Artificial intelligence-based behavior prediction method and apparatus, and related device

Abstract

The application relates to the technical field of artificial intelligence, and provides a behavior prediction method and device based on artificial intelligence and related equipment, the method comprises the following steps: obtaining an original data set from a data source of a plurality of target parameters, and performing first preprocessing on the original data set to obtain a target data table; calling a rule type label, obtaining a first prediction result from the target data table based on the rule type label through a label management system; extracting a feature factor set from the target data table, and inputting the feature factor set into a pre-trained classification prediction model to obtain a second prediction result; and performing second preprocessing on the first prediction result and the second prediction result to obtain a target prediction result. According to the rule engine and the classification prediction model, the target prediction result is obtained, and the accuracy of a procurement fraud prediction result is improved.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and specifically to a behavior prediction method, apparatus, and related equipment based on artificial intelligence. Background Technology

[0002] With the diversification of products, the procurement process, as a key operational link in the business system, is prone to fraud. Existing technologies use expert experience to analyze and process procurement big data to identify abnormal procurement processes that do not comply with external bidding regulations, internal management requirements, or common sense, and to screen procurement projects with potential fraud risks.

[0003] However, using expert experience to analyze and process procurement big data cannot identify whether the procurement big data has been circumvented by procurement fraudsters, resulting in low accuracy of procurement big data and consequently low accuracy of procurement fraud prediction.

[0004] Therefore, it is necessary to propose a method that can accurately predict procurement fraud. Summary of the Invention

[0005] In view of the above, it is necessary to propose a behavior prediction method, device and related equipment based on artificial intelligence, which obtains target prediction results based on rule engine and classification prediction model, thereby improving the accuracy of procurement fraud prediction results.

[0006] A first aspect of the present invention provides an artificial intelligence-based behavior prediction method, the method comprising:

[0007] Parse the received procurement fraud prediction request to obtain the data source for multiple target parameters;

[0008] The original dataset is obtained from the data source of the multiple target parameters, and the original dataset is preprocessed to obtain the target data table.

[0009] Invoke the rule-based tags, and based on the rule-based tags, obtain the first prediction result from the target data table through the tag management system;

[0010] The feature factor set is extracted from the target data table and input into the pre-trained classification prediction model to obtain the second prediction result;

[0011] The first prediction result and the second prediction result are subjected to a second preprocessing to obtain the target prediction result.

[0012] Optionally, the first preprocessing of the original dataset to obtain the target data table includes:

[0013] Identify the procurement business system identifier code corresponding to each of the original data tables in the original dataset;

[0014] The original dataset is categorized according to the identification code of the procurement business system;

[0015] The original datasets from the same procurement business system are designated as the first dataset, and the original datasets from the remaining different procurement business systems are designated as the second dataset.

[0016] The original data tables in the first dataset are merged to obtain a first data table, and the original data tables in the second dataset are merged to obtain a second data table;

[0017] Merge the first data table and the second data table to obtain the third data table;

[0018] The third data table is cleaned to obtain the target data table.

[0019] Optionally, the step of merging the original data tables in the second dataset to obtain the second data table includes:

[0020] Identify the field names of each original data table in the second dataset;

[0021] Merge multiple columns with the same field name from multiple original data tables, and delete the columns involved in the merger to obtain a new table;

[0022] Calculate the total number of field names in the new table, and obtain the preset total number of empty fields based on the total number of field names in the new table;

[0023] The new table is expanded based on the preset total number of empty fields to obtain a second data table.

[0024] Optionally, the step of calling the rule class tag and obtaining the first prediction result based on the rule class tag through a preset rule engine includes:

[0025] The data in the target data table is converted into rule class tags that can be recognized by a preset rule engine;

[0026] The preset rule engine is invoked to perform intelligent matching of rule class tags and obtain the first prediction result.

[0027] Optionally, extracting the feature factor set from the target data table includes:

[0028] Obtain the project information and preset processing script of the procurement project corresponding to the procurement project identifier code in the procurement fraud prediction request;

[0029] The project information is preprocessed to obtain the target project information;

[0030] The target project information is processed using the preset processing script to obtain a set of feature factors.

[0031] Optionally, the second preprocessing of the first prediction result and the second prediction result to obtain the target prediction result includes:

[0032] Compare the first prediction result with the second prediction result;

[0033] Delete the first prediction result that is the same as the second prediction result to obtain the target prediction result.

[0034] Optionally, the data source for parsing the received procurement fraud prediction request to obtain multiple target parameters includes:

[0035] Parse the message of the procurement fraud prediction request to obtain the message information carried by the message;

[0036] Obtain the procurement fraud prediction requirements from the aforementioned message information;

[0037] The procurement project identifier code in the procurement fraud prediction requirement is matched with the system identifier code in the preset database to obtain the first interface of the procurement business system corresponding to the procurement project and the second interface of the third-party system that interacts with the procurement business system.

[0038] Call the first interface and the second interface to obtain the data source of the corresponding target parameters.

[0039] A second aspect of the present invention provides an artificial intelligence-based behavior prediction device, the device comprising:

[0040] The parsing module is used to parse the received procurement fraud prediction request and obtain the data source of multiple target parameters;

[0041] The first preprocessing module is used to obtain the original dataset from the data source of the multiple target parameters, and perform a first preprocessing on the original dataset to obtain the target data table;

[0042] The construction module is used to call rule-based tags and, based on the rule-based tags, obtain a first prediction result from the target data table through the tag management system;

[0043] An extraction and input module is used to extract a set of feature factors from the target data table and input the set of feature factors into a pre-trained classification prediction model to obtain a second prediction result;

[0044] The second preprocessing module is used to perform a second preprocessing on the first prediction result and the second prediction result to obtain the target prediction result.

[0045] A third aspect of the present invention provides an electronic device comprising a processor and a memory, wherein the processor is configured to implement the aforementioned artificial intelligence-based behavior prediction method when executing a computer program stored in the memory.

[0046] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned artificial intelligence-based behavior prediction method.

[0047] In summary, the AI-based behavior prediction method, apparatus, and related equipment of this invention perform a first preprocessing on the original dataset, invoke rule-based labels, obtain a first prediction result from the target data table through a label management system based on the rule-based labels, and input the feature factor set into a pre-trained classification prediction model to obtain a second prediction result. The first prediction result is obtained based on the institutional regulations of the procurement project and is based on a rule engine, while the second prediction result is obtained based on the process of the procurement project and is based on a classification prediction model. By obtaining the target prediction result based on the rule engine and the classification prediction model, the accuracy of the procurement fraud prediction result is improved. Attached Figure Description

[0048] Figure 1 This is a flowchart of the behavior prediction method based on artificial intelligence provided in Embodiment 1 of the present invention.

[0049] Figure 2 This is a structural diagram of the behavior prediction device based on artificial intelligence provided in Embodiment 2 of the present invention.

[0050] Figure 3 This is a schematic diagram of the structure of the electronic device provided in Embodiment 3 of the present invention. Detailed Implementation

[0051] To better understand the above-mentioned objects, features, and advantages of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments of the present invention and the features thereof can be combined with each other.

[0052] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0053] Example 1

[0054] Figure 1 This is a flowchart of the behavior prediction method based on artificial intelligence provided in Embodiment 1 of the present invention.

[0055] In this embodiment, the AI-based behavior prediction method can be applied to electronic devices. For electronic devices that require AI-based behavior prediction, the AI-based behavior prediction function provided by the method of this invention can be directly integrated into the electronic device, or it can run in the electronic device in the form of a software development kit (SDK).

[0056] The embodiments of this invention can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0057] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, as well as machine learning and deep learning.

[0058] like Figure 1 As shown, the behavior prediction method based on artificial intelligence specifically includes the following steps. Depending on different needs, the order of the steps in this flowchart can be changed, and some can be omitted.

[0059] S11, parse the received procurement fraud prediction request to obtain the data source of multiple target parameters, wherein the procurement fraud prediction request includes a procurement project identification code.

[0060] In this embodiment, when an enterprise or user performs procurement fraud prediction, it initiates a procurement fraud prediction request to the server through a client. Specifically, the client can be a smartphone, iPad, or other existing smart device, and the server can be a procurement fraud prediction subsystem. During the procurement fraud prediction process, the client can send a procurement fraud prediction request to the procurement fraud prediction subsystem, which is used to receive the procurement fraud prediction request sent by the client.

[0061] In this embodiment, the data source for the target parameter can be at least one third-party system that interacts with the procurement business system, or at least one subsystem of the procurement business system.

[0062] In an optional embodiment, the data source for parsing the received procurement fraud prediction request to obtain multiple target parameters includes:

[0063] Parse the message of the procurement fraud prediction request to obtain the message information carried by the message;

[0064] Obtain the procurement fraud prediction requirements from the aforementioned message information;

[0065] Data sources for multiple target parameters are obtained based on the procurement fraud prediction requirements.

[0066] Specifically, the procurement fraud prediction requirement includes a procurement project identification code, which is used to uniquely identify the identity of the procurement business.

[0067] Furthermore, the data source for obtaining multiple target parameters based on the procurement fraud prediction requirement includes:

[0068] The procurement project identifier code in the procurement fraud prediction requirement is matched with the system identifier code in the preset database to obtain the first interface of the procurement business system corresponding to the procurement project and the second interface of the third-party system that interacts with the procurement business system.

[0069] Call the first interface and the second interface to obtain the data source of the corresponding target parameters.

[0070] In this embodiment, since the associated data involved in the procurement business system may come from the procurement business system, third-party systems that interact with the procurement business system, and subsystems of the procurement business system, the data source of the corresponding target parameters is obtained by calling the interface of the corresponding procurement business system. It is not necessary to obtain the data source of all target parameters of all procurement business systems, which is more targeted and improves the accuracy of the data source of the target parameters.

[0071] S12, obtain the original dataset from the data source of the multiple target parameters, and perform a first preprocessing on the original dataset to obtain the target data table.

[0072] In this embodiment, the original dataset refers to the original data obtained directly from the data source of the target parameter, and the target data table refers to the data obtained by processing each original data table in the original dataset.

[0073] In an optional embodiment, obtaining the raw dataset from the data source of the plurality of target parameters includes:

[0074] The procurement fraud prediction request is analyzed to obtain the data source collection requirements and corresponding collection interfaces for each target parameter.

[0075] Call the data collection interface of each target parameter, and use web crawling technology to obtain the raw dataset from the data source of each target parameter according to the data collection requirements of each target parameter's data source.

[0076] In this embodiment, the procurement fraud prediction requirement also includes the data source collection requirements and corresponding collection interfaces for each target parameter. The web crawler technology is existing technology and will not be described in detail here.

[0077] In an optional embodiment, the first preprocessing of the original dataset to obtain the target data table includes:

[0078] Identify the procurement business system identifier code corresponding to each of the original data tables in the original dataset;

[0079] The original dataset is categorized according to the identification code of the procurement business system;

[0080] The original datasets from the same procurement business system are designated as the first dataset, and the original datasets from the remaining different procurement business systems are designated as the second dataset.

[0081] The original data tables in the first dataset are merged to obtain a first data table, and the original data tables in the second dataset are merged to obtain a second data table;

[0082] Merge the first data table and the second data table to obtain the third data table;

[0083] The third data table is cleaned to obtain the target data table.

[0084] In this embodiment, the preset cleaning strategy corresponding to the procurement business system is used to clean the third data table to obtain the target data table, which includes:

[0085] Specifically, the cleaning strategy can be pre-set according to the format requirements of the procurement business system and the third-party system that interacts with the procurement business system. By cleaning the third data table through the preset cleaning strategy, the accuracy of the target data table is improved.

[0086] In this embodiment, the procurement business system identifier is used to uniquely identify the system to which each original data table in the original dataset belongs.

[0087] In this embodiment, since the original data tables in the first dataset originate from the same procurement business system or a subsystem of the same procurement business system, and the data table structures of the original data tables in the first dataset are the same, a union process is directly performed on all the original data tables in the first dataset to obtain the first data table.

[0088] Further, the step of merging the original data tables in the second dataset to obtain the second data table includes:

[0089] Identify the field names of each original data table in the second dataset;

[0090] Merge multiple columns with the same field name from multiple original data tables, and delete the columns involved in the merger to obtain a new table;

[0091] Calculate the total number of field names in the new table, and obtain the preset total number of empty fields based on the total number of field names in the new table;

[0092] The new table is expanded based on the preset total number of empty fields to obtain a second data table.

[0093] In this embodiment, the field names of each original data table are the names specified for each column in the table.

[0094] In this embodiment, the original data tables in the second dataset come from different procurement business systems. The data table structures of different procurement business systems are different, so the original data tables in the second dataset need to be processed.

[0095] In this embodiment, the total number of empty fields can be preset. The total number of empty fields is different for different data tables. For example, the second dataset contains original data table A, original data table B, and original data table C. Original data table A contains 20 field names, original data table B contains 20 field names, and original data table C contains 15 field names. There are 15 identical field names in original data table A and original data table B. The identical field names in original data table A and original data table B are merged, and original data table C is also merged to obtain a new table. The new table contains 40 field names. The preset empty field 5 corresponding to the 40 field names is obtained to obtain the second data table.

[0096] In this implementation, during the process of obtaining the second data table from the original data table in the second dataset, the new table is expanded based on the preset total number of empty fields. This makes it easier to directly add new field names to the second data table without reprocessing, thus improving the utilization rate of the second data table.

[0097] S13, invoke rule class tags, and based on the rule class tags, obtain the first prediction result from the target data table through the tag management system.

[0098] In this embodiment, rule-based tags refer to the relationships between various field names in the target data table. The target data table contains rule-based tags: Personnel Information - Project Information - Supplier Information. The first prediction result is the prediction result corresponding to the rule-based tag found in the target data table, that is, the associated data corresponding to Personnel Information - Project Information - Supplier Information is retrieved from the target data table.

[0099] In an optional embodiment, the step of invoking rule class tags and obtaining a first prediction result based on the rule class tags through a preset rule engine includes:

[0100] The data in the target data table is converted into rule class tags that can be recognized by a preset rule engine;

[0101] The preset rule engine is invoked to perform intelligent matching of rule class tags and obtain the first prediction result.

[0102] In this embodiment, rule class tags are pre-stored in the database. After obtaining the target data table, the rule class tags are called, and the data in the target data table is converted into rule class tags that can be recognized by the preset rule engine. The data is then matched with the called rule class tags to obtain the first prediction result.

[0103] In this embodiment, a rule engine can be pre-configured. Specifically, the rule engine performs intelligent matching of rule class tags, which is existing technology and will not be described in detail here.

[0104] In this embodiment, the first prediction result is obtained by classifying the target data table based on rule class tags through a preset rule engine, specifically by considering the institutional regulations of the procurement project.

[0105] S14, extract the feature factor set from the target data table, and input the feature factor set into the pre-trained classification prediction model to obtain the second prediction result.

[0106] In this embodiment, the feature factor refers to the feature factor obtained by processing the field name that is related to the prediction of procurement fraud. For example, the field name of personnel information-project information-supplier information is determined as a feature factor.

[0107] In an optional embodiment, extracting the feature factor set from the target data table includes:

[0108] Obtain the project information and preset processing script of the procurement project corresponding to the procurement project identifier code in the procurement fraud prediction request;

[0109] The project information is preprocessed to obtain the target project information;

[0110] The target project information is processed using the preset processing script to obtain a set of feature factors.

[0111] In this embodiment, the target project information is obtained by preprocessing the project information. Specifically, the data corresponding to the field names in the project information that are related to procurement fraud is preprocessed, for example, by correction and format conversion.

[0112] In this embodiment, a classification prediction model can be pre-trained. After the classification prediction model is trained, the extracted feature factor set is input into the classification prediction model to determine whether there is fraud information in the data information corresponding to each feature factor. The prediction result is output to obtain a second prediction result, which is obtained from the process of the procurement project.

[0113] In this embodiment, the input to the classification prediction model is a set of feature factors. The set of feature factors is used as a training set to train a preset neural network model to obtain the classification prediction model. The specific training process is not limited here.

[0114] S15, perform a second preprocessing on the first prediction result and the second prediction result to obtain the target prediction result.

[0115] In this embodiment, the second preprocessing is obtained by deduplicating the first prediction result and the second prediction result.

[0116] In an optional embodiment, the second preprocessing of the first prediction result and the second prediction result to obtain the target prediction result includes:

[0117] Compare the first prediction result with the second prediction result;

[0118] Delete the first prediction result that is the same as the second prediction result to obtain the target prediction result.

[0119] Furthermore, the method also includes:

[0120] The target prediction result is sent to the client, and the target prediction result is displayed on the client's interface.

[0121] Furthermore, the method also includes:

[0122] The target prediction result is added to the training set of the training classification prediction model, and the classification prediction model is retrained.

[0123] In this embodiment, since the first prediction result is obtained based on rule-based class labels and the second prediction result is obtained based on a pre-trained classification prediction model, the first and second prediction results are compared. For example, if both the first and second prediction results contain the statement that user A has engaged in procurement fraud, the statement that user A has engaged in procurement fraud is deleted from the first prediction result, while the statement that user A has engaged in procurement fraud is retained from the second prediction result. By combining the rule engine and the classification prediction model, the completeness and accuracy of the target prediction result are improved.

[0124] In summary, the AI-based behavior prediction method described in this embodiment performs a first preprocessing on the original dataset, calls rule-based labels, obtains a first prediction result from the target data table based on the rule-based labels through a label management system, and inputs the feature factor set into a pre-trained classification prediction model to obtain a second prediction result. The first prediction result is obtained based on the institutional regulations of the procurement project and is based on a rule engine, while the second prediction result is obtained based on the process of the procurement project and is based on a classification prediction model. By obtaining the target prediction result based on the rule engine and the classification prediction model, the accuracy of the procurement fraud prediction result is improved.

[0125] Example 2

[0126] Figure 2 This is a structural diagram of the behavior prediction device based on artificial intelligence provided in Embodiment 2 of the present invention.

[0127] In some embodiments, the AI-based behavior prediction device 20 may include multiple functional modules composed of program code segments. The program code of each program segment in the AI-based behavior prediction device 20 may be stored in the memory of an electronic device and executed by the at least one processor to perform (see details). Figure 1 (Description) Functionality based on artificial intelligence for behavior prediction.

[0128] In this embodiment, the AI-based behavior prediction device 20 can be divided into multiple functional modules according to the functions it performs. These functional modules may include: a parsing module 201, a first preprocessing module 202, a construction module 203, an extraction and input module 204, a second preprocessing module 205, a sending module 206, and an adding module 207. The module referred to in this invention is a series of computer-readable instruction segments that can be executed by at least one processor and perform a fixed function, stored in memory. In this embodiment, the functions of each module will be detailed in subsequent embodiments.

[0129] The parsing module 201 is used to parse the received procurement fraud prediction request and obtain the data source of multiple target parameters, wherein the procurement fraud prediction request includes a procurement project identification code.

[0130] In this embodiment, when an enterprise or user performs procurement fraud prediction, it initiates a procurement fraud prediction request to the server through a client. Specifically, the client can be a smartphone, iPad, or other existing smart device, and the server can be a procurement fraud prediction subsystem. During the procurement fraud prediction process, the client can send a procurement fraud prediction request to the procurement fraud prediction subsystem, which is used to receive the procurement fraud prediction request sent by the client.

[0131] In this embodiment, the data source for the target parameter can be at least one third-party system that interacts with the procurement business system, or at least one subsystem of the procurement business system.

[0132] In an optional embodiment, the parsing module 201 parses the received procurement fraud prediction request to obtain the data sources for multiple target parameters, including:

[0133] Parse the message of the procurement fraud prediction request to obtain the message information carried by the message;

[0134] Obtain the procurement fraud prediction requirements from the aforementioned message information;

[0135] Data sources for multiple target parameters are obtained based on the procurement fraud prediction requirements.

[0136] Specifically, the procurement fraud prediction requirement includes a procurement project identification code, which is used to uniquely identify the identity of the procurement business.

[0137] Furthermore, the data source for obtaining multiple target parameters based on the procurement fraud prediction requirement includes:

[0138] The procurement project identifier code in the procurement fraud prediction requirement is matched with the system identifier code in the preset database to obtain the first interface of the procurement business system corresponding to the procurement project and the second interface of the third-party system that interacts with the procurement business system.

[0139] Call the first interface and the second interface to obtain the data source of the corresponding target parameters.

[0140] In this embodiment, since the associated data involved in the procurement business system may come from the procurement business system, third-party systems that interact with the procurement business system, and subsystems of the procurement business system, the data source of the corresponding target parameters is obtained by calling the interface of the corresponding procurement business system. It is not necessary to obtain the data source of all target parameters of all procurement business systems, which is more targeted and improves the accuracy of the data source of the target parameters.

[0141] The first preprocessing module 202 is used to obtain the original dataset from the data source of the multiple target parameters, and perform a first preprocessing on the original dataset to obtain the target data table.

[0142] In this embodiment, the original dataset refers to the original data obtained directly from the data source of the target parameter, and the target data table refers to the data obtained by processing each original data table in the original dataset.

[0143] In an optional embodiment, the first preprocessing module 202 obtains the raw dataset from the data source of the plurality of target parameters by:

[0144] The procurement fraud prediction request is analyzed to obtain the data source collection requirements and corresponding collection interfaces for each target parameter.

[0145] Call the data collection interface of each target parameter, and use web crawling technology to obtain the raw dataset from the data source of each target parameter according to the data collection requirements of each target parameter's data source.

[0146] In this embodiment, the procurement fraud prediction requirement also includes the data source collection requirements and corresponding collection interfaces for each target parameter. The web crawler technology is existing technology and will not be described in detail here.

[0147] In an optional embodiment, the first preprocessing module 202 performs a first preprocessing on the original dataset to obtain a target data table, including:

[0148] Identify the procurement business system identifier code corresponding to each of the original data tables in the original dataset;

[0149] The original dataset is categorized according to the identification code of the procurement business system;

[0150] The original datasets from the same procurement business system are designated as the first dataset, and the original datasets from the remaining different procurement business systems are designated as the second dataset.

[0151] The original data tables in the first dataset are merged to obtain a first data table, and the original data tables in the second dataset are merged to obtain a second data table;

[0152] Merge the first data table and the second data table to obtain the third data table;

[0153] The third data table is cleaned to obtain the target data table.

[0154] In this embodiment, the preset cleaning strategy corresponding to the procurement business system is used to clean the third data table to obtain the target data table, which includes:

[0155] Specifically, the cleaning strategy can be pre-set according to the format requirements of the procurement business system and the third-party system that interacts with the procurement business system. By cleaning the third data table through the preset cleaning strategy, the accuracy of the target data table is improved.

[0156] In this embodiment, the procurement business system identifier is used to uniquely identify the system to which each original data table in the original dataset belongs.

[0157] In this embodiment, since the original data tables in the first dataset originate from the same procurement business system or a subsystem of the same procurement business system, and the data table structures of the original data tables in the first dataset are the same, a union process is directly performed on all the original data tables in the first dataset to obtain the first data table.

[0158] Further, the step of merging the original data tables in the second dataset to obtain the second data table includes:

[0159] Identify the field names of each original data table in the second dataset;

[0160] Merge multiple columns with the same field name from multiple original data tables, and delete the columns involved in the merger to obtain a new table;

[0161] Calculate the total number of field names in the new table, and obtain the preset total number of empty fields based on the total number of field names in the new table;

[0162] The new table is expanded based on the preset total number of empty fields to obtain a second data table.

[0163] In this embodiment, the field names of each original data table are the names specified for each column in the table.

[0164] In this embodiment, the original data tables in the second dataset come from different procurement business systems. The data table structures of different procurement business systems are different, so the original data tables in the second dataset need to be processed.

[0165] In this embodiment, the total number of empty fields can be preset. The total number of empty fields is different for different data tables. For example, the second dataset contains original data table A, original data table B, and original data table C. Original data table A contains 20 field names, original data table B contains 20 field names, and original data table C contains 15 field names. There are 15 identical field names in original data table A and original data table B. The identical field names in original data table A and original data table B are merged, and original data table C is also merged to obtain a new table. The new table contains 40 field names. The preset empty field 5 corresponding to the 40 field names is obtained to obtain the second data table.

[0166] In this implementation, during the process of obtaining the second data table from the original data table in the second dataset, the new table is expanded based on the preset total number of empty fields. This makes it easier to directly add new field names to the second data table without reprocessing, thus improving the utilization rate of the second data table.

[0167] The construction module 203 is used to call the rule class label and, based on the rule class label, obtain the first prediction result from the target data table through the label management system.

[0168] In this embodiment, rule-based tags refer to the relationships between various field names in the target data table. The target data table contains rule-based tags: Personnel Information - Project Information - Supplier Information. The first prediction result is the prediction result corresponding to the rule-based tag found in the target data table, that is, the associated data corresponding to Personnel Information - Project Information - Supplier Information is retrieved from the target data table.

[0169] In an optional embodiment, the construction module 203 calls rule class tags, and based on the rule class tags, obtains a first prediction result through a preset rule engine, including:

[0170] The data in the target data table is converted into rule class tags that can be recognized by a preset rule engine;

[0171] The preset rule engine is invoked to perform intelligent matching of rule class tags and obtain the first prediction result.

[0172] In this embodiment, rule class tags are pre-stored in the database. After obtaining the target data table, the rule class tags are called, and the data in the target data table is converted into rule class tags that can be recognized by the preset rule engine. The data is then matched with the called rule class tags to obtain the first prediction result.

[0173] In this embodiment, a rule engine can be pre-configured. Specifically, the rule engine performs intelligent matching of rule class tags, which is existing technology and will not be described in detail here.

[0174] In this embodiment, the first prediction result is obtained by classifying the target data table based on rule class tags through a preset rule engine, specifically by considering the institutional regulations of the procurement project.

[0175] The extraction and input module 204 is used to extract a set of feature factors from the target data table and input the set of feature factors into a pre-trained classification prediction model to obtain a second prediction result.

[0176] In this embodiment, the feature factor refers to the feature factor obtained by processing the field name that is related to the prediction of procurement fraud. For example, the field name of personnel information-project information-supplier information is determined as a feature factor.

[0177] In an optional embodiment, the extraction and input module 204 extracts a set of feature factors from the target data table, including:

[0178] Obtain the project information and preset processing script of the procurement project corresponding to the procurement project identifier code in the procurement fraud prediction request;

[0179] The project information is preprocessed to obtain the target project information;

[0180] The target project information is processed using the preset processing script to obtain a set of feature factors.

[0181] In this embodiment, the target project information is obtained by preprocessing the project information. Specifically, the data corresponding to the field names in the project information that are related to procurement fraud is preprocessed, for example, by correction and format conversion.

[0182] In this embodiment, a classification prediction model can be pre-trained. After the classification prediction model is trained, the extracted feature factor set is input into the classification prediction model to determine whether there is fraud information in the data information corresponding to each feature factor. The prediction result is output to obtain a second prediction result, which is obtained from the process of the procurement project.

[0183] In this embodiment, the input to the classification prediction model is a set of feature factors. The set of feature factors is used as a training set to train a preset neural network model to obtain the classification prediction model. The specific training process is not limited here.

[0184] The second preprocessing module 205 is used to perform a second preprocessing on the first prediction result and the second prediction result to obtain the target prediction result.

[0185] In this embodiment, the second preprocessing is obtained by deduplicating the first prediction result and the second prediction result.

[0186] In an optional embodiment, the second preprocessing module 205 performs a second preprocessing on the first prediction result and the second prediction result to obtain the target prediction result, including:

[0187] Compare the first prediction result with the second prediction result;

[0188] Delete the first prediction result that is the same as the second prediction result to obtain the target prediction result.

[0189] The sending module 206 is used to send the target prediction result to the client and display the target prediction result on the client's interface.

[0190] Add module 207, which is used to add the target prediction result to the training set of the training classification prediction model and retrain the classification prediction model.

[0191] In this embodiment, since the first prediction result is obtained based on rule-based class labels and the second prediction result is obtained based on a pre-trained classification prediction model, the first prediction result and the second prediction result are compared. For example, if the first prediction result contains the statement that user A has engaged in procurement fraud, and the second prediction result also contains the statement that user A has engaged in procurement fraud, the statement that user A has engaged in procurement fraud in the first prediction result is deleted, while the statement that user A has engaged in procurement fraud in the second prediction result is retained. By combining the rule engine and the classification prediction model, the target prediction result is obtained, which improves the completeness and accuracy of the target prediction result.

[0192] In summary, the AI-based behavior prediction device described in this embodiment performs a first preprocessing on the original dataset, calls rule-based labels, obtains a first prediction result from the target data table based on the rule-based labels through a label management system, and inputs the feature factor set into a pre-trained classification prediction model to obtain a second prediction result. The first prediction result is obtained based on the rules and regulations of the procurement project and is based on a rule engine, while the second prediction result is obtained based on the process of the procurement project and is based on a classification prediction model. By obtaining the target prediction result based on the rule engine and the classification prediction model, the accuracy of the procurement fraud prediction result is improved.

[0193] Example 3

[0194] See Figure 3The diagram shown is a structural schematic of an electronic device provided in Embodiment 3 of the present invention. In a preferred embodiment of the present invention, the electronic device 3 includes a memory 31, at least one processor 32, at least one communication bus 33, and a transceiver 34.

[0195] Those skilled in the art should understand that Figure 3 The structure of the electronic device shown does not constitute a limitation of the embodiments of the present invention. It can be a bus structure or a star structure. The electronic device 3 may also include more or fewer other hardware or software than shown, or different component arrangements.

[0196] In some embodiments, the electronic device 3 is an electronic device capable of automatically performing numerical calculations and / or information processing according to pre-set or stored instructions. Its hardware includes, but is not limited to, microprocessors, application-specific integrated circuits (ASICs), programmable gate arrays (FPGAs), digital processors, and embedded devices. The electronic device 3 may also include client devices, including, but not limited to, any electronic product capable of human-computer interaction with a client via a keyboard, mouse, remote control, touchpad, or voice control device, such as personal computers, tablet computers, smartphones, and digital cameras.

[0197] It should be noted that the electronic device 3 is merely an example. Other existing or future electronic products that are suitable for this invention should also be included within the scope of protection of this invention and are incorporated herein by reference.

[0198] In some embodiments, the memory 31 is used to store program code and various data, such as the AI-based behavior prediction device 20 installed in the electronic device 3, and to achieve high-speed, automatic access to programs or data during the operation of the electronic device 3. The memory 31 includes read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0199] In some embodiments, the at least one processor 32 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits packaged with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The at least one processor 32 is the control unit of the electronic device 3, connecting various components of the entire electronic device 3 via various interfaces and lines. It executes programs or modules stored in the memory 31 and calls data stored in the memory 31 to perform various functions and process data of the electronic device 3.

[0200] In some embodiments, the at least one communication bus 33 is configured to enable communication between the memory 31 and the at least one processor 32, etc.

[0201] Although not shown, the electronic device 3 may also include a power supply (such as a battery) to power the various components. Optionally, the power supply may be logically connected to the at least one processor 32 via a power management device, thereby enabling functions such as charging, discharging, and power consumption management. The power supply may also include one or more DC or AC power sources, a recharging device, a power fault detection circuit, a power converter or inverter, a power status indicator, or any other components. The electronic device 3 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0202] It should be understood that the embodiments described are for illustrative purposes only and are not limited to this structure in the scope of the patent application.

[0203] The integrated unit implemented as a software functional module described above can be stored in a computer-readable storage medium. This software functional module, stored in a storage medium, includes several instructions to cause a computer device (which may be a personal computer, electronic device, or network device, etc.) or processor to execute portions of the methods described in the various embodiments of the present invention.

[0204] In a further embodiment, combined with Figure 2 The at least one processor 32 can execute the operating device of the electronic device 3 and various installed applications (such as the artificial intelligence-based behavior prediction device 20), program code, etc., for example, the various modules mentioned above.

[0205] The memory 31 stores program code, and the at least one processor 32 can call the program code stored in the memory 31 to execute related functions. For example, Figure 2 The modules described herein are program codes stored in the memory 31 and executed by the at least one processor 32, thereby realizing the functions of the modules to achieve the purpose of behavior prediction based on artificial intelligence.

[0206] For example, the program code can be divided into one or more modules / units, which are stored in the memory 31 and executed by the processor 32 to complete this application. The one or more modules / units can be a series of computer-readable instruction segments capable of performing a specific function, which describe the execution process of the program code in the electronic device 3. For example, the program code can be divided into a parsing module 201, a first preprocessing module 202, a construction module 203, an extraction and input module 204, a second preprocessing module 205, a sending module 206, and an adding module 207.

[0207] In one embodiment of the present invention, the memory 31 stores a plurality of computer-readable instructions, which are executed by the at least one processor 32 to achieve a behavior prediction function based on artificial intelligence.

[0208] Specifically, the specific implementation method of the above instructions by the at least one processor 32 can be referred to Figure 1 The descriptions of the relevant steps in the corresponding embodiments are not repeated here.

[0209] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0210] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0211] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0212] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered illustrative and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims. Furthermore, it is clear that the word "comprising" does not exclude other elements, and the singular does not exclude the plural. Multiple elements or devices recited in the present invention may also be implemented by a single element or device in software or hardware. The terms "first," "second," etc., are used to denote names and do not indicate any particular order.

[0213] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A behavior prediction method based on artificial intelligence, characterized in that, The method includes: Parse the received procurement fraud prediction request to obtain the data source for multiple target parameters; The process involves obtaining raw datasets from the data sources of the multiple target parameters, and performing a first preprocessing step on the raw datasets to obtain target data tables. This includes: identifying the procurement business system identifier code corresponding to each raw data table in the raw dataset; classifying the raw datasets according to the procurement business system identifier code; identifying raw datasets from the same procurement business system as a first dataset, and identifying the remaining raw datasets from different procurement business systems as a second dataset; merging the raw data tables in the first dataset to obtain a first data table, and merging the raw data tables in the second dataset to obtain a second data table; merging the first data table and the second data table to obtain a third data table; and cleaning the third data table to obtain the target data table. Invoke the rule-based tags, and based on the rule-based tags, obtain the first prediction result from the target data table through the tag management system; Extracting a feature factor set from the target data table includes: obtaining the project information of the procurement project corresponding to the procurement project identifier code in the procurement fraud prediction request and a preset processing script; preprocessing the data corresponding to the field names in the project information that are related to procurement fraud to obtain target project information; processing the target project information using the preset processing script to obtain a feature factor set, and inputting the feature factor set into a pre-trained classification prediction model to obtain a second prediction result; The first prediction result and the second prediction result are subjected to a second preprocessing to obtain a target prediction result, including: comparing the first prediction result with the second prediction result; deleting the first prediction result that is the same as the second prediction result to obtain the target prediction result.

2. The behavior prediction method based on artificial intelligence as described in claim 1, characterized in that, The step of merging the original data tables in the second dataset to obtain the second data table includes: Identify the field names of each original data table in the second dataset; Merge multiple columns with the same field name from multiple original data tables, and delete the columns involved in the merger to obtain a new table; Calculate the total number of field names in the new table, and obtain the preset total number of empty fields based on the total number of field names in the new table; The new table is expanded based on the preset total number of empty fields to obtain a second data table.

3. The behavior prediction method based on artificial intelligence as described in claim 1, characterized in that, The invocation of rule class tags, and the acquisition of a first prediction result based on the rule class tags through a preset rule engine, includes: The data in the target data table is converted into rule class tags that can be recognized by a preset rule engine; The preset rule engine is invoked to perform intelligent matching of rule class tags and obtain the first prediction result.

4. The behavior prediction method based on artificial intelligence as described in claim 1, characterized in that, The data sources for parsing the received procurement fraud prediction request and obtaining multiple target parameters include: Parse the message of the procurement fraud prediction request to obtain the message information carried by the message; Obtain the procurement fraud prediction requirements from the aforementioned message information; The procurement project identifier code in the procurement fraud prediction requirement is matched with the system identifier code in the preset database to obtain the first interface of the procurement business system corresponding to the procurement project and the second interface of the third-party system that interacts with the procurement business system. Call the first interface and the second interface to obtain the data source of the corresponding target parameters.

5. A behavior prediction device based on artificial intelligence, characterized in that, The apparatus is used to implement the artificial intelligence-based behavior prediction method as described in any one of claims 1 to 4, the apparatus comprising: The parsing module is used to parse the received procurement fraud prediction request and obtain the data source of multiple target parameters; The first preprocessing module is used to obtain the original dataset from the data source of the multiple target parameters, and perform a first preprocessing on the original dataset to obtain the target data table; The construction module is used to call rule-based tags and, based on the rule-based tags, obtain a first prediction result from the target data table through the tag management system; An extraction and input module is used to extract a set of feature factors from the target data table and input the set of feature factors into a pre-trained classification prediction model to obtain a second prediction result; The second preprocessing module is used to perform a second preprocessing on the first prediction result and the second prediction result to obtain the target prediction result.

6. An electronic device, characterized in that, The electronic device includes a processor and a memory, wherein the processor is configured to implement the artificial intelligence-based behavior prediction method as described in any one of claims 1 to 4 when executing a computer program stored in the memory.

7. A computer-readable storage medium storing a computer program thereon, characterized in that, When the computer program is executed by the processor, it implements the artificial intelligence-based behavior prediction method as described in any one of claims 1 to 4.

Images