Government procurement risk supervision prediction method and system based on artificial intelligence learning
Through the government procurement risk supervision and prediction method based on artificial intelligence, a procurement risk prediction model is established, which solves the challenges of risk monitoring and preventing dishonest behavior in government procurement, realizes automatic risk monitoring and rapid response, and improves prevention capabilities.
Patent Information
- Application Number
- CN202410244482.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-05
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-03-05
AI Technical Summary
In the process of government procurement, how to monitor risks in accordance with laws and regulations and prevent dishonest behaviors and improve prevention capabilities has become an important challenge.
Adopt government procurement risk supervision and prediction methods based on artificial intelligence learning, and establish procurement risk prediction models to conduct risk prediction and grade assessment of effective procurement projects, generate risk warnings and provide targeted processing guidance to ensure full-process risks are carried out at different procurement stages. Monitor and prevent.
It realizes automatic risk monitoring of government procurement projects, promptly detects dishonesty problems, reduces losses caused by dishonest behavior, improves the prevention capabilities within the government, and quickly reminds relevant responsible persons to take measures to avoid the expansion of risks.
Smart Images

Figure CN117993571B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of procurement management, and in particular to a government procurement risk supervision and prediction method and system based on artificial intelligence learning. Background Art
[0002] In recent years, in order to improve the ability of government procurement projects in emergency response and social stability, various places have invested a lot of resources in the procurement of related project equipment and equipment. The government often implements hundreds of thousands of projects through bidding and procurement, involving huge amounts of money, becoming the largest procurement force in the province's cities and having a high influence in provincial cities across the country. However, in the huge scale of procurement, how to conduct procurement in accordance with laws and regulations, improve internal prevention capabilities, and prevent dishonest behavior in the procurement process has become an issue that all purchasers need to pay attention to. Therefore, the present invention proposes a government procurement risk supervision prediction method and system based on artificial intelligence learning. Summary of the invention
[0003] The present invention provides a government procurement risk supervision and prediction method and system based on artificial intelligence learning, establishes a procurement risk prediction model, realizes the whole process risk monitoring of procurement projects, can timely discover procurement dishonesty problems, prevent and reduce losses caused by dishonest behavior in the procurement process as much as possible, effectively improve the prevention capabilities within the government, and quickly remind the relevant responsible persons of the risky procurement projects when risks are discovered, so as to facilitate the adoption of corresponding measures to avoid further expansion of risks, and generate risk project processing results based on risk level assessment results and the responsibilities and authorities of each link of the risky procurement project, so as to provide targeted guidance and specifications for the processing of risky procurement projects.
[0004] The present invention provides a government procurement risk supervision and prediction method based on artificial intelligence learning, comprising:
[0005] Step 1: Based on the procurement risk prediction model, predict the risk of valid procurement projects, obtain the abnormal risk value of the project, and determine whether the valid procurement project is a risky procurement project;
[0006] Step 2: When a valid procurement project is determined to be a risky procurement project, a risk level assessment is conducted on the risky procurement project. Based on the risk level assessment results, a risk warning is generated and sent to the target person;
[0007] Step 3: Generate risk project processing results based on the risk level assessment results and the responsibilities and authorities of each link of the risk procurement project.
[0008] Preferably, in a government procurement risk supervision and prediction method based on artificial intelligence learning, step 1 includes:
[0009] Obtain a list of valid procurement items at preset time intervals;
[0010] Based on the procurement risk prediction model, risk prediction is performed on multiple valid procurement projects in the valid procurement project list in a preset order to obtain abnormal risk values of the projects;
[0011] Based on the abnormal risk value of the project, risk assessment is performed on multiple valid procurement projects;
[0012] When the abnormal risk value of a project is greater than the preset value, the valid procurement project is determined to be a risky procurement project;
[0013] Otherwise, the valid procurement project is determined to be a non-risk procurement project.
[0014] Preferably, in a government procurement risk supervision and prediction method based on artificial intelligence learning, it also includes:
[0015] Collect global historical procurement data corresponding to a large number of historical risk procurement projects, and obtain the processing results of each historical risk procurement project;
[0016] Determine the person in charge of each procurement sub-project of the historical risk procurement project, obtain the list of responsible persons, compare the penalty list in the processing results with the list of responsible persons, and determine the problem sub-projects corresponding to the risk procurement project;
[0017] According to the problem sub-projects, risk mark the procurement data corresponding to each problem sub-project in the historical procurement data to determine the risk data corresponding to each procurement project;
[0018] Extract risk features from risk data to obtain feature data sets, and build training sets and test sets based on the feature data sets;
[0019] The deep learning neural network is trained and tested based on the training set and the test set to obtain the procurement risk prediction model.
[0020] Preferably, in a government procurement risk supervision and prediction method based on artificial intelligence learning, step 2 includes:
[0021] Obtain the risk value corresponding to the risky procurement project, determine the risk level corresponding to the risky procurement project in combination with the preset level classification, and generate a risk warning of the corresponding level based on the risk level;
[0022] At the same time, obtain the list of project responsible persons corresponding to the risky procurement project, and generate a list of target persons for risk warning based on the list of project responsible persons;
[0023] Risk warnings are sent synchronously based on the target person list, and warning records are generated for storage.
[0024] Preferably, in a government procurement risk supervision and prediction method based on artificial intelligence learning, while obtaining a list of project responsible persons corresponding to risky procurement projects and generating a list of target persons for risk warning based on the list of project responsible persons, it also includes:
[0025] Obtain the risk value corresponding to the risky procurement project, determine the risk level corresponding to the risky procurement project in combination with the preset level classification, and generate a risk warning of the corresponding level based on the risk level.
[0026] The present invention provides a government procurement risk supervision and prediction system based on artificial intelligence learning, comprising:
[0027] The risk judgment module is used to predict the risks of effective procurement projects based on the procurement risk prediction model, obtain the abnormal risk value of the project, and judge whether the effective procurement project is a risky procurement project;
[0028] The risk warning module is used to evaluate the risk level of a valid procurement project when it is determined that the project is a risky procurement project. According to the risk level evaluation results, a risk warning is generated and sent to the target person;
[0029] The processing generation module is used to generate risk project processing results based on the risk level assessment results and the responsibilities and authorities of each link of the risk procurement project.
[0030] Preferably, in a government procurement risk supervision and prediction system based on artificial intelligence learning, the risk judgment module includes:
[0031] An autonomous acquisition unit, used to obtain a list of valid procurement items at preset time intervals;
[0032] An intelligent prediction unit, used to perform risk prediction on multiple valid procurement projects in the valid procurement project list in a preset order based on a procurement risk prediction model, and obtain abnormal risk values of the projects;
[0033] Intelligent judgment unit, used to make risk judgments on multiple valid procurement projects based on the abnormal risk value of the project;
[0034] When the abnormal risk value of a project is greater than the preset value, the valid procurement project is determined to be a risky procurement project;
[0035] Otherwise, the valid procurement project is determined to be a non-risk procurement project.
[0036] Preferably, in a government procurement risk supervision and prediction system based on artificial intelligence learning, it also includes:
[0037] The forecast preparation model is used to collect global historical procurement data corresponding to a large number of historical risk procurement projects and obtain the processing results of each historical risk procurement project;
[0038] Determine the person in charge of each procurement sub-project of the historical risk procurement project, obtain the list of responsible persons, compare the penalty list in the processing results with the list of responsible persons, and determine the problem sub-projects corresponding to the risk procurement project;
[0039] According to the problem sub-projects, risk mark the procurement data corresponding to each problem sub-project in the historical procurement data to determine the risk data corresponding to each procurement project;
[0040] Extract risk features from risk data to obtain feature data sets, and build training sets and test sets based on the feature data sets;
[0041] The deep learning neural network is trained and tested based on the training set and the test set to obtain the procurement risk prediction model.
[0042] Preferably, in a government procurement risk supervision and prediction system based on artificial intelligence learning, the risk warning module includes:
[0043] The warning person determination unit is used to obtain the list of project responsible persons corresponding to the risky procurement project, and based on the list of project responsible persons, generate a list of target persons for risk warning sending;
[0044] The warning sending record unit is used to synchronously send risk warnings according to the target person list and generate warning records for storage.
[0045] Preferably, in a government procurement risk supervision and prediction system based on artificial intelligence learning, the risk warning module also includes:
[0046] The level signal generating unit is used to obtain the risk value corresponding to the risk procurement project, determine the risk level corresponding to the risk procurement project in combination with the preset level classification, and generate a risk warning of the corresponding level according to the risk level.
[0047] Compared with the prior art, the present invention has at least the following beneficial effects:
[0048] The present invention establishes a procurement risk prediction model, performs risk prediction on effective procurement projects, obtains abnormal risk values of projects, determines whether effective procurement projects are risky procurement projects, and realizes automatic monitoring of risks of government procurement projects. At different stages of the progress of procurement projects, the risk prediction model is used to automatically predict the risks of various procurement projects, and the whole process risk monitoring of procurement projects is realized. Procurement dishonesty problems can be discovered in time, and losses caused by dishonest behaviors in the procurement process can be prevented and reduced as much as possible, effectively improving the prevention capabilities within the government, and when an effective procurement project is determined to be a risky procurement project, a risk level assessment is performed on the risky procurement project. According to the risk level assessment results, a risk warning is generated and sent to the target person. When risks are discovered, the relevant responsible persons of the risky procurement project are quickly reminded, so that corresponding measures can be taken to avoid further expansion of risks. At the same time, according to the risk level assessment results and the responsibilities and authorities of each link of the risky procurement project, the risk project processing results are generated, and targeted guidance and specifications are provided for the processing of risky procurement projects.
[0049] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures specifically pointed out in this application document.
[0050] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0051] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:
[0052] Figure 1 This is a flow chart of a government procurement risk supervision and prediction method based on artificial intelligence learning of the present invention;
[0053] Figure 2 This is a flowchart of step 1 of a government procurement risk supervision and prediction method based on artificial intelligence learning of the present invention;
[0054] Figure 3 This is a flowchart of step 2 of a government procurement risk supervision and prediction method based on artificial intelligence learning of the present invention;
[0055] Figure 4 This is a flowchart of step 3 of a government procurement risk supervision and prediction method based on artificial intelligence learning of the present invention;
[0056] Figure 5 This is a schematic diagram of a government procurement risk supervision and prediction system based on artificial intelligence learning according to the present invention. DETAILED DESCRIPTION
[0057] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.
[0058] Embodiment 1:
[0059] The present invention provides a government procurement risk supervision and prediction method based on artificial intelligence learning, such as Figure 1 As shown, including:
[0060] Step 1: Based on the procurement risk prediction model, predict the risk of valid procurement projects, obtain the abnormal risk value of the project, and determine whether the valid procurement project is a risky procurement project;
[0061] Step 2: When a valid procurement project is determined to be a risky procurement project, a risk level assessment is conducted on the risky procurement project. Based on the risk level assessment results, a risk warning is generated and sent to the target person;
[0062] Step 3: Generate risk project processing results based on the risk level assessment results and the responsibilities and authorities of each link of the risk procurement project.
[0063] In this embodiment, risky procurement projects refer to procurement projects involving dishonest behaviors such as non-standard procurement processes, mismatch between procurement amounts, and mismatch between the quality of purchased materials and the value of the materials.
[0064] In this embodiment, the abnormal project risk value refers to the probability that each procurement project has risks.
[0065] In this embodiment, the risk level includes multiple levels. The thresholds corresponding to different risk levels can be changed according to actual usage, and the number of risk levels and the standards corresponding to each risk level can be adjusted.
[0066] In this embodiment, the risk project processing result refers to determining the processing methods and control measures for different responsible persons (including the purchaser and the person in charge) in the risk project based on the risk assessment results of the risk project.
[0067] The beneficial effects of the above technical solution are as follows: the present invention establishes a procurement risk prediction model, performs risk prediction on effective procurement projects, obtains abnormal risk values of projects, determines whether effective procurement projects are risky procurement projects, and realizes automatic monitoring of risks of government procurement projects. At different stages of the progress of procurement projects, the risk prediction model is used to automatically predict the risks of various procurement projects, and the whole process of risk monitoring of procurement projects is realized. Procurement dishonesty problems can be discovered in time, and losses caused by dishonest behaviors in the procurement process can be prevented and reduced as much as possible, effectively improving the prevention capabilities within the government, and when determining that an effective procurement project is a risky procurement project, the risk level of the risky procurement project is evaluated, and according to the risk level evaluation results, a risk warning is generated and sent to the target person. When risks are discovered, the relevant responsible persons of the risky procurement project are quickly reminded, so as to facilitate the adoption of corresponding measures to avoid further expansion of risks. At the same time, according to the risk level evaluation results and the responsibilities and authorities of each link of the risky procurement project, the risk project processing results are generated to provide targeted guidance and specifications for the processing of risky procurement projects.
[0068] Embodiment 2:
[0069] On the basis of Example 1, step 1, such as Figure 2 As shown, including:
[0070] Step 101: Obtain a list of valid procurement items at a preset time interval;
[0071] Step 102: Based on the procurement risk prediction model, risk prediction is performed on multiple valid procurement projects in the valid procurement project list in a preset order to obtain abnormal risk values of the projects;
[0072] Step 103: Based on the abnormal risk value of the project, risk judgment is performed on multiple valid procurement projects respectively;
[0073] When the abnormal risk value of a project is greater than the preset value, the valid procurement project is determined to be a risky procurement project;
[0074] Otherwise, the valid procurement project is determined to be a non-risk procurement project.
[0075] In this embodiment, the valid procurement project list refers to a list generated by all procurement projects that are in progress when the preset monitoring time arrives or that have been completed within the current preset monitoring time and the previous preset monitoring time.
[0076] In this embodiment, the project abnormal risk value refers to the probability of the existence of irregular or dishonest behavior risks in the effective procurement projects predicted according to the procurement risk prediction model.
[0077] In this embodiment, the preset order refers to the risk prediction order of different types of procurement projects set in advance. For example, the risk prediction order of engineering projects is higher than that of daily necessities procurement projects.
[0078] The beneficial effects of the above technical solution are as follows: the present invention obtains a list of valid procurement projects at preset time intervals, realizes automatic screening of a large number of recorded government procurement projects, avoids repeated monitoring of completed projects, and realizes regular automatic monitoring of government procurement projects, effectively improving the prevention capabilities of risky procurement projects and minimizing risks.
[0079] Embodiment 3:
[0080] Based on Example 1, a government procurement risk supervision and prediction method based on artificial intelligence learning also includes:
[0081] Collect global historical procurement data corresponding to a large number of historical risk procurement projects, and obtain the processing results of each historical risk procurement project;
[0082] Determine the person in charge of each procurement sub-project of the historical risk procurement project, obtain the list of responsible persons, compare the penalty list in the processing results with the list of responsible persons, and determine the problem sub-projects corresponding to the risk procurement project;
[0083] According to the problem sub-projects, risk mark the procurement data corresponding to each problem sub-project in the historical procurement data to determine the risk data corresponding to each procurement project;
[0084] Extract risk features from risk data to obtain feature data sets, and build training sets and test sets based on the feature data sets;
[0085] The deep learning neural network is trained and tested based on the training set and the test set to obtain the procurement risk prediction model.
[0086] In this embodiment, historical risk procurement projects refer to historical risk projects whose processing results have been determined. The historical risk procurement projects include multiple risk procurement projects, including multiple procurement project types and multiple project stages where risks are discovered in the projects, ensuring the diversity of data used in model training and improving the generalization ability and risk prediction ability of the model.
[0087] In this embodiment, the procurement sub-project refers to a large number of procurement materials due to historical risk procurement projects. The procurement task corresponding to a project is divided into multiple small procurement projects, and different procurement sub-projects are executed at different stages of the project.
[0088] In this embodiment, the list of responsible persons refers to a list of persons in charge of each procurement sub-project in the historical risk procurement project and the purchaser.
[0089] In this embodiment, the problematic sub-project refers to a procurement sub-project in a historical risk procurement project that is determined to have risks.
[0090] In this embodiment, risk feature extraction is performed on risk data, including: feature extraction of procurement data, including bidding project name, procurement amount, procurement method, etc.;
[0091] And encode non-numeric features, for example, use one-hot encoding to convert categorical features into numerical features.
[0092] The beneficial effects of the above technical solution are as follows: the present invention obtains and collects historical procurement data corresponding to a large number of historical risk procurement projects, and obtains the processing results of each historical risk procurement project; determines the person in charge corresponding to each procurement sub-project of the historical risk procurement project, obtains a list of responsible persons, compares the penalty list in the processing results with the list of responsible persons, and determines the problem sub-projects corresponding to the risk procurement projects; according to the problem sub-projects, the procurement data corresponding to each problem sub-project in the historical procurement data is risk-marked, and the risk data corresponding to each procurement project is determined; risk features are extracted from the risk data to obtain a feature data set, and a training set and a test set are established based on the feature data set; the deep learning neural network is trained and tested based on the training set and the test set to obtain a procurement risk prediction model, and the construction of a risk prediction model for the entire process of procurement project risk based on an artificial intelligence neural network is completed.
[0093] Embodiment 4:
[0094] On the basis of Example 3, a large number of global historical procurement data corresponding to historical risk procurement projects are collected, and the processing results of each historical risk procurement project are obtained, including:
[0095] Obtain the historical procurement plans corresponding to each historical procurement project and the actual project progress corresponding to each historical procurement project;
[0096] Based on the historical procurement plan, the global historical procurement data corresponding to each historical procurement project is segmented to determine the phased historical procurement data corresponding to each procurement stage in the actual project process;
[0097] Determine whether each procurement stage of each historical procurement project has corresponding staged historical procurement data;
[0098] If it does not exist, clear the global historical procurement data corresponding to the historical procurement item;
[0099] If so, compare the phased historical procurement data of different historical procurement projects to obtain the difference comparison results, and obtain the data association relationship between the historical procurement link data corresponding to each procurement link in each procurement phase of the historical procurement project, and compare the data association relationship of different historical procurement projects to obtain the association comparison results;
[0100] Based on the difference comparison results and the correlation comparison results, all global historical procurement data are deduplicated to obtain the final global historical procurement data, and the processing results of the historical risk procurement projects corresponding to the final global historical procurement data are obtained.
[0101] In this embodiment, the historical procurement plan refers to a procurement plan corresponding to a historical procurement project. The procurement plan divides the historical procurement project into multiple procurement stages, and one procurement stage corresponds to one procurement sub-project.
[0102] In this embodiment, the actual project progress refers to the actual progress of the project when the historical procurement project is determined to be a risky project.
[0103] In this embodiment, the global historical procurement data refers to all historical procurement data corresponding to the historical procurement project; the phased historical procurement data refers to the historical procurement data corresponding to each procurement phase of the historical procurement project.
[0104] In this embodiment, the difference comparison result refers to the result obtained after comparing the total historical procurement data corresponding to each procurement stage of different historical procurement projects, and the result may be the same or different.
[0105] In this embodiment, the association comparison result refers to the comparison result of the data association relationship between the procurement data of different historical procurement projects, and the result may be the same or different.
[0106] In this embodiment, the historical procurement link data refers to the historical procurement data corresponding to different procurement links in each procurement stage.
[0107] In this embodiment, the data association relationship refers to the association relationship between procurement data between different procurement links in each procurement stage.
[0108] In this embodiment, the final global historical purchase data refers to the global historical purchase data from which duplicate data and incomplete data have been deleted.
[0109] The beneficial effects of the above technical solution are as follows: respectively obtaining the historical procurement plans corresponding to each historical procurement project and the actual project progress corresponding to each historical procurement project; based on the historical procurement plans, respectively performing data segmentation on the global historical procurement data corresponding to each historical procurement project, and determining the phased historical procurement data corresponding to each procurement stage in the actual project process; judging whether there is corresponding phased historical procurement data for each procurement stage of each historical procurement project; if not, clearing the global historical procurement data corresponding to the historical procurement project; if so, comparing the phased historical procurement data of different historical procurement projects to obtain difference comparison results, and respectively obtaining the data association relationship between the historical procurement link data corresponding to each procurement link in each procurement stage of the historical procurement project, comparing the data association relationship of different historical procurement projects to obtain association comparison results; according to the difference comparison results and the association comparison results, deduplication of all global historical procurement data to obtain the final global historical procurement data, and obtaining the processing results of the historical risk procurement projects corresponding to the final global historical procurement data, completing the cleaning of model training and test data, and ensuring the integrity and accuracy of model training and test data.
[0110] Embodiment 5:
[0111] Based on Example 1, step 2, such as Figure 3 As shown, including:
[0112] Step 211: Obtain the risk value corresponding to the risky procurement project, determine the risk level corresponding to the risky procurement project in combination with the preset level classification, and generate a risk warning of the corresponding level according to the risk level;
[0113] Step 212: Obtain a list of project responsible persons corresponding to the risky procurement project, and generate a list of target persons to whom risk warnings are to be sent based on the list of project responsible persons;
[0114] Step 202: synchronously send risk warnings according to the target person list, and generate warning records for storage.
[0115] In this embodiment, the list of project responsible persons refers to the list of persons in charge of the risk procurement project.
[0116] In this embodiment, the target person refers to the person who receives the risk warning.
[0117] The beneficial effects of the above technical solution are as follows: the present invention determines the risk level of risky procurement projects, and generates risk warnings of corresponding levels according to the risk level, so that the target task can quickly determine the severity of the risk of the procurement project according to the warning notification, and quickly remind the relevant responsible persons of the risky procurement project, which is conducive to the target person to quickly take risk control measures after receiving the risk warning; obtain the list of project responsible persons corresponding to the risky procurement project, and generate a list of target persons for sending risk warnings based on the list of project responsible persons; send risk warnings synchronously according to the list of target persons, which ensures the accuracy of risk warning notifications, facilitates the target person to take corresponding measures to avoid further expansion of risks, and generates warning records for storage, which is convenient for data viewing in the subsequent responsibility authentication process.
[0118] Embodiment 6:
[0119] Based on Example 1, step 3, as Figure 4 As shown, including:
[0120] Step 301: According to the risk level assessment result, obtain the first historical processing result corresponding to the same type of historical risk procurement projects of the corresponding risk level, and generate a processing result set;
[0121] Step 302: Obtain the risk marking result of the risk procurement project based on the risk prediction model of the procurement risk prediction model, and determine the first problem procurement stage corresponding to the risk procurement project according to the risk marking result;
[0122] Step 303: respectively comparing the first problem procurement stage of the risk procurement project with the second problem procurement stage of similar historical risk procurement projects to obtain risk similarity;
[0123] Step 304: based on the risk similarity, data is eliminated from the processing result set to obtain a result reference set;
[0124] Step 305: Obtain problem data corresponding to the first problem procurement stage, determine the risk links corresponding to the first problem procurement stage based on the problem data, and determine the risk occurrence time of each risk link according to the procurement sequence corresponding to the risk link;
[0125] Step 306: Based on the risk link, refer to the result reference set to obtain the prediction processing result corresponding to the risk procurement project;
[0126] Step 307: According to the risk occurrence time and in combination with the preset project management strategy, the prediction processing result is adjusted to obtain the risk project processing result and send it to the project management end.
[0127] In this embodiment, the same type of historical risk procurement projects refers to historical risk procurement projects with the same risk level as the current risk procurement project.
[0128] In this embodiment, the first historical processing result refers to the processing results corresponding to all similar historical risk procurement projects.
[0129] In this embodiment, the processing result set refers to a data set constructed based on the first processing result.
[0130] In this embodiment, the risk marking result refers to the marking result corresponding to the risky procurement data selected by the risky procurement project predicted by the procurement risk prediction model.
[0131] In this embodiment, the first problem procurement stage refers to the procurement stage where risks exist in risk procurement projects; the second problem procurement stage refers to the procurement stage where risks exist in similar historical risk procurement projects, and one procurement stage corresponds to one procurement sub-project.
[0132] In this embodiment, risk similarity refers to the similarity of the location where risks exist when comparing a risky procurement project with different similar historical risky procurement projects, where the location is limited to the procurement stage.
[0133] In this embodiment, the result reference set refers to a processing result set from which the first historical processing results corresponding to similar historical risk procurement projects whose risk similarity is less than a preset value are eliminated.
[0134] In this embodiment, the problem data refers to procurement data predicted to have risks in the procurement stage where the risky procurement project has risks (ie, the first problem procurement stage).
[0135] In this embodiment, the risk link refers to the procurement link with risks predicted in the first problem procurement stage.
[0136] In this embodiment, the predicted processing result refers to the processing result corresponding to the risk procurement project inferred based on the processing results of historical risk procurement projects with the same risk level as the risk procurement project.
[0137] In this embodiment, the risk occurrence time refers to the start time of execution of each risk link (i.e., the start time of procurement).
[0138] In this embodiment, the preset project management strategy refers to the government department's management rules and regulations for procurement and the penalty system.
[0139] In this embodiment, the risk project processing result refers to the risk project penalty result that is ultimately sent to the project management end for reference.
[0140] The beneficial effects of the above technical solution are as follows: the present invention obtains the first historical processing result corresponding to the same type of historical risk procurement projects of the corresponding risk level according to the risk level assessment result, and generates a processing result set; obtains the risk marking result of the risk procurement project based on the risk prediction of the procurement risk prediction model, and determines the first problem procurement stage corresponding to the risk procurement project according to the risk marking result; respectively compares the first problem procurement stage of the risk procurement project with the second problem procurement stage of the same type of historical risk procurement project to obtain the risk similarity; based on the risk similarity, performs data elimination on the processing result set to obtain a result reference set; obtains the problem data corresponding to the first problem procurement stage, determines the risk link corresponding to the first problem procurement stage based on the problem data, and determines the risk occurrence time of each risk link according to the procurement sequence corresponding to the risk link; based on the risk link, refers to the result reference set to obtain the predicted processing result corresponding to the risk procurement project; according to the risk occurrence time, in combination with the preset project management strategy, adjusts the predicted processing result, obtains the risk project processing result, and sends it to the project management end, thereby realizing the automatic generation of penalties for risk procurement projects, providing a reference for the penalties for risk procurement projects, and providing targeted guidance and specifications for the determination of penalty results.
[0141] Embodiment 7:
[0142] On the basis of Example 6, based on the risk link, refer to the result reference set to obtain the prediction processing results corresponding to the risk project, including:
[0143] The second historical processing result in the result reference set is divided according to the second problem procurement stage of the third historical risk procurement project corresponding to the second historical processing result, to obtain historical processing sub-results corresponding to different second problem procurement stages;
[0144] Compare the historical processing sub-results of the same second problem procurement stage of different second historical processing results to obtain the stage processing result characteristics;
[0145] Determine the historical risk links corresponding to each second-issue procurement stage of different similar historical risk procurement projects, and determine the similarity of data characteristics between each historical risk link and the risk link of the risk project;
[0146] Based on the similarity of data features, determine the reference processing results corresponding to each risk link of the risk procurement project;
[0147] Based on the procurement stage where the risk link corresponding to each reference processing result is located, the corresponding stage processing result characteristics are obtained, and the reference results are standardized according to the stage processing result characteristics to obtain the processing results corresponding to each risk link;
[0148] Obtain a list of purchasers corresponding to each risk link of the risky procurement project, combine it with the list of project responsible persons, generate a penalty list, add the processing results to the penalty list according to the corresponding relationship between the risk links, and obtain the predicted processing results corresponding to the risky procurement project.
[0149] In this embodiment, the second historical processing result refers to the processing result included in the result reference set.
[0150] In this embodiment, the third historical risk procurement project refers to the historical risk procurement project corresponding to the second historical processing result.
[0151] In this embodiment, the historical processing sub-result refers to the processing results of the responsible persons corresponding to different procurement stages with risks in the second historical processing results.
[0152] In this embodiment, the stage processing result characteristics refer to the penalty characteristics of different procurement stages with risks.
[0153] In this embodiment, the historical risk link refers to the procurement link with risks in the second problem procurement stage.
[0154] In this embodiment, determining the reference processing result based on the data feature similarity means taking the processing result corresponding to the historical risk link with the highest data feature similarity to the risk link of the risk project as the reference processing result of the risk link.
[0155] Beneficial effects of the above technical solution: the present invention divides the second historical processing result in the result reference set according to the second problem procurement stage of the third historical risk procurement project corresponding to the second historical processing result, and obtains historical processing sub-results corresponding to different second problem procurement stages; compares the historical processing sub-results of the same second problem procurement stage of different second historical processing results to obtain stage processing result characteristics; determines the historical risk links corresponding to each second problem procurement stage of different similar historical risk procurement projects, and determines the data feature similarity between each historical risk link and the risk link of the risk project; determines the reference processing results corresponding to each risk link of the risk procurement project based on the data feature similarity; obtains the corresponding stage processing result characteristics based on the procurement stage where the risk link corresponding to each reference processing result is located, and standardizes the reference results according to the stage processing result characteristics to obtain the processing results corresponding to each risk link; obtains the list of purchasers corresponding to each risk link of the risk procurement project, generates a penalty list in combination with the list of project responsible persons, and adds the processing results to the penalty list according to the corresponding relationship between the risk links to obtain the predicted processing results corresponding to the risk procurement project, and realizes the prediction of risk procurement project processing through the procurement link where the risk occurs and the historical processing results, which provides a basis for the generation of the final risk project processing results.
[0156] Embodiment 8:
[0157] Based on Example 7, the reference results are standardized according to the characteristics of the stage processing results, including:
[0158] Based on the characteristics of the stage processing results, determine the corresponding responsibilities and authorities of the responsible persons at different levels in each procurement stage, as well as their corresponding responsibility levels, and generate the stage processing reference standards;
[0159] Obtain the data difference between the risk data corresponding to the reference result and the risk data corresponding to the risk link. Based on the data difference, determine the target responsibilities corresponding to different responsible persons in each risk link. According to the target responsibilities and combined with the stage processing reference standards, adjust the reference results to obtain standardized reference results.
[0160] In this embodiment, the degree of responsibility refers to the proportion of responsibility shared by persons of different levels, responsibilities and authorities in the entire procurement sub-project at each procurement stage.
[0161] The beneficial effects of the above technical solution: The present invention determines the responsibilities and authorities corresponding to the responsible persons of different levels in each procurement stage, as well as their corresponding degree of responsibility, based on the characteristics of the stage processing results, generates stage processing reference standards, and provides responsibility sharing standards for the penalties of the responsible persons in different procurement stages of risk projects. Then, the data difference between the risk data corresponding to the reference results and the risk data corresponding to the risk links is obtained. Based on the data difference, the target responsibilities corresponding to different responsible persons in each risk link are determined, and according to the target responsibilities and combined with the stage processing reference standards, the reference results are adjusted to obtain standardized reference results. According to the stage processing reference standards corresponding to each procurement stage, the processing results of the responsible persons of each risk link in the stage are adjusted to ensure the rationality of the reference results, improve the accuracy of the reference results, and help to increase the probability of the reference results being adopted.
[0162] Embodiment 9:
[0163] On the basis of Example 6, according to the time of risk occurrence and in combination with the preset project management strategy, the prediction processing result is adjusted to obtain the risk project processing result, including:
[0164] According to the risk transmission time and the forecast processing results, determine the main and secondary responsible persons for the risk procurement project;
[0165] According to the project importance corresponding to the risk link corresponding to the sub-responsible person and the amount corresponding to the risk data, combined with the preset project management strategy, the sub-responsible person is graded to obtain the responsibility classification result;
[0166] The main responsible person is regarded as the highest responsibility level, and the final responsibility result is generated based on the responsibility level results;
[0167] Compare the final responsibility result with the responsibility certification result in the predicted processing result to obtain the responsibility certification difference, determine the person responsible for the difference based on the responsibility certification difference, and obtain the same type of processing result in the predicted processing result based on the final responsibility result corresponding to the person responsible for the difference;
[0168] Adjust the predicted processing results of the person responsible for the difference according to the similar processing results to obtain the final processing result, and take the predicted processing results of the person not responsible for the difference as the final processing result;
[0169] Based on the final processing results, the risk project processing results are generated.
[0170] In this embodiment, the main responsible person refers to the person who bears the main responsibility for the risks of the risk procurement project; the secondary responsible person refers to the person who bears other responsibility for the risks of the risk procurement project other than the main responsible person.
[0171] In this embodiment, the final responsibility result refers to the responsibility classification result of all responsible persons.
[0172] In this embodiment, the responsibility authentication difference refers to the difference between the responsibility classification determination results of different responsible persons in the prediction processing results and the responsibility classification determination results of different responsible persons in the final responsibility results.
[0173] In this embodiment, the person responsible for the difference refers to the person responsible for the difference in the responsibility authentication result.
[0174] In this embodiment, the same type of processing results refers to the processing results corresponding to the same responsibility level as the predicted processing results and the final responsibility results corresponding to the difference responsible persons.
[0175] The beneficial effects of the above technical solution are as follows: the present invention determines the main responsible person and the secondary responsible person of the risk procurement project according to the risk sending time and the prediction processing result, thereby realizing the primary classification of the responsibility of the responsible person for the risk procurement project; and according to the project importance corresponding to the risk link corresponding to the secondary responsible person and the amount corresponding to the risk data, combined with the preset project management strategy, the secondary responsible person is classified to obtain the responsibility classification result, thereby realizing the secondary classification of the responsibility of the secondary responsible person; the main responsible person is taken as the highest responsibility classification, and the final responsibility result is generated in combination with the responsibility classification result, so that the responsibility division of the risk procurement project is more refined, which is conducive to improving the refinement of the risk procurement project processing result; the final responsibility result is compared with the responsibility certification result in the prediction processing result to obtain the responsibility certification difference, and the difference responsible person is determined based on the responsibility certification difference, and the same type of processing result is obtained in the prediction processing result based on the final responsibility result corresponding to the difference responsible person; the prediction processing result of the difference responsible person is adjusted according to the same type of processing result to obtain the final processing result, and the prediction processing result of the non-difference responsible person is taken as the final processing result; based on the final processing result, the risk project processing result is generated to obtain, and the accuracy of the reference result is improved again, which is conducive to improving the probability of the reference result being adopted.
[0176] In embodiment 10:
[0177] The present invention provides a government procurement risk supervision and prediction system based on artificial intelligence learning, comprising:
[0178] The risk judgment module is used to predict the risks of effective procurement projects based on the procurement risk prediction model, obtain the abnormal risk value of the project, and judge whether the effective procurement project is a risky procurement project;
[0179] The risk warning module is used to evaluate the risk level of a valid procurement project when it is determined that the project is a risky procurement project. According to the risk level evaluation results, a risk warning is generated and sent to the target person;
[0180] The processing generation module is used to generate risk project processing results based on the risk level assessment results and the responsibilities and authorities of each link of the risk procurement project.
[0181] The beneficial effects of the above technical solution are as follows: the present invention establishes a procurement risk prediction model, performs risk prediction on effective procurement projects, obtains abnormal risk values of projects, determines whether effective procurement projects are risky procurement projects, and realizes automatic monitoring of risks of government procurement projects. At different stages of the progress of procurement projects, the risk prediction model is used to automatically predict the risks of various procurement projects, and the whole process of risk monitoring of procurement projects is realized. Procurement dishonesty problems can be discovered in time, and losses caused by dishonest behaviors in the procurement process can be prevented and reduced as much as possible, effectively improving the prevention capabilities within the government, and when determining that an effective procurement project is a risky procurement project, the risk level of the risky procurement project is evaluated, and according to the risk level evaluation results, a risk warning is generated and sent to the target person. When risks are discovered, the relevant responsible persons of the risky procurement project are quickly reminded, so as to facilitate the adoption of corresponding measures to avoid further expansion of risks. At the same time, according to the risk level evaluation results and the responsibilities and authorities of each link of the risky procurement project, the risk project processing results are generated to provide targeted guidance and specifications for the processing of risky procurement projects.
[0182] Embodiment 11:
[0183] Based on Example 10, the risk judgment module includes:
[0184] An autonomous acquisition unit, used to obtain a list of valid procurement items at preset time intervals;
[0185] An intelligent prediction unit, used to perform risk prediction on multiple valid procurement projects in the valid procurement project list in a preset order based on a procurement risk prediction model, and obtain abnormal risk values of the projects;
[0186] Intelligent judgment unit, used to make risk judgments on multiple valid procurement projects based on the abnormal risk value of the project;
[0187] When the abnormal risk value of a project is greater than the preset value, the valid procurement project is determined to be a risky procurement project;
[0188] Otherwise, the valid procurement project is determined to be a non-risk procurement project.
[0189] The beneficial effects of the above technical solution are as follows: the present invention obtains a list of valid procurement projects at preset time intervals, realizes automatic screening of a large number of recorded government procurement projects, avoids repeated monitoring of completed projects, and realizes regular automatic monitoring of government procurement projects, effectively improving the prevention capabilities of risky procurement projects and minimizing risks.
[0190] Embodiment 12:
[0191] Based on Example 10, a government procurement risk supervision and prediction system based on artificial intelligence learning also includes:
[0192] The forecast preparation model is used to collect global historical procurement data corresponding to a large number of historical risk procurement projects and obtain the processing results of each historical risk procurement project;
[0193] Determine the person in charge of each procurement sub-project of the historical risk procurement project, obtain the list of responsible persons, compare the penalty list in the processing results with the list of responsible persons, and determine the problem sub-projects corresponding to the risk procurement project;
[0194] According to the problem sub-projects, risk mark the procurement data corresponding to each problem sub-project in the historical procurement data to determine the risk data corresponding to each procurement project;
[0195] Extract risk features from risk data to obtain feature data sets, and build training sets and test sets based on the feature data sets;
[0196] The deep learning neural network is trained and tested based on the training set and the test set to obtain the procurement risk prediction model.
[0197] The beneficial effects of the above technical solution are as follows: the present invention obtains and collects historical procurement data corresponding to a large number of historical risk procurement projects, and obtains the processing results of each historical risk procurement project; determines the person in charge corresponding to each procurement sub-project of the historical risk procurement project, obtains a list of responsible persons, compares the penalty list in the processing results with the list of responsible persons, and determines the problem sub-projects corresponding to the risk procurement projects; according to the problem sub-projects, the procurement data corresponding to each problem sub-project in the historical procurement data is risk-marked, and the risk data corresponding to each procurement project is determined; risk features are extracted from the risk data to obtain a feature data set, and a training set and a test set are established based on the feature data set; the deep learning neural network is trained and tested based on the training set and the test set to obtain a procurement risk prediction model, and the construction of a risk prediction model for the entire process of procurement project risk based on an artificial intelligence neural network is completed.
[0198] Embodiment 13:
[0199] Based on Example 10, the risk warning module includes:
[0200] A level signal generating unit is used to obtain the risk value corresponding to the risk procurement project, determine the risk level corresponding to the risk procurement project in combination with the preset level classification, and generate a risk warning of the corresponding level according to the risk level;
[0201] The warning person determination unit is used to obtain the list of project responsible persons corresponding to the risky procurement project, and based on the list of project responsible persons, generate a list of target persons for risk warning sending;
[0202] The warning sending record unit is used to synchronously send risk warnings according to the target person list and generate warning records for storage.
[0203] The beneficial effects of the above technical solution are as follows: the present invention determines the risk level of risky procurement projects, and generates risk warnings of corresponding levels according to the risk level, so that the target task can quickly determine the severity of the risk of the procurement project according to the warning notification, and quickly remind the relevant responsible persons of the risky procurement project, which is conducive to the target person to quickly take risk control measures after receiving the risk warning; obtain the list of project responsible persons corresponding to the risky procurement project, and generate a list of target persons for sending risk warnings based on the list of project responsible persons; send risk warnings synchronously according to the list of target persons, which ensures the accuracy of risk warning notifications, facilitates the target person to take corresponding measures to avoid further expansion of risks, and generates warning records for storage, which is convenient for data viewing in the subsequent responsibility authentication process.
[0204] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
Claims
1. A government procurement risk supervision and prediction method based on artificial intelligence learning, characterized in that: include: Step 1: Based on the procurement risk prediction model, predict the risk of valid procurement projects, obtain the abnormal risk value of the project, and determine whether the valid procurement project is a risky procurement project; Step 2: When a valid procurement project is determined to be a risky procurement project, a risk level assessment is conducted on the risky procurement project. Based on the risk level assessment results, a risk warning is generated and sent to the target person; Step 3: Generate risk project processing results based on risk level assessment results and the responsibilities and authorities of each link of risk procurement projects; Wherein, step 3 includes: Step 301: According to the risk level assessment result, obtain the first historical processing result corresponding to the same type of historical risk procurement projects of the corresponding risk level, and generate a processing result set; Step 302: Obtain the risk marking result of the risk procurement project based on the risk prediction model of the procurement risk prediction model, and determine the first problem procurement stage corresponding to the risk procurement project according to the risk marking result; Step 303: respectively comparing the first problem procurement stage of the risk procurement project with the second problem procurement stage of similar historical risk procurement projects to obtain risk similarity; Step 304: based on the risk similarity, data is eliminated from the processing result set to obtain a result reference set; Step 305: Obtain problem data corresponding to the first problem procurement stage, determine the risk links corresponding to the first problem procurement stage based on the problem data, and determine the risk occurrence time of each risk link according to the procurement sequence corresponding to the risk link; Step 306: Based on the risk link, refer to the result reference set to obtain the prediction processing result corresponding to the risk procurement project; Step 307: According to the risk occurrence time and in combination with the preset project management strategy, the prediction processing result is adjusted to obtain the risk project processing result and send it to the project management end; Among them, based on the risk link, refer to the result reference set to obtain the prediction processing results corresponding to the risk project, including: The second historical processing result in the result reference set is divided according to the second problem procurement stage of the third historical risk procurement project corresponding to the second historical processing result, to obtain historical processing sub-results corresponding to different second problem procurement stages; Compare the historical processing sub-results of the same second problem procurement stage of different second historical processing results to obtain the stage processing result characteristics; Determine the historical risk links corresponding to each second-issue procurement stage of different similar historical risk procurement projects, and determine the similarity of data characteristics between each historical risk link and the risk link of the risk project; Based on the similarity of data features, determine the reference processing results corresponding to each risk link of the risk procurement project; Based on the procurement stage where the risk link corresponding to each reference processing result is located, the corresponding stage processing result characteristics are obtained, and the reference results are standardized according to the stage processing result characteristics to obtain the processing results corresponding to each risk link; Obtain a list of purchasers corresponding to each risk link of the risky procurement project, combine it with the list of project responsible persons, generate a penalty list, add the processing results to the penalty list according to the corresponding relationship between the risk links, and obtain the predicted processing results corresponding to the risky procurement project.
2. According to claim 1, a government procurement risk supervision and prediction method based on artificial intelligence learning is characterized in that: Step 1 includes: Obtain a list of valid procurement items at preset time intervals; Based on the procurement risk prediction model, risk prediction is performed on multiple valid procurement projects in the valid procurement project list in a preset order to obtain abnormal risk values of the projects; Based on the abnormal risk value of the project, risk assessment is performed on multiple valid procurement projects; When the abnormal risk value of a project is greater than the preset value, the valid procurement project is determined to be a risky procurement project; Otherwise, the valid procurement project is determined to be a non-risk procurement project.
3. According to claim 1, a government procurement risk supervision and prediction method based on artificial intelligence learning is characterized in that: Also includes: Collect global historical procurement data corresponding to a large number of historical risk procurement projects, and obtain the processing results of each historical risk procurement project; Determine the person in charge of each procurement sub-project of the historical risk procurement project, obtain the list of responsible persons, compare the penalty list in the processing results with the list of responsible persons, and determine the problem sub-projects corresponding to the risk procurement project; According to the problem sub-projects, risk mark the procurement data corresponding to each problem sub-project in the historical procurement data to determine the risk data corresponding to each procurement project; Extract risk features from risk data to obtain feature data sets, and build training sets and test sets based on the feature data sets; The deep learning neural network is trained and tested based on the training set and the test set to obtain the procurement risk prediction model.
4. According to claim 1, a government procurement risk supervision and prediction method based on artificial intelligence learning is characterized in that: Step 2 includes: Obtain the list of project responsible persons corresponding to the risky procurement project, and generate a list of target persons for risk warning based on the list of project responsible persons; Risk warnings are sent synchronously based on the target person list, and warning records are generated for storage.
5. According to claim 1, a government procurement risk supervision and prediction method based on artificial intelligence learning is characterized in that: When obtaining the list of project responsible persons corresponding to the risky procurement project and generating a list of target persons for risk warning based on the list of project responsible persons, it also includes: Obtain the risk value corresponding to the risky procurement project, determine the risk level corresponding to the risky procurement project in combination with the preset level classification, and generate a risk warning of the corresponding level based on the risk level.
6. A government procurement risk supervision and prediction system based on artificial intelligence learning, characterized in that: include: The risk judgment module is used to predict the risks of effective procurement projects based on the procurement risk prediction model, obtain the abnormal risk value of the project, and judge whether the effective procurement project is a risky procurement project; The risk warning module is used to evaluate the risk level of a valid procurement project when it is determined that the project is a risky procurement project. According to the risk level evaluation results, a risk warning is generated and sent to the target person; The processing generation module is used to generate risk project processing results based on the risk level assessment results and the responsibilities and authorities of each link of the risk procurement project; Among them, the processing generation module generates the method of risk project processing results according to the risk level assessment results and the responsibilities and authorities of each link of the risk procurement project, including: According to the risk level assessment result, the first historical processing result corresponding to the same type of historical risk procurement projects of the corresponding risk level is obtained, and a processing result set is generated; Obtain the risk marking results of the risk procurement project based on the risk prediction model, and determine the first problem procurement stage corresponding to the risk procurement project according to the risk marking results; Compare the first-question procurement stage of the risk procurement project with the second-question procurement stage of similar historical risk procurement projects to obtain risk similarity; Based on the risk similarity, data is eliminated from the processing result set to obtain a result reference set; Obtain the problem data corresponding to the first problem procurement stage, determine the risk links corresponding to the first problem procurement stage based on the problem data, and determine the risk occurrence time of each risk link according to the procurement sequence corresponding to the risk link; Based on the risk link, refer to the result reference set to obtain the forecast processing results corresponding to the risk procurement project; According to the time of risk occurrence and in combination with the preset project management strategy, the forecast processing results are adjusted to obtain the risk project processing results and send them to the project management end; Among them, based on the risk link, refer to the result reference set to obtain the prediction processing results corresponding to the risk project, including: The second historical processing result in the result reference set is divided according to the second problem procurement stage of the third historical risk procurement project corresponding to the second historical processing result, to obtain historical processing sub-results corresponding to different second problem procurement stages; Compare the historical processing sub-results of the same second problem procurement stage of different second historical processing results to obtain the stage processing result characteristics; Determine the historical risk links corresponding to each second-issue procurement stage of different similar historical risk procurement projects, and determine the similarity of data characteristics between each historical risk link and the risk link of the risk project; Based on the similarity of data features, determine the reference processing results corresponding to each risk link of the risk procurement project; Based on the procurement stage where the risk link corresponding to each reference processing result is located, the corresponding stage processing result characteristics are obtained, and the reference results are standardized according to the stage processing result characteristics to obtain the processing results corresponding to each risk link; Obtain a list of purchasers corresponding to each risk link of the risky procurement project, combine it with the list of project responsible persons, generate a penalty list, add the processing results to the penalty list according to the corresponding relationship between the risk links, and obtain the predicted processing results corresponding to the risky procurement project.
7. According to claim 6, a government procurement risk supervision and prediction system based on artificial intelligence learning is characterized in that: Risk judgment module, including: An autonomous acquisition unit, used to obtain a list of valid procurement items at preset time intervals; An intelligent prediction unit, for performing risk prediction on multiple valid procurement projects in the valid procurement project list in a preset order based on a procurement risk prediction model, and obtaining an abnormal risk value of the project; Intelligent judgment unit, used to make risk judgments on multiple valid procurement projects based on the abnormal risk value of the project; When the abnormal risk value of a project is greater than the preset value, the valid procurement project is determined to be a risky procurement project; Otherwise, the valid procurement project is determined to be a non-risk procurement project.
8. According to claim 6, a government procurement risk supervision and prediction system based on artificial intelligence learning is characterized in that: Also includes: The forecast preparation model is used to collect global historical procurement data corresponding to a large number of historical risk procurement projects and obtain the processing results of each historical risk procurement project; Determine the person in charge of each procurement sub-project of the historical risk procurement project, obtain the list of responsible persons, compare the penalty list in the processing results with the list of responsible persons, and determine the problem sub-projects corresponding to the risk procurement project; According to the problem sub-projects, risk mark the procurement data corresponding to each problem sub-project in the historical procurement data to determine the risk data corresponding to each procurement project; Extract risk features from risk data to obtain feature data sets, and build training sets and test sets based on the feature data sets; The deep learning neural network is trained and tested based on the training set and the test set to obtain the procurement risk prediction model.
9. According to claim 6, a government procurement risk supervision and prediction system based on artificial intelligence learning is characterized in that: Risk early warning module, including: The warning person determination unit is used to obtain the list of project responsible persons corresponding to the risky procurement project, and based on the list of project responsible persons, generate a list of target persons for risk warning sending; The warning sending record unit is used to synchronously send risk warnings according to the target person list and generate warning records for storage.
10. The government procurement risk supervision and prediction system based on artificial intelligence learning according to claim 9 is characterized in that: The risk warning module also includes: The level signal generating unit is used to obtain the risk value corresponding to the risk procurement project, determine the risk level corresponding to the risk procurement project in combination with the preset level classification, and generate a risk warning of the corresponding level according to the risk level.
