of suppliers during background investigation and evaluate the association risks between suppliers based on the basic information; A bidding behavior correlation analysis module (3), which is used to obtain supplier behavior data through historical bidding data, and to evaluate the correlation risk of supplier bidding behavior based on the behavior data; A bidding document analysis module (4), used to conduct risk assessment on abnormal bidding documents of suppliers by combining paper bidding documents and electronic bidding documents; The bidding and quotation analysis module (5) is used to conduct risk assessment on abnormal bidding and quotation of suppliers based on the bidding and quotation situations of different suppliers; The risk report generation module (6) is used to generate a bidding risk report by combining the supplier's credit risk assessment results, associated risk assessment results, bidding behavior risk assessment results, bidding document risk assessment results and bidding quotation risk assessment results.
Citation Information
Patent Citations
Bidding risk early warning system and method
CN111091271A
Intelligent identification system for agreed bidding and bidding collusion
CN112561670A
Bidding behavior detection method and device based on electronic bidding file comparison
CN114492323A
Enterprise bidding behavior monitoring method and system based on bid invitation and bidding field, and medium
CN115018598A
Autonomous bidder solicitation and selection system
US20210304297A1
Cited By
Intelligent enterprise risk management and control system and method based on data analysis technology
CN119130163A
An enterprise risk intelligent control system and method based on data analysis technology
CN119130163B
Risk discrimination system and method, computer storage medium and electronic equipment
CN119850209A
Supplier fund bidding risk assessment method and electronic equipment
CN119918950A
Bid invitation information feedback comparison system based on big data
CN119991032A
