An artificial intelligence-based multi-dimensional data acquisition and analysis system and method

Abstract

The application discloses a multi-dimensional data collection and analysis system and method based on artificial intelligence, relates to the technical field of multi-dimensional data analysis, and collects historical student subsidy file data from different data sources through file import and API interface calling; the collected historical student subsidy file data is classified and stored; artificial intelligence algorithms are used to analyze the data, obtain corresponding rules, obtain current to-be-detected subsidy file data, and analyze the to-be-detected file; the use of student cards and the situation of working-study are regularly sampled and detected to analyze whether there are missed subsidy objects and subsidy objects that do not meet the requirements; the application introduces a living difficulty index and a dynamically changing sampling detection frequency, solves the problem that whether there are missed subsidy objects in students who do not submit subsidy applications cannot be accurately determined in time, introduces an abnormal living standard evaluation value, and improves the accuracy of subsidy file data auditing results.

Description

Technical Field

[0001] This invention relates to the field of multi-dimensional data analysis technology, specifically a multi-dimensional data acquisition and analysis system and method based on artificial intelligence. Background Technology

[0002] With the development of big data technology, the demand for the collection, fusion, and analysis of multi-source heterogeneous data across various industries has exploded. Traditional data analysis systems suffer from significant technical limitations; the single data source leads to biased analysis results, failing to construct comprehensive data profiles and thus unable to meet the real-time, accuracy, and intelligent requirements of modern application scenarios. In student financial aid management, aid files contain a large amount of multi-dimensional data, including student personal information, family circumstances, and historical aid records. The eligibility review for student financial aid in schools requires the evaluation of numerous data types, which must be obtained from different systems, each with different formats and quantities, making the integration and analysis of all data complex. Traditional aid file management methods primarily rely on… Relying on manual operations leads to inefficiencies, poor accuracy, and an inability to effectively utilize historical data to summarize patterns in data collection, processing, analysis, and document review. Existing data analysis technologies, particularly in some student aid management systems that employ simple rule matching, are ill-suited to complex and ever-changing aid scenarios. Furthermore, after the initial eligibility review of aid recipients, it is difficult to track and analyze any anomalies, or to promptly identify students who did not apply for aid but whose circumstances changed and who now require it. Relying solely on student application documents cannot guarantee the authenticity of the submitted information or the accuracy of the review results. Therefore, there is an urgent need for a method capable of analyzing and processing multidimensional data to improve the efficiency and automation of aid information management. Summary of the Invention

[0003] The purpose of this invention is to provide a multi-dimensional data acquisition and analysis system and method based on artificial intelligence to solve the problems raised in the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a multi-dimensional data acquisition and analysis method based on artificial intelligence, comprising the following steps: S1. Obtain student basic information, family situation, and historical financial aid records from different data sources through file import and API interface calls; S2. Classify and store the collected historical student financial aid data; use artificial intelligence algorithms to analyze the classified and stored historical financial aid data to obtain the correspondence between family circumstances and financial aid types; S3. Obtain the current funding file data to be tested, and use the set thresholds and corresponding patterns to review and analyze whether the file to be tested meets the funding requirements and whether the family situation type matches the type of funding applied for. S4. By regularly sampling and testing students' student card usage and work-study activities, we can analyze whether there are any missed recipients among students who have not submitted financial aid applications, and at the same time, we can analyze whether there are any ineligible recipients based on their living standards.

[0005] Furthermore, in step S1: after authorization, basic student information and family information are obtained from the student financial aid management system and student registration system through file import and database connection; historical financial aid records of students are obtained from the database of the civil affairs department of the student's place of origin through API interface call; the basic student information includes the student's name and student ID; the family information includes the number of working people in the family m and family poverty certificate documents; the historical financial aid records include historical financial aid records and student loan application information; and the student's student card usage frequency is obtained from the student card transaction system in real time through API interface, and the student's work-study registration records are obtained from the student work-study records at regular intervals.

[0006] Furthermore, in step S2: the acquired data is classified according to the set information classification criteria, and the information obtained from different databases is divided into student basic information, family situation information, and historical financial aid record information. The acquired data is then classified and stored according to the classification results. A combination of relational and non-relational databases is used for data storage. The relational database stores student basic information, while the non-relational database stores family situation information and historical financial aid record information. Indexes are created for the data in both databases for querying specific student data. In the relational database, primary key indexes are created for the three unique identifier fields of student name, ID number, and student ID to locate specific student data, and ordinary indexes are created for the two types of data: family labor force quantity and family poverty certificate. In the non-relational database, a single-field index is created for student unique identifiers such as student ID to quickly retrieve corresponding document data based on the student ID. A composite index is created by combining data classification tags such as family hardship certificates and award certificates with student unique identifiers. The composite index optimizes multi-condition queries and significantly improves query efficiency when data needs to be queried simultaneously based on student ID and data classification tags. The Apriori algorithm is used to analyze the historically archived funding data after classified storage, so as to analyze the correspondence between the information of students' family situations and the types of grants obtained by the students; the absolute poverty line and the minimum subsistence security line set by the civil affairs department are used to divide the family situations into absolutely poor families, minimum subsistence security poor families and relatively poor families according to the poverty certificates; the number of family labor forces is divided into three categories, namely m≤m1, m1<m≤m2, and m>m2; and the historical grant types are divided into national grant, provincial grant and campus grant; the Apriori algorithm is used to analyze the historical data, so as to find out the grant types corresponding to different types of family situations, and obtain the correspondence rule between the family situation types and the grant types, where the correspondence rule is as follows: Absolutely poor family, with the number of family labor forces m≤m1 → obtaining the national grant; Minimum subsistence security poor family, with the number of family labor forces m1<m≤m2 → obtaining the provincial grant; Relatively poor family, with the number of family labor forces m>m2 → obtaining the campus grant.

[0007] Further, in step S3: the to-be-detected current funding archive data is acquired, and it is judged whether the submitted funding archive data meets the funding requirements; a threshold is set for judging the student funding archives that do not meet the funding requirements, and the maximum threshold for the number of family labor forces is set as m4; if it is detected that there is a case of m≥m4 in the to-be-detected student funding archive, it is judged that the student funding archive does not meet the funding requirements, an abnormality prompt indicating that there is a student funding archive that does not meet the funding requirements is issued, and the student funding archive is marked as unapproved; if it is detected that all data analysis results in the to-be-detected student funding archive show m<m4, it is judged that the student funding archive meets the funding requirements, and the student funding archive is marked as approved; for the student funding archives marked as approved, the to-be-detected archives are further reviewed and analyzed in combination with the summarized correspondence rule between the family situation types and the grant types, and an abnormality prompt of mismatched funding situations is issued for the case where the family situation type does not match the applied grant type; for example, after analysis, a student's archive data is judged as: a minimum subsistence security poor family with the number of family labor forces m1<m≤m2, but the applied grant type is the campus grant, it can be judged that the family situation type does not match the applied grant type, and an abnormality prompt of mismatched funding situations is issued to remind the staff to verify the actual situation of the student and adjust the grant type; this greatly saves the manpower consumption for reviewing the funding archives, and improves the review efficiency; moreover, the probability of matching the student's family situation with the grant type is improved according to the summarized rule, and the balance of funding resource allocation is guaranteed.

[0008] Furthermore, in step S4: after the file analysis and review of students who have applied for funding is completed, the student card usage and work-study activities of students are sampled and tested according to the sampling frequency q; when sampling and testing all students, the single sampling period is k days; the number of times the student card is used L within k days is obtained from the student card transaction system; and the daily work-study duration {Δt1, Δt2, ..., Δt} of each student is obtained from the work-study records. k Set the maximum daily work-study time threshold Δt. max and the minimum threshold Δt for daily work-study hours min ; will Δt i ≥Δt max The number of days is denoted as k2, where i = 1, 2, ..., k; the student's living hardship index is calculated using the following formula: ; Where W represents the hardship index; The sampling frequency q for sampling and detecting students' student card usage and work-study activities is a dynamically changing value; the sampling frequency q is calculated according to the following formula: ; Where q0 represents the basic sampling frequency; the above formula indicates that the actual sampling frequency is... The results after rounding up show that the sampling frequency during the sampling and detection of students' student card usage and work-study activities is related to the obtained living hardship index. The sampling frequency is a dynamic value. As the living hardship index W increases, the sampling frequency increases. When students with living hardship are identified, the sampling frequency can be automatically adjusted, which improves the efficiency of identifying missed aid recipients from risk. Set a threshold W for the life hardship index. max The obtained living hardship index was analyzed, and the results are as follows: If W <W max This indicates that the frequency of students' student card usage and the duration of their work-study programs were normal during this sampling period, and it was determined that no students experienced any sudden financial hardship. If W≥W max If the frequency of student card use and the duration of work-study are too high during the current sampling period, it is determined that the student is experiencing a sudden financial hardship. In this case, the student's student ID is extracted and the student is marked as a key focus. For students marked as key focus, the sampling period for which they are marked as key focus is recorded as the first sampling period. The living hardship index of key focus students will be analyzed in the next two sampling tests.

[0009] Furthermore, in the subsequent sampling and detection, the observation period is set from the sampling period recorded as the first sampling. A key observation mode for the key observation objects is triggered. The initial value of the total number of samples in the key observation mode is set to c, which is a dynamically changing value. The W value obtained from each sampling and detection is analyzed, and the obtained W value is compared with the set threshold W. max A comparison is performed, and the analysis results of each of the c′ samplings are compared. If the analysis results of c′ consecutive samplings are all W≥W, the comparison is successful. max If the result of 2 consecutive c′ comparisons is W≥W, then subtract 1 from the value of c; max If the key observation mode is lifted, the key observation target will be identified as the target of funding; if the c′ consecutive comparative analysis results are all W <W max If the result is W, then increment the value of c by 1; if 2c′ consecutive comparisons are both W, then increment the value of c by 1. <W max If so, the key observation mode will be directly lifted, and the key observation target will be judged as an entity that does not need funding; After the number of samplings for the key observation mode reaches c and the analysis is completed, the sampling results from the c sampling tests are defined as W≥W. max The number of times is denoted as c1. By analyzing the value of c1, we determine whether the key focus group is in need of funding. The analysis results are as follows: If c1 < 2c′, it is determined that the target of this focus is not the target of funding. If c1≥2c′, the student in question is identified as needing financial aid. Further analysis is performed by comparing the student's basic information with that of students whose financial aid applications have been reviewed. If the student has submitted a financial aid application and passed the review, no additional notification is issued. If the student has not submitted a financial aid application, a notification is issued indicating that a student has been overlooked. The concept of financial hardship is introduced to determine if any students who have not submitted applications have been overlooked, thus addressing the problem of not being able to promptly identify students facing financial difficulties due to unforeseen circumstances.

[0010] Furthermore, when sampling and testing students' student card usage and work-study activities, for students already determined to meet the funding requirements, an additional analysis of abnormal living standard assessment values ​​will be added, including Δt. i <Δt min The number of days is denoted as k3. The abnormal standard of living assessment value is calculated according to the following formula: ; wherein M is the abnormal living standard assessment value; a higher value of k3 / k indicates a higher probability that the student's daily work-study time is lower than the minimum daily work-study time threshold; wherein a higher value of 3k / L indicates a lower frequency of the student using the student card, and using the student card for on-campus dining and shopping is much more cost-effective than off-campus, so the lower the frequency of student card use, the more it indicates that the student does not need to be frugal in terms of dining and shopping; the higher the abnormal living standard assessment value obtained by combining the two, the more it indicates that the student still does not need to be frugal even when the work-study time is not long, which means that the student has a high living standard and may not meet the conditions requiring funding support; setting an abnormal living standard assessment value threshold M0; performing comparative analysis on the calculated abnormal living standard assessment value and the set threshold, and the analysis results are as follows: If M<M0, it indicates that the student does not have a living standard exceeding the family's affordability; If M≥M0, it indicates that the student has a living standard exceeding the family's affordability, then the abnormal living standard assessment value of this student will be analyzed emphatically in the next two sampling detections of student card usage and work-study situation. If it is judged for three consecutive times that the student has a living standard exceeding the family's affordability, an abnormal prompt for student funding file review is issued, to prompt the staff to verify the authenticity of the information in the student's funding file and re-review the student's funding file. By introducing the abnormal living standard assessment value for the funded students who have passed the review, and judging whether the student's living standard is consistent with the submitted family situation according to the detection of student card usage and work-study situation, and verifying and re-reviewing the information submitted by the funded students with abnormal living standard assessment values, the accuracy of the review result of the funding file data is improved.

[0011] A multi-dimensional data collection and analysis system based on artificial intelligence, the system includes a multi-dimensional data collection module, a data analysis and storage module, a funding file analysis and review module, and an abnormality identification and feedback module; The multi-dimensional data collection module obtains basic student information, family situation and historical funding records from different data sources by means of file import and API interface call; The data analysis and storage module classifies and stores the collected historical student funding file data; uses an artificial intelligence algorithm to analyze the classified and stored historical funding file data to obtain the corresponding rule between the family situation and the funding type; The funding file analysis and review module obtains the current funding file data to be detected, and uses the set threshold and the obtained corresponding rule to review and analyze whether the file to be detected meets the funding requirements and whether the family situation type matches the applied funding type; The anomaly detection and feedback module periodically samples and detects students' student card usage and work-study activities to analyze whether there are any missed recipients among students who have not submitted funding applications, and at the same time analyzes whether there are any ineligible recipients based on their living standards.

[0012] Furthermore, the multi-dimensional data collection module includes a student personal and family information collection unit, a historical financial aid record collection unit, and a student card and work-study information collection unit. The student personal and family information collection unit obtains basic student information and family information from the student financial aid management system and the student registration system through file import and database connection. The historical financial aid record collection unit obtains the student's historical financial aid record information from the database of the civil affairs department of the student's place of origin through API interface calls. The student card and work-study information collection unit obtains the student's student card usage status from the student card transaction system in real time through API interface, and periodically obtains the student's work-study registration record from the student work-study record.

[0013] Furthermore, the data analysis and storage module includes a data classification and storage unit and a data analysis unit. The data classification and storage unit classifies the acquired data according to the set information classification criteria, dividing the information obtained from different databases into student basic information, family situation information, and historical financial aid records, and stores the acquired data according to the classification results. The data storage method combines relational and non-relational databases, using relational databases to store highly structured student basic information and non-relational databases to store family situation information and historical financial aid records. The data analysis unit uses the Apriori algorithm to analyze the classified and stored historical financial aid archive data, analyzing the correspondence between student family situation information and the types of financial aid received by students.

[0014] Furthermore, the funding file analysis and review module sets a threshold to determine student funding files that do not meet the funding requirements, and marks the student funding files that are judged to be ineligible as approved. For the student funding files marked as approved, the module further reviews and analyzes the files to be tested by combining the summarized correspondence between family situation types and funding types, and issues an abnormal prompt of funding mismatch for cases where the family situation type does not match the applied funding type.

[0015] Furthermore, the anomaly identification and feedback module includes an omission recipient detection unit and an abnormal recipient detection unit. The omission recipient detection unit conducts regular sampling tests by sampling the student card usage and work-study activities of all students, and introduces a living hardship index to analyze the students' living conditions to determine whether there are any omission recipients among the students who have not submitted a grant application. The abnormal recipient detection unit introduces an abnormal living standard assessment value to evaluate the living standards of students who have passed the grant review, and verifies the information of students whose living standards continuously exceed their family's affordability and re-examines their grant files.

[0016] Compared with the prior art, the beneficial effects of the present invention are: This invention collects historical student aid records from multiple data sources and stores the data using a combination of relational and non-relational databases, classifying and indexing key data to improve the efficiency of retrieving specific student data. It analyzes the classified historical aid record data using the Apriori algorithm, summarizing the aid types corresponding to different family circumstances and identifying the correspondence between family circumstances and aid types. This allows for rapid determination of whether the family circumstances match the applied aid type when analyzing new aid records, issuing an anomaly alert for mismatches. This significantly reduces manpower consumption in reviewing aid records and improves review efficiency. Furthermore, the summarized patterns increase the probability of matching student family circumstances with aid types, ensuring a balanced allocation of aid resources. This invention also proposes a periodic... This invention presents a technical solution for sampling and detecting students' student card usage and work-study activities. It introduces a hardship index to determine if any students who haven't submitted aid applications have been overlooked, thus solving the problem of not being able to promptly identify students facing hardship. Furthermore, the sampling frequency for student card usage and work-study activities is linked to the acquired hardship index, with the sampling frequency being a dynamically changing value. This allows for automatic adjustment of the sampling frequency when students experiencing hardship are identified, improving the efficiency of identifying overlooked aid recipients from the initial risk assessment. Additionally, the invention introduces an abnormal living standard assessment value for approved aid recipients. By detecting abnormal living standards, it determines whether a student's living standard matches their submitted family circumstances. Information submitted by aid recipients whose living standards consistently exceed their family's affordability is verified and re-examined, improving the accuracy of aid application review results. Attached Figure Description

[0017] Figure 1This is a flowchart illustrating a multi-dimensional data acquisition and analysis method based on artificial intelligence according to the present invention. Figure 2 This is a schematic diagram of the process for determining missing funding recipients in a multi-dimensional data acquisition and analysis method based on artificial intelligence according to the present invention. Figure 3 This is a schematic diagram of the structure of a multi-dimensional data acquisition and analysis system based on artificial intelligence according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] like Figures 1-3 As shown, this invention provides a technical solution: a multi-dimensional data acquisition and analysis method based on artificial intelligence, such as... Figure 1 As shown, it includes the following steps: S1. Obtain student basic information, family situation, and historical financial aid records from different data sources through file import and API interface calls; S2. Classify and store the collected historical student financial aid data; use artificial intelligence algorithms to analyze the classified and stored historical financial aid data to obtain the correspondence between family circumstances and financial aid types; S3. Obtain the current funding file data to be tested, and use the set thresholds and corresponding patterns to review and analyze whether the file to be tested meets the funding requirements and whether the family situation type matches the type of funding applied for. S4. By regularly sampling and testing students' student card usage and work-study activities, we can analyze whether there are any missed recipients among students who have not submitted financial aid applications, and at the same time, we can analyze whether there are any ineligible recipients based on their living standards.

[0020] In step S1: after authorization, basic student information and family situation information are obtained from the student financial aid management system and the student status management system through file import and database connection; historical financial aid record information of students is obtained from the student-origin civil affairs department database through API interface call; the basic student information includes student name and student number; the family situation information includes the number of family laborers m and the family poverty certification document; the historical financial aid records include historical financial aid records and student loan application status; and the student card usage frequency of the student is obtained in real time from the student card transaction system, and the work-study registration record of the student is obtained periodically from the student work-study record.

[0021] In step S2: the acquired data is classified according to the set information classification standard, the information acquired from different databases is divided into basic student information, family situation information and historical financial aid record information, and the acquired data is classified and stored according to the classification result; during data storage, a storage method combining a relational database and a non-relational database is adopted for data storage, the relational database is used to store the basic student information, and the non-relational database is used to store the family situation information and the historical financial aid record information; indexes are established for the data in the two databases for querying specific student data; in the relational database, primary key indexes are established for the three unique identification fields of student name, ID number and student number, so as to locate the data of a specific student, and common indexes are established for the two types of data of the number of family laborers and the family poverty certification; The Apriori algorithm is used to analyze the classified and stored historical financial aid archive data, and the corresponding relationship between the student family situation information and the types of financial aid obtained by the students is analyzed; according to the absolute poverty line and the minimum living security line set by the civil affairs department, the family situations are divided into absolute poor families, minimum living security poor families and relatively poor families according to the poverty certification; the number of family laborers is divided into three categories, namely m≤m1, m1<m≤m2, and m>m2; and the historical financial aid types are divided into national grants, provincial grants and campus grants; the Apriori algorithm is used to analyze historical data, the financial aid types corresponding to different categories of family situations are analyzed, and the corresponding rules between family situation types and financial aid types are obtained, and the corresponding rules are as follows: Absolute poor family, number of family laborers m≤m1 → obtaining national grant; Minimum living security poor family, number of family laborers m1<m≤m2 → obtaining provincial grant; Relatively poor family, number of family laborers m>m2 → obtaining campus grant.

[0022] In step S3: acquire the current grant file data to be detected, and judge whether the submitted grant file data meets the grant requirements; set a threshold for judging student grant files that do not meet the grant requirements, and set the maximum threshold for the number of family working members as m4; if it is detected that the condition of m≥m4 exists in the student grant file to be detected, it is judged that the student grant file does not meet the grant requirements, then an abnormal prompt indicating that there is a student grant file that does not meet the grant requirements is issued, and the student grant file is marked as failed in review; if it is detected that all data analysis results in the student grant file to be detected satisfy m<m4, it is judged that the student grant file meets the grant requirements, and the student grant file is marked as passed in review; for the student grant files marked as passed in review, further review and analysis are performed on the files to be detected in combination with the summarized corresponding rule between family situation types and grant types, and an abnormal prompt of mismatched grant situations is issued for the situation where the family situation type does not conform to the applied grant type; this greatly saves the manpower consumption for reviewing grant files, improves the review efficiency, and increases the matching probability between students' family situations and grant types according to the summarized rule, so as to ensure the balance of grant resource allocation.

[0023] In step S4: after the analysis and review of the files of students who have applied for grants is completed, perform sampling detection on the student card usage and work-study situation of the students according to the sampling frequency q; when performing sampling detection on the student card usage and work-study situation of all students, a single sampling period is k days; acquire the usage times L of the student card within k days from the student card transaction system; and acquire the daily work-study duration of each student {Δt1, Δt2, …, Δt k}; set the maximum daily work-study duration threshold Δt max and the minimum daily work-study duration threshold Δt min ; perform comparative analysis between the sampled data and the set thresholds; set the days where Δt i ≥Δt max is recorded as k2, where i=1, 2, …, k; calculate the student's life hardship index according to the following formula: ; where W represents the life hardship index; the sampling frequency q for sampling and detecting the student's student card usage and work-study situation is a dynamically changing value; the sampling frequency q is calculated according to the following formula: ; where q0 represents the base sampling frequency; set the life hardship index threshold W max , analyze the obtained life hardship index, such Figure 2As shown, the analysis results are as follows: If W <W max This indicates that the frequency of students' student card usage and the duration of their work-study programs were normal during this sampling period, and it was determined that no students experienced any sudden financial hardship. If W≥W max If the frequency of student card use and the duration of work-study are too high during the current sampling period, it is determined that the student is experiencing a sudden financial hardship. In this case, the student's student ID is extracted and the student is marked as a key focus. For students marked as key focus, the sampling period for which they are marked as key focus is recorded as the first sampling period, and the living hardship index of key focus students is analyzed in subsequent sampling tests.

[0024] In the subsequent sampling and detection, the observation period begins from the sampling period designated as the first sampling. This triggers a focused observation mode for key observation objects. The initial total number of samples in this focused observation mode is set to c, which is a dynamically changing value. The W value obtained from each sampling and detection is analyzed, and the obtained W value is compared with the set threshold W. max A comparison is performed, and the analysis results of each of the c′ samplings are compared. If the analysis results of c′ consecutive samplings are all W≥W, the comparison is successful. max If the result of 2 consecutive c′ comparisons is W≥W, then subtract 1 from the value of c; max If the key observation mode is lifted, the key observation target will be identified as the target of funding; if the c′ consecutive comparative analysis results are all W <W max If the result is W, then increment the value of c by 1; if 2c′ consecutive comparisons are both W, then increment the value of c by 1. <W max If so, the key observation mode will be directly lifted, and the key observation target will be judged as an entity that does not need funding; After the number of samplings for the key observation mode reaches c and the analysis is completed, the sampling results from the c sampling tests are defined as W≥W. max The number of times is denoted as c1. By analyzing the value of c1, we determine whether the key focus group is in need of funding. The analysis results are as follows: If c1 < 2c′, it is determined that the target of this focus is not the target of funding. If c1≥2c′, the student in question is identified as needing financial aid. Further, the student's basic information is extracted and compared with the student's basic information after the financial aid file has been reviewed. If the student has submitted a financial aid application and the application has been approved, no additional notification is issued. If the student has not submitted a financial aid application, a notification is issued indicating that a student has been overlooked.

[0025] When sampling and detecting the usage of student cards and work-study situations of students, for students who have been judged to meet the funding requirements, an analysis of the abnormal living standard evaluation value is added, and Δt i <Δt min The number of days of is recorded as k3, and the abnormal living standard evaluation value is calculated according to the following formula: ; Wherein, M is the abnormal living standard evaluation value; Set the abnormal living standard evaluation value threshold M0; compare and analyze the calculated abnormal living standard evaluation value with the set threshold, and the analysis results are as follows: If M<M0, it indicates that the student does not have a living standard exceeding the family's affordability; If M≥M0, it indicates that the student has a living standard exceeding the family's affordability. In the next two sampling detections of the student's student card usage and work-study situation, the abnormal living standard evaluation value of this student will be analyzed emphatically. If it is judged for three consecutive times that the student has a living standard exceeding the family's affordability, an abnormality prompt for student funding file review will be issued, prompting the staff to verify the authenticity of the information in the student's funding file, and re-review the student's funding file.

[0026] A multi-dimensional data collection and analysis system based on artificial intelligence, as shown in Figure 3 , the system includes a multi-dimensional data collection module, a data analysis and storage module, a funding file analysis and review module, and an abnormality identification and feedback module; The multi-dimensional data collection module obtains student basic information, family situation and historical funding records from different data sources by means of file import and API interface call; The data analysis and storage module classifies and stores the collected historical student funding file data; uses artificial intelligence algorithms to analyze the classified and stored historical funding file data to obtain the corresponding rule between family situations and funding types; The funding file analysis and review module obtains the current funding file data to be detected, and uses the set threshold and the obtained corresponding rule to review and analyze whether the file to be detected meets the funding requirements and whether the family situation type matches the applied funding type; The abnormality identification and feedback module regularly samples and detects the usage of student cards and work-study situations of students, analyzes whether there are omitted funding objects among students who have not submitted funding applications, and at the same time analyzes whether there are unqualified funding objects according to the living standard.

[0027] The multi-dimensional data collection module includes a student personal and family information collection unit, a historical financial aid record collection unit, and a student card and work-study information collection unit. The student personal and family information collection unit obtains basic student information and family information from the student financial aid management system and the student registration system through file import and database connection. The historical financial aid record collection unit obtains students' historical financial aid record information from the database of the civil affairs department of the student's place of origin through API interface calls. The student card and work-study information collection unit obtains students' student card usage information in real time from the student card transaction system through API interface, and periodically obtains students' work-study registration records from the student work-study record system.

[0028] The data analysis and storage module includes a data classification and storage unit and a data analysis unit. The data classification and storage unit categorizes the acquired data according to set information classification standards, dividing information from different databases into student basic information, family information, and historical financial aid records. The data is then stored according to these categories. A combination of relational and non-relational databases is used for data storage. Relational databases are used to store the highly structured student basic information, while non-relational databases are used to store family information and historical financial aid records. The data analysis unit uses the Apriori algorithm to analyze the categorized historical financial aid records, identifying the correspondence between student family information and the types of financial aid received by students.

[0029] The funding file analysis and review module uses thresholds to determine which student funding files do not meet the funding requirements. Files deemed ineligible are marked as approved. For files marked as approved, the module further analyzes and reviews them based on the observed patterns between family circumstances and funding types. If the family circumstances do not match the applied funding type, an error message indicating a mismatch in funding information is issued.

[0030] The anomaly detection and feedback module includes a unit for detecting missing recipients of financial aid and a unit for detecting abnormal recipients of financial aid. The unit for detecting missing recipients of financial aid conducts regular sampling and testing of all students' student card usage and work-study activities, and introduces a living hardship index to analyze students' living conditions and determine whether there are any missing recipients among students who have not submitted financial aid applications. The unit for detecting abnormal recipients of financial aid introduces an abnormal living standard assessment value to evaluate the living standards of students who have passed the financial aid review, and verifies the information of students whose living standards consistently exceed their families' affordability and re-examines their financial aid files.

[0031] Embodiment 1: In step 2, the Apriori algorithm is used to analyze the historically categorized and stored financial aid archive data, so as to analyze the correspondence between students' family situation information and the types of financial aid obtained by the students; the absolute poverty line and the minimum subsistence allowance line set by the civil affairs department are used to divide the family situations into absolutely poor families, minimum subsistence allowance poor families and relatively poor families according to the poverty certificates; the number of family laborers is divided into three categories, namely m≤1, 1<m≤2, and m>2; and the historical financial aid types are divided into national grant, provincial grant and campus grant; then, the Apriori algorithm is used to analyze the historical data, to figure out the financial aid types corresponding to different categories of family situations, and obtain the correspondence rule between family situation types and financial aid types, which is as follows: Absolutely poor family, number of family laborers m≤1 → obtaining the national grant; Minimum subsistence allowance poor family, number of family laborers 1<m≤2 → obtaining the provincial grant; Relatively poor family, number of family laborers m>2 → obtaining the campus grant.

[0032] In step S3: obtain the to-be-tested financial aid archive data, where the number of family laborers is m=1, and the applied financial aid type is the national grant; judge whether the submitted financial aid archive data meets the financial aid requirements; set a threshold for judging the student financial aid archives that do not meet the financial aid requirements, and set the maximum threshold for the number of family laborers as m4=4; then the data analysis result shows that m<m4, it is judged that the student's financial aid archive meets the financial aid requirements, and the student's financial aid archive is marked as approved; combined with the summarized correspondence rule between family situation types and financial aid types, further review and analysis are performed on the to-be-tested archive, and the analysis result is: absolutely poor family, number of family laborers m≤1 → obtaining the national grant; the family situation type is consistent with the applied financial aid type.

[0033] In step S4: after the archive analysis and review for the students who have applied for financial aid is completed, sampling and detection are performed on the student card usage situation and work-study situation of the students according to the sampling frequency q; when performing sampling and detection on the student card usage situation and work-study situation of all students, a single sampling period is k=10 days; the number of uses of the student card within k days obtained from the student card transaction system is L=25; and the daily work-study duration of each student obtained from the work-study records is {5, 4, 6, 6, 6, 7, 6, 4, 6, 6}; the maximum threshold of daily work-study duration Δt is set max =6h and the minimum threshold of daily work-study duration Δt min =1h; let Δt i ≥Δt max The number of days that satisfy the above condition is recorded as k2=7, and the student's life hardship index is calculated according to the following formula: ; Where W represents the hardship index; W=1.53; The sampling frequency q for monitoring students' student card usage and work-study activities is a dynamically changing value; the sampling frequency q is calculated using the following formula: ; Where q0 = 1 time / month represents the basic sampling frequency; q = 26 times; Set a threshold W for the life hardship index. max =1.2. Analysis of the obtained living difficulties index yields the following results: W=1.53≥W max If the frequency of student card use and the duration of work-study are too high during the current sampling period, it is judged that the student is experiencing sudden financial hardship and there may be changes in the family situation. In this case, the student's student ID is extracted and marked as a key focus. For students marked as key focus, the sampling period marked as key focus is recorded as the first sampling period, and the living hardship index of key focus students is analyzed in the next two sampling tests. In the subsequent sampling and detection, the observation period begins from the sampling period designated as the first sampling, triggering a key observation mode for the key observation objects. The initial total number of samples in the key observation mode is set to c=10, where c is a dynamically changing value. The W value obtained from each sampling and detection is analyzed, and the obtained W value is compared with the set threshold W. max A comparison is performed, and for each sample c′=3 times, the analysis results of the 3 samples are compared and analyzed. If the results of 3 consecutive comparative analyses are all W≥W, the comparison is considered complete. max If the result of six consecutive comparisons is W≥W, then subtract 1 from the value of c; max If the key observation mode is lifted, the key observation target will be identified as the target of funding; if the results of three consecutive comparative analyses are all W... <W max If the result is W, then increment the value of c by 1; if the result of 6 consecutive comparisons is W, then increment the value of c by 1. <W max If so, the key observation mode will be directly lifted, and the key observation target will be judged as an entity that does not need funding; After the number of samplings for the key observation mode reaches c and the analysis is completed, the sampling results from the c sampling tests are defined as W≥W. max The number of times is denoted as c1. By analyzing the value of c1, we determine whether the key focus group is in need of funding. The analysis results are as follows: If c1 < 2c′, it is determined that the target of this focus is not the target of funding. If c1≥2c', it is determined that the key attention object is an object in need of funding; further, the basic student information of the key attention object is extracted and compared with the basic student information that has been audited through the funding file. If the student has submitted a funding application file and the audit is passed, no additional prompt will be issued; if the student has not submitted a funding application file, a prompt indicating that there is a missing funding object will be issued.

[0034] When sampling and detecting the usage of student cards and work-study situations of students, for students who have been judged to meet the funding requirements, an analysis of the abnormal living standard evaluation value is added, and Δt i <Δt min the number of days is recorded as k3=2, and the abnormal living standard evaluation value is calculated according to the following formula: the abnormal consumption evaluation value is calculated according to the following formula: ; where M is the abnormal living standard evaluation value; M=1.4; The abnormal living standard evaluation value threshold M0=2.1 is set; the calculated abnormal living standard evaluation value is compared and analyzed with the set threshold, and the analysis results are as follows: If M<M0, it indicates that the student does not have a living standard exceeding the affordability of the family.

[0035] It is apparent to those skilled in the art that the present invention is not limited to the details of the above-described exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention. Therefore, the embodiments are to be regarded as illustrative in all respects, and not restrictive. The scope of the present invention is defined by the appended claims rather than the foregoing description, so that all changes that come within the meaning and range of equivalency of the claims are intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.

Claims

1. A multi-dimensional data acquisition and analysis method based on artificial intelligence, characterized in that: comprising the following steps: S1. Obtain basic student information, family status and historical subsidy records from different data sources through file import and API interface invocation; S2. Classify and store the collected historical student subsidy archive data; analyze the classified and stored historical subsidy archive data by using an artificial intelligence algorithm to obtain the corresponding rule between family status and subsidy type; S3. Acquire the current to-be-detected subsidy archive data, and perform review and analysis on whether the to-be-detected archive meets the subsidy requirements and whether the family status type matches the applied subsidy type by using the set threshold and the obtained corresponding rule; S4. Regularly perform sampling detection on the student card usage status and work-study status of students, analyze whether there are missed subsidy recipients among students who have not submitted subsidy applications, and simultaneously analyze whether there are non-compliant subsidy recipients according to living standards.

2. The multi-dimensional data acquisition and analysis method based on artificial intelligence according to claim 1, characterized in that: In step S1: after authorization, obtain basic student information and family status information from the student subsidy management system and the student status management system through file import and database connection; obtain historical subsidy record information of students from the database through API interface invocation; the basic student information includes student name and student number; the family status information includes the number of family working members m and the family poverty certification document; the historical subsidy records include historical subsidy records and student loan application status; and obtain the student card usage frequency of students in real time from the student card transaction system through the API interface, and regularly obtain the work-study registration records of students from the student work-study records.

3. The multi-dimensional data acquisition and analysis method based on artificial intelligence according to claim 1, characterized in that: In step S2: classify the acquired data according to the set information classification standard, divide the information acquired from different databases into basic student information, family status information and historical subsidy record information, and classify and store the acquired data according to the classification result; when storing data, adopt a storage mode combining a relational database and a non-relational database for data storage, use the relational database to store basic student information, and use the non-relational database to store family status information and historical subsidy record information; establish indexes for the data in the two databases for querying specific student data; Analyze the classified and stored historical subsidy archive data by using the Apriori algorithm, and analyze the corresponding relationship between the student family status information and the subsidy type and subsidy obtained by the student; divide the family status into absolute poor family, subsistence allowance poor family and relatively poor family according to the poverty certification by using the set absolute poverty line and subsistence allowance line; divide the number of family working members into three categories, namely m≤m1, m1<m≤m2, and m>m2; and divide the historical subsidy types into national grant, provincial grant and campus grant; analyze the historical data by using the Apriori algorithm, figure out the subsidy types corresponding to different categories of family status, and obtain the corresponding rule between the family status type and the subsidy type, where the corresponding rule is as follows: Absolute poor family, number of family working members m≤m1→obtain national grant; Subsistence allowance poor family, number of family working members m1<m≤m2→obtain provincial grant; Relatively poor families where the number of family laborers m is greater than m2 → obtain the on-campus financial aid grant.

4. The multi-dimensional data acquisition and analysis method based on artificial intelligence according to claim 3, characterized in that: In step S3: acquire the current to-be-detected financial aid file data, and judge whether the submitted financial aid file data meets the financial aid requirements; set a threshold for judging the student financial aid files that do not meet the financial aid requirements, and set the maximum threshold for the number of family laborers as m4; if it is detected that the condition of m ≥ m4 exists in the to-be-detected student financial aid file, it is determined that the student financial aid file does not meet the financial aid requirements, an abnormality prompt indicating that there is a student financial aid file that does not meet the financial aid requirements is issued, and the student financial aid file is marked as not approved for review; if it is detected that all data analysis results in the to-be-detected student financial aid file satisfy m < m4, it is determined that the student financial aid file meets the financial aid requirements, and the student financial aid file is marked as approved for review; for the student financial aid file marked as approved for review, further review and analysis are performed on the to-be-detected file in combination with the summarized correspondence rule between family situation types and financial aid types, and an abnormality prompt of mismatched financial aid situation is issued for the case where the family situation type does not match the applied financial aid type.

5. The multi-dimensional data acquisition and analysis method based on artificial intelligence according to claim 1, characterized in that: In step S4: After the file analysis and review of students who have applied for funding are completed, the student card usage and work-study activities of students are sampled and tested according to the sampling frequency q; when sampling and testing all students, the single sampling period is k days; the number of times the student card is used L within k days is obtained from the student card transaction system; and the daily work-study duration {Δt1, Δt2, ..., Δt} of each student is obtained from the work-study records. k Set the maximum daily work-study time threshold Δt. max and the minimum threshold Δt for daily work-study hours min ; will Δt i ≥Δt max The number of days is denoted as k2, where i = 1, 2, ..., k; the student's living hardship index is calculated using the following formula: ； Wherein, W represents the living difficulty index; The sampling frequency q for sampling and detecting the usage status of the student's student card and the work-study status is a dynamically changing value; the sampling frequency q is calculated according to the following formula: ； Wherein, q0 represents the basic sampling frequency; Set a threshold W for the life hardship index. max The obtained living hardship index was analyzed, and the results are as follows: If W <W max This indicates that the frequency of students' student card usage and the duration of their work-study programs were normal during this sampling period, and it was determined that no students experienced any sudden financial hardship. If W≥W max If the frequency of student card use and the duration of work-study are too high during the current sampling period, it is determined that the student is experiencing a sudden financial hardship. In this case, the student's student ID is extracted and the student is marked as a key focus. For students marked as key focus, the sampling period for which they are marked as key focus is recorded as the first sampling period, and the living hardship index of key focus students is analyzed in subsequent sampling tests.

6. The multi-dimensional data acquisition and analysis method based on artificial intelligence according to claim 5, characterized in that: In the subsequent sampling and detection, the observation period begins from the sampling period designated as the first sampling. This triggers a focused observation mode for key observation objects. The initial total number of samples in this focused observation mode is set to c, which is a dynamically changing value. The W value obtained from each sampling and detection is analyzed, and the obtained W value is compared with the set threshold W. max A comparison is performed, and the analysis results of each of the c′ samplings are compared. If the analysis results of c′ consecutive samplings are all W≥W, the comparison is successful. max If the result of 2 consecutive c′ comparisons is W≥W, then subtract 1 from the value of c; max If the key observation mode is lifted in advance, the key observation target will be identified as the target of funding; if the c′ consecutive comparative analysis results are all W <W max If the result is W, then increment the value of c by 1; if 2c′ consecutive comparisons are both W, then increment the value of c by 1. <W max If so, the key observation mode will be lifted in advance, and the key observation target will be judged as someone who does not need funding; For cases where the key observation mode is not terminated in advance, after the number of samplings under the key observation mode reaches c and the analysis is completed, the sampling results from the c sampling tests will be defined as W≥W. max The number of times is denoted as c1. By analyzing the value of c1, we determine whether the key focus group is in need of funding. The analysis results are as follows: If c1 < 2c', it is determined that the key concern object is not an object in need of financial aid; If c1 ≥ 2c', it is determined that the key concern object is an object in need of financial aid; further extract the basic student information of the key concern object and compare it with the basic student information of students whose financial aid files have been reviewed, if this student has submitted a financial aid application file and the application is approved, no additional prompt will be issued; if this student has not submitted a financial aid application file, a prompt indicating that there is a missed financial aid recipient will be issued.

7. The multi-dimensional data acquisition and analysis system based on artificial intelligence according to claim 6, characterized in that: When sampling and testing students' student card usage and work-study activities, for students already determined to meet the eligibility requirements for financial aid, an additional analysis of abnormal living standard assessment values ​​will be added, including Δt. i <Δt min The number of days is denoted as k3. The abnormal standard of living assessment value is calculated according to the following formula: ； Wherein, M is the abnormal living standard assessment value; Set the abnormal living standard assessment value threshold M0; perform comparative analysis on the calculated abnormal living standard assessment value and the set threshold, and the analysis results are as follows: If M < M0, it means that this student does not have a living standard exceeding the family's affordability; If M ≥ M0, it means that this student has a living standard exceeding the family's affordability, then the abnormal living standard assessment value of this student will be analyzed emphatically in the next two sampling detections of the student card usage status and work-study status. If it is judged for three consecutive times that this student has a living standard exceeding the family's affordability, an abnormality prompt for student financial aid file review will be issued, prompting the staff to verify the authenticity of the information in this student's financial aid file, and re-review this student's financial aid file.

8. A multi-dimensional data acquisition and analysis system based on artificial intelligence, characterized in that: The system includes a multi-dimensional data acquisition module, a data analysis and storage module, a financial aid file analysis and review module, and an abnormality identification and feedback module; The multi-dimensional data acquisition module acquires basic student information, family situations and historical financial aid records from different data sources by means of file import and API interface invocation; The data analysis and storage module will collect historical student financial aid records, classify and store them; and use artificial intelligence algorithms to analyze the classified and stored historical financial aid records to obtain the correspondence between family circumstances and financial aid types. The funding file analysis and review module obtains the funding file data to be tested, and uses the set thresholds and corresponding patterns to review and analyze whether the file to be tested meets the funding requirements and whether the family situation type matches the type of funding applied for. The anomaly detection and feedback module periodically samples and detects students' student card usage and work-study activities to analyze whether there are any missed recipients among students who have not submitted funding applications, and at the same time analyzes whether there are any ineligible recipients based on their living standards.

9. The multi-dimensional data acquisition and analysis system based on artificial intelligence according to claim 8, characterized in that: The multi-dimensional data collection module includes a student personal and family information collection unit, a historical financial aid record collection unit, and a student card and work-study status collection unit. The student personal and family information collection unit obtains basic student information and family information from the student financial aid management system and the student registration system through file import and database connection. The historical financial aid record collection unit obtains students' historical financial aid record information from the database of the civil affairs department of the student's place of origin through API interface calls. The student card and work-study status collection unit obtains students' student card usage information in real time from the student card transaction system through API interface, and periodically obtains students' work-study registration records from the student work-study record system. The data analysis and storage module includes a data classification and storage unit and a data analysis unit. The data classification and storage unit classifies the acquired data according to the set information classification criteria, dividing the information obtained from different databases into student basic information, family information, and historical financial aid records, and stores the acquired data according to the classification results. The data storage method combines relational and non-relational databases, using relational databases to store the highly structured student basic information and non-relational databases to store family information and historical financial aid records. The data analysis unit uses the Apriori algorithm to analyze the classified and stored historical financial aid records, analyzing the correspondence between student family information and the types of financial aid received by students.

10. The multi-dimensional data acquisition and analysis system based on artificial intelligence according to claim 8, characterized in that: The funding file analysis and review module sets a threshold to determine student funding files that do not meet the funding requirements, and marks the student funding files that do not meet the funding requirements as approved. For the student funding files marked as approved, the module further reviews and analyzes the files to be tested based on the summarized correspondence between family situation types and funding types, and issues an abnormal prompt of funding mismatch for cases where the family situation type does not match the applied funding type. The anomaly identification and feedback module includes a missing aid recipient detection unit and an abnormal aid recipient detection unit. The missing aid recipient detection unit conducts regular sampling tests by monitoring the student card usage and work-study activities of all students, and introduces a living hardship index to analyze the students' living conditions to determine whether there are any missing aid recipients among students who have not submitted aid applications. The abnormal aid recipient detection unit introduces an abnormal living standard assessment value to evaluate the living standards of students who have passed the aid review, and verifies the information of students whose living standards continuously exceed their family's affordability and re-examines their aid files.

Images