Briquette behavior recognition method, device and equipment and storage medium

Abstract

The invention discloses a bidding behavior recognition method and device, equipment and a storage medium, and relates to the technical field of bidding analysis, and the method comprises the steps: judging whether the final retainers of all bidding documents are the same or not based on the attribute information of all bidding documents, and determining whether the bidding documents have a bidding behavior or not according to a judgment result; determining the similarity between the text and the picture content in each bidding document, determining the comprehensive similarity based on the corresponding text similarity and the picture content similarity, and determining whether the bidding document has a surrounding bidding behavior or not according to the size relationship between the comprehensive similarity and a preset threshold value; analyzing semantic errors in the bidding documents, and determining whether bidding documents with the same semantic errors exist or not to judge whether the bidding documents have a surrounding bidding behavior or not; extracting quotation information in each bidding document, and determining whether the bidding document has the surrounding bidding behavior or not according to the quotation information and a predetermined project budget; and determining whether to send alarm information to a preset purchaser terminal or not according to the judgment result of each bidding behavior. And the bidding behavior is efficiently identified.

Description

Technical Field

[0001] The present invention relates to the technical field of bidding analysis, and in particular to a method, device, equipment and storage medium for identifying bid rigging and collusion. Background Art

[0002] With the rapid development of information technology and the widespread application of emerging technologies such as big data, cloud computing, and artificial intelligence, informatization has become an important trend in social development. In this context, discipline inspection and supervision work is also facing an urgent need to transform from traditional models to digital and intelligent ones. Traditional discipline inspection and supervision work is limited by the heavy task of manually processing bids, data, and information, and often faces challenges such as heavy workload, low processing efficiency, and inevitable human errors. Summary of the invention

[0003] In view of this, the purpose of the present invention is to provide a method, device, equipment and storage medium for identifying bid-rigging behavior, which can be based on information extraction, image recognition, similarity comparison and other means, and is no longer limited to the manual processing of bids by staff under normal circumstances, effectively alleviating the heavy workload of staff, significantly improving processing efficiency, and greatly reducing errors caused by human factors, achieving accurate and efficient bid review, and quickly capturing potential problems in the bidding process. The specific scheme is as follows:

[0004] In a first aspect, the present application discloses a method for identifying bid-rigging behavior, comprising:

[0005] Extracting attribute information of each bid document, determining whether the final savers of each bid document are the same based on the attribute information, and determining whether there is bid rigging and collusion in the bid document according to the corresponding determination result, so as to obtain a first bid rigging and collusion determination result;

[0006] Determine the similarity of the text and image content in each of the bid documents respectively, determine the comprehensive similarity based on the corresponding text similarity and image content similarity, and determine whether the bid document involves bid rigging according to the relationship between the comprehensive similarity and a preset threshold value, so as to obtain a second bid rigging judgment result;

[0007] Analyze the semantic errors in each of the bids, and determine whether the bids contain the bid-rigging behavior by determining whether there are bids with the same semantic errors, so as to obtain a third bid-rigging behavior determination result;

[0008] Extracting quotation information from each of the bids, and determining whether the bid involves bid rigging based on the quotation information and a predetermined project budget, so as to obtain a fourth bid rigging judgment result;

[0009] Determine whether to send corresponding alarm information to the preset buyer terminal based on the judgment results of each bid-rigging behavior.

[0010] Optionally, the attribute information includes the original author of the bid document, the final saver, the revision number and version number in the revision history, the company to which it belongs, and the saving time;

[0011] Accordingly, determining whether there is bid-rigging behavior in the bid document according to the corresponding judgment result includes:

[0012] If the final depositor of each of the bid documents is the same, it is determined that the bid documents involve bid rigging;

[0013] If the final depositors of the bid documents are different, determining whether the bid documents are written by a single author;

[0014] If the bid is written by a single author and there are no signs of multiple editing and revisions in the bid, it will be determined that the bid is suspected of bid rigging.

[0015] Optionally, the respectively determining the similarity of the text and the image content in each of the bid documents, and determining the comprehensive similarity based on the corresponding text similarity and image content similarity, includes:

[0016] Standardize the format of the tender documents and divide the contents of the bidding documents and the tender documents into several paragraphs;

[0017] Deleting the paragraphs in each of the said bids that are similar to the said bidding documents to obtain the bids after deletion;

[0018] Performing word segmentation processing on the text in each of the removed bids, and determining the word segmentation similarity between the word segmentation results corresponding to each of the removed bids based on the cosine similarity algorithm, so as to determine the text similarity between each of the bids based on the word segmentation similarity;

[0019] Extracting features from the image content of each of the bid documents through a computer vision model to obtain extracted features, and calculating the image content similarity between the bid documents through cosine similarity based on the extracted features;

[0020] The comprehensive similarity is determined according to the text similarity and the image content similarity.

[0021] Optionally, determining whether there is bid rigging in the bid document according to the magnitude relationship between the comprehensive similarity and a preset threshold value includes:

[0022] If the comprehensive similarity is greater than the preset threshold, it is determined that the bid document contains the bid-rigging behavior;

[0023] If the comprehensive similarity is less than or equal to a preset threshold, and the comprehensive similarity is greater than a target threshold, it is determined that the bid document is suspected of bid rigging.

[0024] Optionally, determining whether there are bids with the same semantic errors to determine whether the bids involve bid rigging includes:

[0025] If there are bids with the same semantic errors, it is determined that the bids contain the aforementioned bid-rigging behavior;

[0026] If there is no bid with the same semantic error, it is determined that the bid does not involve the bid rigging behavior.

[0027] Optionally, determining whether the bid document contains the bid-rigging behavior according to the quotation information and a predetermined project budget includes:

[0028] Comparing and analyzing the quotation information in each of the bids and the project budget, if only one of the bids has quotation information lower than the project budget, and the quotation information of the other bids is higher than the project budget, then determining whether the bid involves bid rigging by determining whether the project budget has been made public;

[0029] If the project budget has been publicly announced, the bid is judged to contain the bid-rigging behavior;

[0030] If the project budget has not been made public, the bid will be deemed to be suspected of bid rigging.

[0031] Optionally, the determining whether to send corresponding warning information to a preset purchaser terminal according to the judgment results of each bid-rigging behavior includes:

[0032] If it is determined that the bid document contains the bid-rigging behavior, a preset first-level warning message is sent to the purchaser terminal, so that the purchaser terminal can invalidate the bid document or blacklist the manufacturer corresponding to the bid document based on the preset first-level warning message;

[0033] If it is determined that the bid document is suspected of bid rigging, a preset second-level alarm message is sent to the purchaser terminal so that the purchaser terminal can re-examine the manufacturer corresponding to the bid document based on the preset second-level alarm message.

[0034] In a second aspect, the present application discloses a device for identifying bid-rigging behavior, comprising:

[0035] A first judgment result acquisition module is used to extract attribute information of each bid document, determine whether the final savers of each bid document are the same based on the attribute information, and determine whether there is bid rigging and collusion in the bid document according to the corresponding judgment result, so as to obtain a first bid rigging and collusion judgment result;

[0036] A second judgment result acquisition module is used to respectively determine the similarity of the text and image content in each of the bid documents, determine the comprehensive similarity based on the corresponding text similarity and image content similarity, and determine whether the bid document has bid rigging behavior according to the relationship between the comprehensive similarity and a preset threshold value, so as to obtain a second bid rigging behavior judgment result;

[0037] A third judgment result acquisition module is used to analyze the semantic errors in each of the bids, and determine whether the bids have the bid-rigging behavior by determining whether there are bids with the same semantic errors, so as to obtain a third bid-rigging behavior judgment result;

[0038] a fourth judgment result acquisition module, configured to extract quotation information from each of the bid documents, and determine whether the bid document contains the bid-rigging behavior according to the quotation information and a predetermined project budget, so as to obtain a fourth bid-rigging behavior judgment result;

[0039] The alarm information sending module is used to determine whether to send corresponding alarm information to the preset buyer terminal based on the judgment results of each bid-rigging behavior.

[0040] In a third aspect, the present application discloses an electronic device, comprising:

[0041] Memory, used to store computer programs;

[0042] A processor is used to execute the computer program to implement the aforementioned method for identifying bid-rigging behavior.

[0043] In a fourth aspect, the present application discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, the method for identifying bid-rigging behavior as described above is implemented.

[0044] When identifying bid-rigging and collusion, the present application first extracts attribute information of each bid document, determines whether the final preservers of each bid document are the same based on the attribute information, and determines whether the bid document has bid-rigging and collusion according to the corresponding judgment result, so as to obtain a first bid-rigging and collusion judgment result; then determines the similarity of the text and image content in each bid document respectively, determines the comprehensive similarity based on the corresponding text similarity and image content similarity, and determines whether the bid document has bid-rigging and collusion according to the size relationship between the comprehensive similarity and a preset threshold value, so as to obtain a second bid-rigging and collusion judgment result; analyzes the semantic errors in each bid document, and determines whether the bid document has bid-rigging and collusion by determining whether there are bid documents with the same semantic errors, so as to obtain a third bid-rigging and collusion judgment result; then extracts the quotation information in each bid document, and determines whether the bid document has bid-rigging and collusion according to the quotation information and a predetermined project budget, so as to obtain a fourth bid-rigging and collusion judgment result; finally, determines whether to send a corresponding alarm message to a preset purchaser terminal according to each bid-rigging and collusion judgment result. It can be seen that this application analyzes multiple dimensions of information such as the attribute information, text and image content, semantics, and quotation analysis of the tender documents. Based on information extraction, image recognition, similarity comparison and other means, it not only focuses on the direct comparison of the tender document content, but also deeply explores potential semantic errors and hidden information, and identifies the improper behaviors such as bid rigging and bid collusion that may exist in the bidding process in a multi-dimensional and high-precision manner. In this way, it is no longer limited to the manual processing of tender documents by staff under normal circumstances, which effectively alleviates the heavy workload of staff, significantly improves processing efficiency, and greatly reduces errors caused by human factors. It not only realizes the accuracy and efficiency of tender document review, but also can quickly capture potential problems in the bidding process, effectively improves the transparency and fairness of the bidding process, and provides strong support for building a healthier and fairer market competition environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0046] Figure 1 A flow chart of a method for identifying bid-rigging behavior disclosed in this application;

[0047] Figure 2 A schematic diagram of a specific method for identifying bid-rigging behavior disclosed in this application;

[0048] Figure 3 This is a schematic diagram of the structure of a device for identifying bid-rigging behavior disclosed in this application;

[0049] Figure 4 This is a structural diagram of an electronic device disclosed in this application. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] Traditional discipline inspection and supervision work is limited by the heavy task of manually processing bids, data and information, and often faces challenges such as heavy workload, low processing efficiency, and inevitable human errors. In order to solve the above technical problems, this application discloses a method, device, equipment and storage medium for identifying bid-rigging behavior, which can no longer be limited to manual processing of bids by staff under normal circumstances, effectively alleviate the heavy workload of staff, significantly improve processing efficiency, and greatly reduce errors caused by human factors, so as to achieve accurate and efficient bid review and quickly capture potential problems in the bidding process.

[0052] See also Figure 1 As shown, an embodiment of the present invention discloses a method for identifying bid-rigging behavior, including:

[0053] Step S11, extracting attribute information of each bid document, determining whether the final preservers of each bid document are the same based on the attribute information, and determining whether there is bid rigging or collusion in the bid document according to the corresponding judgment result, so as to obtain a first bid rigging judgment result.

[0054] In this embodiment, the application first extracts the attribute information of the bid, reads the attribute details of each supplier's bid through Apache POI, OpenOfficeAPI, Microsoft Graph API and other technologies, and efficiently parses and extracts the key attribute information in the bid, including but not limited to the original author of the bid, the final saver, the revision number and version number in the revision history, the company to which it belongs, and the specific storage time and other detailed data. Then, according to the extracted attribute information, it is determined whether the final savers of each bid are the same. If the final savers of each bid are the same, it is determined that there is bid collusion in the bid; if the final savers of each bid are different, it is determined whether the bid is a bid written by a single author; if the bid is a bid written by a single author and there are no traces of multiple edits and revisions in the bid, it is determined that the bid is suspected of bid collusion. Specifically, during the detailed review process, if it is identified that the final saver or submitter of multiple bid documents points to the same entity, it is highly alerted that there is a potential risk of bid collusion, that is, different suppliers may collude with each other through secret agreements or illegal means to manipulate the bidding process and undermine the fairness and transparency of the bidding results. In addition, further analysis can show that a certain bid document is dominated by a single author and lacks obvious signs of multiple editing and revisions, which is also worthy of vigilance. Such situations may mean that the preparation process of the bid document is not sufficient or rigorous, or even suspected of directly copying and pasting content from other bid documents. In this way, compared with traditional manual recognition methods, it not only significantly improves the accuracy and efficiency of information retrieval, but also provides great convenience for subsequent in-depth analysis and detailed comparison of bid documents, ensuring the comprehensive traceability of information.

[0055] Step S12, respectively determine the similarity of the text and image content in each of the bids, determine the comprehensive similarity based on the corresponding text similarity and image content similarity, and determine whether the bid involves bid rigging based on the relationship between the comprehensive similarity and a preset threshold value, so as to obtain a second bid rigging judgment result.

[0056] In this embodiment, the similarity of the text and image content in each bid is determined respectively, the format of each bid is unified, and the content of the bidding document and each bid is divided into several paragraphs; the paragraphs similar to the bidding document in each bid are deleted to obtain the bid after removal; the text in each bid after removal is processed by word segmentation, and the word segmentation similarity between the word segmentation results corresponding to each bid after removal is determined based on the cosine similarity algorithm, so as to determine the text similarity between each bid based on the word segmentation similarity; the image content of each bid is extracted by a computer vision large model to obtain the extracted features, and the image content similarity between each bid is calculated by cosine similarity based on the extracted features; the comprehensive similarity is determined based on the text similarity and the image content similarity. If the comprehensive similarity is greater than the preset threshold, it is determined that the bid has the suspicion of bid collusion; if the comprehensive similarity is less than or equal to the preset threshold, and the comprehensive similarity is greater than the target threshold, it is determined that the bid is suspected of bid collusion.

[0057] Specifically, when comparing the similarity of the text content of the tender documents, the similarity of the text content of the tender documents is first evaluated. First, the format of the tender documents is unified (both Word or PDF); then the tender documents and each tender document are divided into multiple logically independent paragraphs or chapters, and the paragraphs in each tender document that are similar to the tender document are excluded; the remaining parts of each tender document are reintegrated and compared two by two. The two paragraphs of text are segmented, and the segmentation results are converted into points in a high-dimensional vector space. The cosine similarity algorithm is used to calculate the cosine value of the angle between the two document vectors to measure the similarity between them. The closer the value is to 1, the higher the similarity between the two tender documents; finally, based on the similarity calculation results, the significant differences and subtle differences between the tender documents are intelligently identified, and they are divided into two categories according to the exact same and similar, which is convenient for staff to quickly review and make decisions. In addition, it should be pointed out that in order to ensure the accuracy of the comparison, the non-differentiated content commonly present in the tender documents is first identified and excluded, especially the common parts such as commercial terms and general requirements, so as to focus on the core differences such as the technical details and procurement plans of each bidding scheme, so as to improve the pertinence and accuracy of the tender content comparison. When comparing the similarity of the content of bid images, we integrate Optical Character Recognition (OCR) technology with the big model capabilities of Computer Vision (CV) to deeply analyze the similarity of the image content, so as to accurately identify whether there are highly similar screenshots. If the combined similarity of the text content and the image content exceeds 80%, it is determined that the bid involves bid rigging; if the similarity is between 60% and 80%, it is determined that the bid involves suspicion of bid rigging.

[0058] Step S13, analyzing the semantic errors in each of the bids, and determining whether there are bids with the same semantic errors to determine whether the bids involve bid rigging, so as to obtain a third bid rigging determination result.

[0059] In this embodiment, the semantic errors in the bids of each supplier are analyzed by natural language processing (NLP) technology to determine whether the same semantic errors appear in the bids of different suppliers, especially those unintentionally repeated and obvious errors, such as redundant expressions such as "including" and "having". If the same or highly similar errors appear independently in different bids, and these errors appear grammatically unconventional, it is determined that the bids have collusion; if there are no bids with the same semantic errors, it is determined that there is no collusion.

[0060] Step S14, extracting the quotation information from each of the bids, and determining whether the bids involve bid rigging based on the quotation information and a predetermined project budget, so as to obtain a fourth bid rigging judgment result.

[0061] In this embodiment, the application uses natural language processing technology to extract the quotation information in the bids of each supplier, and compares and analyzes the quotation information in each bid and the project budget. If there is only one bid with quotation information lower than the project budget, and the quotation information of other bids is higher than the project budget, then it is determined whether the project budget has been made public to judge whether there is bid-rigging behavior in the bid; if the project budget has been made public, it is determined that there is bid-rigging behavior in the bid; if the project budget has not been made public, it is determined that there is suspicion of bid-rigging behavior in the bid. Specifically, the quotation information in the bids of each supplier is extracted, and combined with the budget comparison and analysis of the procurement project, if only one supplier's quotation is lower than the budget range, and the quotations of all other suppliers are higher than the budget, there is an obvious suspicion of bid-rigging, that is, some suppliers may not participate in the bidding with real competitive intentions, but submit quotations higher than the budget for specific purposes (such as helping one party win the bid).

[0062] Step S15: Determine whether to send corresponding alarm information to the preset purchaser terminal according to the judgment results of each bid-rigging behavior.

[0063] In this embodiment, the present application obtains different judgment results of bid rigging and collusion through the judgment of the above four steps, and then determines whether to send corresponding alarm information to the preset purchaser terminal according to the judgment results of each bid rigging and collusion behavior. If it is determined that the bid document involves bid rigging and collusion, a preset first-level alarm information is sent to the purchaser terminal, so that the purchaser terminal can invalidate the bid document or blacklist the manufacturer corresponding to the bid document based on the preset first-level alarm information; if it is determined that the bid document is suspected of bid rigging and collusion, a preset second-level alarm information is sent to the purchaser terminal, so that the purchaser terminal can re-examine the manufacturer corresponding to the bid document based on the preset second-level alarm information.

[0064] In general, once this application finds that the final editor of each supplier's bid is the same person, it means that there may be serious bid-rigging behavior, and the system will immediately trigger a serious alarm; if the bid shows the distinct characteristics of being completed by a single author and lacks significant signs of multiple editing and revisions, there is reason to suspect that its content may be directly copied and pasted from other bids, and a secondary alarm will be issued at this time. At the same time, similarity analysis of the bid content is also indispensable. When the combined similarity of text and image content exceeds the threshold of 80%, the system will issue a serious alarm; if the similarity exceeds 60%, a secondary alarm will be triggered. In addition, if the same semantic errors appear in the bid, this often indicates that the possibility of reuse of the bid content is very high, and the system will issue a serious alarm. In the case of public budget, if only one supplier's quotation is lower than the budget range, and the rest of the suppliers' quotations are higher, such a situation undoubtedly points to obvious bid-rigging behavior, and the system will immediately issue a serious alarm; if the budget is not public, the same behavior cannot be directly characterized as bid-rigging, but there is also suspicion of bid-rigging, and the system will issue a secondary alarm accordingly.

[0065] In summary, this application analyzes multiple dimensions of information such as the attribute information, text and image content, semantics, and quotation analysis of the tender documents. Based on information extraction, image recognition, similarity comparison and other means, it not only focuses on the direct comparison of the tender document content, but also deeply explores potential semantic errors and hidden information, so as to identify the improper behaviors such as bid rigging and bid collusion that may exist in the bidding process in a multi-dimensional and high-precision manner. In this way, it is no longer limited to the manual processing of tender documents by staff under normal circumstances, which effectively alleviates the heavy workload of staff, significantly improves processing efficiency, and greatly reduces errors caused by human factors. It not only realizes the precision and efficiency of tender document review, but also can quickly capture potential problems in the bidding process, effectively improves the transparency and fairness of the bidding process, and provides strong support for building a healthier and fairer market competition environment.

[0066] Based on the above embodiment, it can be seen that the present application can accurately and efficiently identify bid rigging behavior. Next, the specific process of identifying bid rigging behavior will be described in detail. Figure 2 As shown, the present application discloses a specific method for identifying bid-rigging behavior.

[0067] First, read the detailed attribute information of each supplier's bid, including but not limited to the original author of the bid, the final saver, the revision number and version number in the revision history, the final save time, etc. Identify whether there are multiple final savers of the bid documents pointing to the same entity. If so, issue a serious alarm and notify the purchaser by email. If not, further analyze whether there is a bid that is dominated by a single author and lacks obvious signs of multiple edits and revisions. If so, issue a minor alarm and notify the purchaser by email.

[0068] Then compare the similarity of the bid text and the image content. Through technical means, the text and image content are compared and analyzed separately. If the combined similarity of the text content and the image content exceeds 80%, a serious alarm is issued and the buyer is notified by email. If the similarity is between 60% and 80%, a minor alarm is issued and the buyer is notified by email.

[0069] Next, the semantic errors in the bids of each supplier are analyzed to determine whether the same semantic errors appear in the bids of different suppliers. A serious alarm is then issued and the buyer is notified by email.

[0070] Finally, extract the quotation information from each supplier's bid, and compare and analyze it with the budget of the procurement project: if only one supplier's quotation is below the budget range, and all other suppliers' quotations are above the budget, there is an obvious suspicion of bid-straightening. If the budget has been announced, it is an obvious bid-straightening behavior, and a serious alarm will be issued and the purchaser will be notified by email; if the budget has not been announced, a minor alarm will be issued and the purchaser will be notified by email.

[0071] Specifically, this application issues different levels of alerts based on the severity of the problems found. Once there is an obvious violation of regulations or disciplines, the system will immediately trigger a severe alert, highlighting the severity and urgency of the problem. At this time, it is strongly recommended that procurement personnel consider canceling the bid or directly blacklisting the manufacturer to maintain a fair and just procurement environment. If the situation is still in doubt, but there is indeed suspicion of violations of regulations and disciplines, a secondary alert will be issued as a preliminary warning and reminder. For these suspicious suppliers, further in-depth analysis will be conducted, for example, a detailed review of the projects that the supplier has bid for in the past to fully assess its credibility and compliance.

[0072] In this way, this application conducts a comprehensive and in-depth analysis of the bids submitted by suppliers. It is not limited to the one-to-one comparison of the basic attributes and contents of the bids, but also creatively integrates multiple technical means such as attribute matching, text comparison, image review and semantic analysis. This process not only focuses on the direct comparison of the content of the bids, but also deeply explores potential semantic errors and possible hidden information, and identifies possible improper behaviors such as bid rigging and collusion in the bidding process in a multi-dimensional and high-precision manner, thereby effectively improving the transparency and fairness of the bidding process.

[0073] See also Figure 3 As shown, an embodiment of the present invention discloses a device for identifying bid-rigging behavior, including:

[0074] The first judgment result acquisition module 11 is used to extract attribute information of each bid document, determine whether the final savers of each bid document are the same based on the attribute information, and determine whether there is bid rigging behavior in the bid document according to the corresponding judgment result, so as to obtain a first bid rigging behavior judgment result;

[0075] The second judgment result acquisition module 12 is used to respectively determine the similarity of the text and image content in each of the bid documents, determine the comprehensive similarity based on the corresponding text similarity and image content similarity, and determine whether the bid document has bid rigging behavior according to the relationship between the comprehensive similarity and a preset threshold value, so as to obtain a second bid rigging behavior judgment result;

[0076] The third judgment result acquisition module 13 is used to analyze the semantic errors in each of the bids, and determine whether the bids have the bid-rigging behavior by determining whether there are bids with the same semantic errors, so as to obtain a third bid-rigging behavior judgment result;

[0077] The fourth judgment result acquisition module 14 is used to extract the quotation information in each of the bid documents, and determine whether the bid document contains the bid-rigging behavior according to the quotation information and the predetermined project budget, so as to obtain a fourth bid-rigging behavior judgment result;

[0078] The alarm information sending module 15 is used to determine whether to send corresponding alarm information to the preset purchaser terminal according to the judgment results of each bid-rigging behavior.

[0079] It can be seen that this application analyzes multiple dimensions of information such as the attribute information, text and image content, semantics, and quotation analysis of the tender documents. Based on information extraction, image recognition, similarity comparison and other means, it not only focuses on the direct comparison of the tender document content, but also deeply explores potential semantic errors and hidden information, and identifies the improper behaviors such as bid rigging and bid collusion that may exist in the bidding process in a multi-dimensional and high-precision manner. In this way, it is no longer limited to the manual processing of tender documents by staff under normal circumstances, which effectively alleviates the heavy workload of staff, significantly improves processing efficiency, and greatly reduces errors caused by human factors. It not only realizes the accuracy and efficiency of tender document review, but also can quickly capture potential problems in the bidding process, effectively improves the transparency and fairness of the bidding process, and provides strong support for building a healthier and fairer market competition environment.

[0080] In some specific embodiments, the first judgment result acquisition module 11 may specifically include:

[0081] a first bid-rigging behavior determination unit, configured to determine that the bid-rigging behavior exists in the bid-rigging if the final savers of the bid-rigging documents are the same;

[0082] a judgment unit, configured to judge whether the bid is written by a single author if the final savers of the bids are different;

[0083] The second bid-rigging behavior determination unit is used to determine that the bid is suspected of bid-rigging behavior if the bid is written by a single author and there are no traces of multiple editing and revisions in the bid.

[0084] In some specific embodiments, the second judgment result acquisition module 12 may specifically include:

[0085] A division unit is used to unify the formats of the tender documents and divide the contents of the tender documents and the tender documents into several paragraphs;

[0086] A paragraph deleting unit, used for deleting paragraphs in each of the bids that are similar to the bidding document, so as to obtain a bid after the deletion;

[0087] A word segmentation similarity determination unit is used to perform word segmentation processing on the text in each of the removed bids, and determine the word segmentation similarity between the word segmentation results corresponding to each of the removed bids based on a cosine similarity algorithm, so as to determine the text similarity between each of the bids based on the word segmentation similarity;

[0088] A content similarity determination unit, configured to extract features of the image content of each of the bid documents through a computer vision model, obtain extracted features, and calculate the image content similarity between the bid documents through cosine similarity based on the extracted features;

[0089] The comprehensive similarity determination unit is used to determine the comprehensive similarity according to the text similarity and the image content similarity.

[0090] In some specific embodiments, the second judgment result acquisition module 12 may specifically include:

[0091] A third bid-rigging behavior determination unit, configured to determine that the bid document contains the bid-rigging behavior if the comprehensive similarity is greater than the preset threshold;

[0092] The fourth bid-rigging behavior determination unit is used to determine that the bid document is suspected of bid-rigging behavior if the comprehensive similarity is less than or equal to a preset threshold and the comprehensive similarity is greater than a target threshold.

[0093] In some specific embodiments, the third judgment result acquisition module 13 may specifically include:

[0094] a fifth bid-rigging behavior determination unit, configured to determine, if there are bids with the same semantic errors, that the bids involve the bid-rigging behavior;

[0095] The sixth bid rigging behavior determination unit is used to determine that the bid does not involve bid rigging if there is no bid with the same semantic error.

[0096] In some specific embodiments, the fourth judgment result acquisition module 14 may specifically include:

[0097] a bid-rigging behavior judgment unit, configured to compare and analyze the quotation information in each of the bids and the project budget, and if only one of the bids has quotation information lower than the project budget, and the quotation information of the other bids is higher than the project budget, then determining whether the bid involves bid-rigging behavior by determining whether the project budget is publicly disclosed;

[0098] A seventh bid-rigging behavior determination unit, configured to determine if the project budget has been publicly disclosed, that the bid document contains the bid-rigging behavior;

[0099] The eighth bid rigging behavior determination unit is used to determine whether the bid document is suspected of bid rigging behavior if the project budget has not been disclosed to the public.

[0100] In some specific embodiments, the alarm information sending module 15 may specifically include:

[0101] A first-level warning information sending unit is used to send a preset first-level warning information to the buyer terminal if it is determined that the bid document contains the bid-rigging behavior, so that the buyer terminal can invalidate the bid document or blacklist the manufacturer corresponding to the bid document based on the preset first-level warning information;

[0102] The secondary alarm information sending unit is used to send preset secondary alarm information to the buyer's terminal if it is determined that the bid is suspected of bid rigging, so that the buyer's terminal can re-examine the manufacturer corresponding to the bid based on the preset secondary alarm information.

[0103] Furthermore, the present application also discloses an electronic device. Figure 4 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content in the diagram cannot be regarded as any limitation on the scope of use of the present application.

[0104] Figure 4 A schematic diagram of the structure of an electronic device 20 provided in an embodiment of the present application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 is used to store a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the method for identifying bid-rigging behavior disclosed in any of the aforementioned embodiments. In addition, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0105] In this embodiment, the power supply 23 is used to provide working voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and the external device, and the communication protocol it follows is any communication protocol that can be applied to the technical solution of the present application, and is not specifically limited here; the input and output interface 25 is used to obtain external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs and is not specifically limited here.

[0106] In addition, the memory 22 as a carrier for resource storage may be a read-only memory, a random access memory, a disk or an optical disk, etc. The resources stored thereon may include an operating system 221, a computer program 222, etc., and the storage method may be temporary storage or permanent storage.

[0107] The operating system 221 is used to manage and control the hardware devices and computer program 222 on the electronic device 20, which can be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program that can be used to complete the bid-rigging behavior identification method performed by the electronic device 20 disclosed in any of the aforementioned embodiments, the computer program 222 can further include a computer program that can be used to complete other specific tasks.

[0108] Furthermore, the present application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, the aforementioned disclosed method for identifying bid-rigging behavior is implemented. The specific steps of the method can be referred to the corresponding contents disclosed in the aforementioned embodiments, and will not be repeated here.

[0109] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0110] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0111] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0112] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0113] The technical solution provided by the present application is introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea. At the same time, for general technicians in this field, according to the idea of ​​the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A method for identifying bid-rigging behavior, characterized in that: include: Extracting attribute information of each bid document, determining whether the final savers of each bid document are the same based on the attribute information, and determining whether there is bid rigging and collusion in the bid document according to the corresponding determination result, so as to obtain a first bid rigging and collusion determination result; Determine the similarity of the text and image content in each of the bid documents respectively, determine the comprehensive similarity based on the corresponding text similarity and image content similarity, and determine whether the bid document involves bid rigging according to the relationship between the comprehensive similarity and a preset threshold value, so as to obtain a second bid rigging judgment result; Analyze the semantic errors in each of the bids, and determine whether the bids contain the bid-rigging behavior by determining whether there are bids with the same semantic errors, so as to obtain a third bid-rigging behavior determination result; Extracting quotation information from each of the bids, and determining whether the bid involves bid rigging based on the quotation information and a predetermined project budget, so as to obtain a fourth bid rigging judgment result; Determine whether to send corresponding alarm information to the preset buyer terminal based on the judgment results of each bid-rigging behavior.

2. The method for identifying bid-rigging behavior according to claim 1, characterized in that: The attribute information includes the original author of the bid, the final saver, the revision number and version number in the revision history, the company to which it belongs, and the saving time; Accordingly, determining whether there is bid-rigging behavior in the bid document according to the corresponding judgment result includes: If the final depositor of each of the bid documents is the same, it is determined that the bid documents involve bid rigging; If the final depositors of the bid documents are different, determining whether the bid documents are written by a single author; If the bid is written by a single author and there are no signs of multiple editing and revisions in the bid, it will be determined that the bid is suspected of bid rigging.

3. The method for identifying bid-rigging behavior according to claim 1, characterized in that: The step of respectively determining the similarity of the text and the image content in each of the bid documents, and determining the comprehensive similarity based on the corresponding text similarity and image content similarity, includes: Standardize the format of the tender documents and divide the contents of the bidding documents and the tender documents into several paragraphs; Deleting the paragraphs in each of the said bids that are similar to the said bidding documents to obtain the bids after deletion; Performing word segmentation processing on the text in each of the removed bids, and determining the word segmentation similarity between the word segmentation results corresponding to each of the removed bids based on the cosine similarity algorithm, so as to determine the text similarity between each of the bids based on the word segmentation similarity; Extracting features from the image content of each of the bid documents through a computer vision model to obtain extracted features, and calculating the image content similarity between the bid documents through cosine similarity based on the extracted features; The comprehensive similarity is determined according to the text similarity and the image content similarity.

4. The method for identifying bid-rigging behavior according to claim 1, characterized in that: The determining whether there is bid rigging in the bid document according to the size relationship between the comprehensive similarity and the preset threshold value includes: If the comprehensive similarity is greater than the preset threshold, it is determined that the bid document contains the bid-rigging behavior; If the comprehensive similarity is less than or equal to a preset threshold, and the comprehensive similarity is greater than a target threshold, it is determined that the bid document is suspected of bid rigging.

5. The method for identifying bid-rigging behavior according to claim 1, characterized in that: The determining whether there are bids with the same semantic errors to determine whether the bids involve bid rigging includes: If there are bids with the same semantic errors, it is determined that the bids contain the aforementioned bid-rigging behavior; If there is no bid with the same semantic error, it is determined that the bid does not involve the bid rigging behavior.

6. The method for identifying bid-rigging behavior according to claim 1, characterized in that: Determining whether the bid document contains the bid-rigging behavior according to the quotation information and the predetermined project budget includes: Comparing and analyzing the quotation information in each of the bids and the project budget, if only one of the bids has quotation information lower than the project budget, and the quotation information of the other bids is higher than the project budget, then determining whether the bid involves bid rigging by determining whether the project budget has been made public; If the project budget has been publicly announced, the bid is judged to contain the bid-rigging behavior; If the project budget has not been made public, the bid will be deemed to be suspected of bid rigging.

7. The method for identifying bid-rigging behavior according to any one of claims 1 to 6, characterized in that: The determining whether to send corresponding warning information to the preset purchaser terminal according to the judgment results of each bid-rigging behavior includes: If it is determined that the bid document contains the bid-rigging behavior, a preset first-level warning message is sent to the purchaser terminal, so that the purchaser terminal can invalidate the bid document or blacklist the manufacturer corresponding to the bid document based on the preset first-level warning message; If it is determined that the bid document is suspected of bid rigging, a preset second-level alarm message is sent to the purchaser terminal so that the purchaser terminal can re-examine the manufacturer corresponding to the bid document based on the preset second-level alarm message.

8. A device for identifying bid-rigging behavior, characterized in that: include: A first judgment result acquisition module is used to extract attribute information of each bid document, determine whether the final savers of each bid document are the same based on the attribute information, and determine whether there is bid rigging and collusion in the bid document according to the corresponding judgment result, so as to obtain a first bid rigging and collusion judgment result; A second judgment result acquisition module is used to respectively determine the similarity of the text and image content in each of the bid documents, determine the comprehensive similarity based on the corresponding text similarity and image content similarity, and determine whether the bid document has bid rigging behavior according to the relationship between the comprehensive similarity and a preset threshold value, so as to obtain a second bid rigging behavior judgment result; A third judgment result acquisition module is used to analyze the semantic errors in each of the bids, and determine whether the bids have the bid-rigging behavior by determining whether there are bids with the same semantic errors, so as to obtain a third bid-rigging behavior judgment result; a fourth judgment result acquisition module, configured to extract quotation information from each of the bid documents, and determine whether the bid document contains the bid-rigging behavior according to the quotation information and a predetermined project budget, so as to obtain a fourth bid-rigging behavior judgment result; The alarm information sending module is used to determine whether to send corresponding alarm information to the preset buyer terminal based on the judgment results of each bid-rigging behavior.

9. An electronic device, characterized in that: include: Memory, used to store computer programs; A processor is used to execute the computer program to implement the steps of the method for identifying bid-rigging behavior as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: Used to store a computer program; wherein, when the computer program is executed by a processor, the steps of the method for identifying bid-rigging behavior as described in any one of claims 1 to 7 are implemented.

Images