Anti-cross-labeling intelligent atlas identification method and device, electronic equipment and storage medium

Abstract

The invention relates to an anti-cross-labeling intelligent atlas identification method and device, electronic equipment and a storage medium. The anti-cross-bidding intelligent atlas identification method comprises the following steps: collecting related data in a bidding and tendering process; the data is cleaned, integrated and formatted, so that subsequent graph construction and analysis are facilitated; a bidding data atlas is constructed based on the collected data, nodes represent data entities including bidders and bidding files, and sides represent the relationships including bidding relationships, association relationships and bidding file repetition similarity degrees among the entities; extracting features related to the trademark behavior from the atlas; training the extracted features by using a machine learning algorithm, and constructing a string mark behavior recognition model; evaluating the trained model, and performing parameter adjustment and optimization according to an evaluation result; and the trained model is used to carry out automatic detection of the string marking behavior. According to the anti-cross-labeling intelligent atlas identification method, cross-labeling behaviors can be automatically, efficiently and accurately identified.

Description

Technical Field

[0001] The present invention relates to the field of data processing and image recognition, and in particular to a method, device, electronic device and storage medium for intelligent graph recognition of bid-rigging behaviors that may occur during a bidding process. Background Art

[0002] Collusion in bidding refers to the act of colluding in bidding and tendering, which is the act of using improper means between the tenderer and the bidder, or between the bidders, to collude on bidding and tendering matters in order to exclude competitors or damage the interests of the tenderer.

[0003] In the bidding and tendering sector, bid rigging seriously undermines fair competition in the market. Traditional methods for identifying bid rigging rely primarily on manual review, but this approach is inefficient, error-prone, and difficult to detect complex bid rigging tactics. Furthermore, it requires a high level of professional expertise and experience from reviewers.

[0004] An intelligent graph is a graphical data structure designed to capture and represent the relationships between an organization's goals, strategies, activities, and outcomes. It can be thought of as a graph-based representation of an organization's purpose, with nodes representing the components of the purpose and edges representing the relationships between those components.

[0005] Therefore, it is urgent to develop an intelligent graph recognition method and device that can automatically, efficiently and accurately identify bid-rigging behavior, which does not rely on manual review, can automatically collect, process and analyze data, and realize automatic detection of bid-rigging behavior. Summary of the Invention

[0006] The technical problem to be solved by the present invention is that in the field of bidding, bid collusion seriously affects fair competition in the market; traditional bid collusion identification methods mainly rely on manual review, but this method is inefficient, prone to errors, and difficult to deal with complex bid collusion methods.

[0007] In order to solve the above technical problems, according to one aspect of the present invention, a method for anti-coercive bidding intelligent graph recognition is provided, which includes the following steps: S1, data collection, collecting relevant data in the bidding process, the relevant data including supplier contact information, authorization content, bidding document content, and document preparation information submitted by suppliers when registering; S2, data preprocessing, cleaning, integrating and formatting the data to facilitate subsequent graph construction and analysis; S3, graph construction, constructing a bidding data graph based on the collected data, the graph includes nodes and edges, nodes represent data entities including bidders and bidding documents, edges represent the connection between entities including bidding relationships, association relationships, and the weight of the data content of the bidding documents The relationship between the degree of complex similarity; S4, feature extraction, extracting features related to bid-collusion behavior from the graph, the features include the bidder's historical bidding records, bidding document preparation tools, time and content similarity, among which the historical bidding records include registration information for participating in different projects at different times; S5, model training, using machine learning algorithms to train the extracted features and build a bid-collusion behavior recognition model; S6, model evaluation and optimization, evaluating the trained model, and adjusting and optimizing parameters according to the evaluation results; S7, bid-collusion recognition, using the trained model to automatically detect bid-collusion behavior, inputting the bidding data to be detected into the trained model for automatic detection of bid-collusion behavior.

[0008] According to an embodiment of the present invention, in step S1, the file compilation information may include: author, user, writer, creation time in PDF document properties, and author, last saver, creation time in Word document properties.

[0009] According to an embodiment of the present invention, in step S5, the extracted features may be trained using a machine learning algorithm including a support vector machine and a random forest.

[0010] According to an embodiment of the present invention, in step S6, the trained model may be evaluated by cross-validation.

[0011] According to the second aspect of the present invention, there is provided a device for intelligent graph recognition to prevent bid collusion, which comprises: a data collection module for collecting relevant data in the bidding process, the relevant data including supplier contact information, authorization content, bidding document content, and document compilation information submitted by suppliers when they register; a data preprocessing module for cleaning, integrating and formatting the data to facilitate subsequent graph construction and analysis; a graph construction module for constructing a bidding data graph based on the collected data, the graph comprising nodes and edges, the nodes representing data entities including bidders and bidding documents, and the edges representing the entities including bidding relationships, association relationships, and bidding documents. The relationship between the degree of repetition and similarity of the content of the data; the feature extraction module is used to extract features related to bid-collusion behavior from the graph, and the features include the bidder's historical bidding records, bidding document preparation tools, time and content similarity; the model training module is used to train the extracted features using machine learning algorithms and build a bid-collusion behavior recognition model; the model evaluation and optimization module is used to evaluate the trained model and adjust and optimize parameters according to the evaluation results; the bid-collusion recognition module is used to automatically detect bid-collusion behavior using the trained model, and input the bidding data to be detected into the trained model to automatically detect bid-collusion behavior.

[0012] According to an embodiment of the present invention, the file compilation information collected by the data collection module may include: author, user, writer, creation time in PDF document properties, author, last saver, content creation time in Word document properties.

[0013] According to an embodiment of the present invention, the model training module may use a machine learning algorithm including a support vector machine and a random forest to train the extracted features.

[0014] According to an embodiment of the present invention, the model evaluation and optimization module can evaluate the trained model through cross-validation.

[0015] According to a third aspect of the present invention, an electronic device is provided, comprising: a memory, a processor, and an anti-label tampering intelligent graph recognition program stored in the memory and executable on the processor, wherein the anti-label tampering intelligent graph recognition program implements the steps of the above-mentioned anti-label tampering intelligent graph recognition method when executed by the processor.

[0016] According to a fourth aspect of the present invention, a computer storage medium is provided, wherein a program for intelligent graph recognition against label tampering is stored on the computer storage medium, and when the program for intelligent graph recognition against label tampering is executed by a processor, the steps of the above-mentioned method for intelligent graph recognition against label tampering are implemented.

[0017] Compared with the prior art, the technical solution provided by the embodiments of the present invention can achieve at least the following beneficial effects:

[0018] The method for anti-label collusion intelligent graph recognition according to the present invention has a high degree of automation, can automatically collect, process and analyze data, and realize automatic detection of label collusion behavior, which greatly improves work efficiency, saves time, and reduces the requirements for the professional ability of auditors.

[0019] The method for anti-bid-rigging intelligent graph recognition according to the present invention has high accuracy. By constructing a bidding data graph and using a machine learning algorithm for pattern recognition, the present invention can accurately identify bid-rigging behavior and reduce the false alarm rate.

[0020] The method for anti-fraud intelligent graph recognition according to the present invention is highly flexible. The present invention can train and optimize the model according to actual needs to adapt to the requirements of fraudulent labeling behavior recognition in different scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings of the embodiments will be briefly introduced below. Obviously, the drawings in the following description only relate to some embodiments of the present invention, but are not intended to limit the present invention.

[0022] Figure 1 The present invention is a flowchart illustrating a method for intelligent pattern recognition to prevent label tampering according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of 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 described 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.

[0024] Unless otherwise defined, technical or scientific terms used herein shall have the ordinary meaning as understood by persons of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in the present patent application specification and claims do not denote any order, quantity, or importance, but are merely used to distinguish different components. Similarly, terms such as "a" or "an" do not denote a limitation of quantity, but rather denote the presence of at least one.

[0025] Figure 1 The present invention is a flowchart illustrating a method for intelligent pattern recognition to prevent label tampering according to an embodiment of the present invention.

[0026] like Figure 1 As shown, the method for anti-cross-label intelligent pattern recognition includes the following steps:

[0027] S1. Data collection: collect relevant data during the bidding process. The relevant data includes supplier contact information, authorization letter content, bidding document content, and document preparation information submitted by suppliers when they register.

[0028] S2. Data preprocessing: cleaning, integrating, and formatting the data to facilitate subsequent map construction and analysis.

[0029] S3. Graph construction: A bidding data graph is constructed based on the collected data. The graph includes nodes and edges. Nodes represent data entities including bidders and bidding documents, and edges represent relationships between entities including bidding relationships, association relationships, and the degree of repetition and similarity of the content of bidding documents.

[0030] S4. Feature extraction: extract features related to bid-rigging behavior from the graph. The features include the bidder's historical bidding records, bidding document preparation tools, time and content similarity. Among them, historical bidding records include registration information for participating in different projects at different times.

[0031] S5. Model training: Use machine learning algorithms to train the extracted features and build a cross-label behavior recognition model.

[0032] S6. Model evaluation and optimization: Evaluate the trained model and adjust and optimize parameters based on the evaluation results.

[0033] S7. Identify bid-rigging. Use the trained model to automatically detect bid-rigging behavior. Input the bidding data to be detected into the trained model to automatically detect bid-rigging behavior.

[0034] According to one or some embodiments of the present invention, in step S1, the file compilation information includes: author, user, writer, creation time in PDF document properties, and author, last saver, content creation time in Word document properties.

[0035] According to one or some embodiments of the present invention, in step S5, the extracted features are trained using a machine learning algorithm including a support vector machine and a random forest.

[0036] The method for anti-bid-rigging intelligent graph recognition according to the present invention has high accuracy. By constructing a bidding data graph and using a machine learning algorithm for pattern recognition, the present invention can accurately identify bid-rigging behavior and reduce the false alarm rate.

[0037] According to one or some embodiments of the present invention, in step S6, the trained model is evaluated by cross-validation.

[0038] According to a second aspect of the present invention, there is provided a device for anti-label tampering intelligent pattern recognition, comprising:

[0039] The data collection module is used to collect relevant data during the bidding process, including supplier contact information, authorization letter content, bidding document content, and document preparation information submitted by suppliers when they register;

[0040] Data preprocessing module, used to clean, integrate and format data to facilitate subsequent map construction and analysis;

[0041] A graph construction module is used to construct a bidding data graph based on the collected data. The graph includes nodes and edges. Nodes represent data entities such as bidders and bidding documents, while edges represent relationships between entities, including bidding relationships, association relationships, and the degree of repetition and similarity of the content of bidding documents.

[0042] The feature extraction module is used to extract features related to bid-rigging behavior from the graph, including the bidder's historical bidding records, bidding document preparation tools, time and content similarity;

[0043] The model training module is used to train the extracted features using machine learning algorithms to build a fraudulent behavior recognition model;

[0044] Model evaluation and optimization module, used to evaluate the trained model and adjust and optimize parameters based on the evaluation results;

[0045] The bid-rigging identification module is used to automatically detect bid-rigging behavior using a trained model. The bidding data to be detected is input into the trained model to automatically detect bid-rigging behavior.

[0046] The method for anti-label collusion intelligent graph recognition according to the present invention has a high degree of automation, can automatically collect, process and analyze data, and realize automatic detection of label collusion behavior, which greatly improves work efficiency, saves time, and reduces the requirements for the professional ability of auditors.

[0047] According to one or some embodiments of the present invention, the file compilation information collected by the data collection module includes: author, user, writer, creation time in PDF document properties, and author, last saver, content creation time in Word document properties.

[0048] According to one or some embodiments of the present invention, the model training module uses machine learning algorithms including support vector machines and random forests to train the extracted features.

[0049] According to one or some embodiments of the present invention, the model evaluation and optimization module evaluates the trained model through cross-validation.

[0050] The method and device for anti-fraud intelligent graph recognition according to the present invention are highly flexible. The present invention can train and optimize the model according to actual needs to adapt to the requirements of fraudulent labeling behavior recognition in different scenarios.

[0051] According to another aspect of the present invention, a device for anti-label tampering intelligent graph recognition is provided, which includes: a memory, a processor, and an anti-label tampering intelligent graph recognition program stored in the memory and runnable on the processor. When the anti-label tampering intelligent graph recognition program is executed by the processor, the steps of the above-mentioned anti-label tampering intelligent graph recognition method are implemented.

[0052] According to the present invention, a computer storage medium is also provided.

[0053] The computer storage medium stores an anti-label tampering intelligent graph recognition program, and when the anti-label tampering intelligent graph recognition program is executed by the processor, the steps of the anti-label tampering intelligent graph recognition method are implemented.

[0054] Among them, the method implemented when the anti-label tampering intelligent graph recognition program running on the processor is executed can refer to the various embodiments of the anti-label tampering intelligent graph recognition method of the present invention, and will not be repeated here.

[0055] The present invention also provides a computer program product.

[0056] The computer program product of the present invention includes an anti-label tampering intelligent graph recognition program, which implements the steps of the anti-label tampering intelligent graph recognition method described above when executed by a processor.

[0057] Among them, the method implemented when the anti-label tampering intelligent graph recognition program running on the processor is executed can refer to the various embodiments of the anti-label tampering intelligent graph recognition method of the present invention, and will not be repeated here.

[0058] Through the description of the above embodiments, those skilled in the art can clearly understand that the above embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better embodiment. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0059] The above description is merely an exemplary embodiment of the present invention and is not intended to limit the scope of protection of the present invention. The scope of protection of the present invention is determined by the appended claims.

Claims

1. A method for anti-label cross-labeling intelligent pattern recognition, comprising the following steps: S1. Data collection: collect relevant data during the bidding process, including supplier contact information, authorization content, bidding document content, and document preparation information submitted by suppliers when registering; S2, data preprocessing, cleaning, integration and formatting of data to facilitate subsequent map construction and analysis; S3. Graph construction: A bidding data graph is constructed based on the collected data. The graph includes nodes and edges. Nodes represent data entities including bidders and bidding documents, and edges represent relationships between entities including bidding relationships, association relationships, and the degree of repetition and similarity of the content of bidding documents. S4, feature extraction, extracting features related to bid-rigging behavior from the graph, the features including the bidder's historical bidding records, bidding document preparation tools, time and content similarity; S5, model training, using machine learning algorithms to train the extracted features and build a fraudulent behavior recognition model; S6, model evaluation and optimization, evaluate the trained model, and adjust and optimize parameters based on the evaluation results; S7. Identify bid-rigging by using the trained model to automatically detect bid-rigging behavior. Input the bidding data to be detected into the trained model to automatically detect bid-rigging behavior.

2. The method for anti-fraud intelligent graph recognition according to claim 1, in step S1, the file compilation information includes: The author, user, writer, and creation time in the PDF document properties; the author, last saver, and content creation time in the Word document properties.

3. The method for anti-fraud intelligent graph recognition as described in claim 1, in step S5, the extracted features are trained using a machine learning algorithm including a support vector machine and a random forest.

4. The method for anti-labeling intelligent graph recognition as described in claim 1, in step S6, the trained model is evaluated by cross-validation.

5. A device for preventing label tampering with intelligent graph recognition, comprising: The data collection module is used to collect relevant data in the bidding process, including supplier contact information, authorization content, bidding document content, and document preparation information submitted by suppliers when registering; Data preprocessing module, used to clean, integrate and format data for subsequent map construction and analysis; A graph construction module is used to construct a bidding data graph based on the collected data. The graph includes nodes and edges. Nodes represent data entities including bidders and bidding documents, and edges represent relationships between entities including bidding relationships, association relationships, and the degree of repetition and similarity of the content of bidding documents. A feature extraction module is used to extract features related to bid-rigging behavior from the graph, including the bidder's historical bidding records, bidding document preparation tools, time and content similarity; Model training module, used to train the extracted features using machine learning algorithms and build a fraudulent behavior recognition model; Model evaluation and optimization module, used to evaluate the trained model and adjust and optimize parameters based on the evaluation results; The bid-rigging identification module is used to automatically detect bid-rigging behavior using a trained model. The bidding data to be detected is input into the trained model to automatically detect bid-rigging behavior.

6. The device for preventing label fraud according to claim 5, wherein: The file compilation information collected by the data collection module includes: the author, user, writer, creation time in the PDF document attributes, and the author, last saver, creation time in the word document attributes.

7. The device for preventing label fraud according to claim 5, wherein: The model training module uses machine learning algorithms including support vector machines and random forests to train the extracted features.

8. The device for preventing label fraud according to claim 5, wherein: The model evaluation and optimization module evaluates the trained model by cross-validation.

9. An electronic device, comprising: A memory, a processor, and an anti-label tampering intelligent graph recognition program stored in the memory and executable on the processor, wherein the anti-label tampering intelligent graph recognition program, when executed by the processor, implements the steps of the anti-label tampering intelligent graph recognition method as described in any one of claims 1 to 4.

10. A computer storage medium, wherein: The computer storage medium stores an anti-label tampering intelligent graph recognition program, and when the anti-label tampering intelligent graph recognition program is executed by the processor, the steps of the anti-label tampering intelligent graph recognition method as described in any one of claims 1 to 4 are implemented.

Images