Economical difficulty student assessment method and system based on data mining

Abstract

The invention discloses an economic difficulty student assessment method and system based on data mining, and relates to the technical field of economic difficulty student assessment. The economic difficulty student assessment method based on data mining comprises the following steps: acquiring consumption behavior time sequence data and source cost constraint data of a plurality of students to be assessed in a set class; based on a pre-trained consumption evolution mapping model and in combination with consumption behavior time sequence data, analyzing consumption comprehensive limited characteristic values of corresponding to-be-evaluated students; according to the method, each student to be assessed is subjected to economic difficulty student assessment processing based on the comprehensive difficulty assessment feature value, the student-seeing support demand feature value of the corresponding student to be assessed is analyzed on the basis of the biogenic cost constraint data, and the comprehensive difficulty assessment feature value of the corresponding student to be assessed is analyzed in combination with the consumption comprehensive limited feature value. Therefore, the actual economic difficulty condition of the student is comprehensively reflected, the accuracy of the evaluation result is improved, and the fairness of the evaluation of the economic difficulty student is effectively improved.

Description

Technical Field

[0001] This invention relates to the field of economically disadvantaged student assessment technology, specifically to an economically disadvantaged student assessment method and system based on data mining. Background Technology

[0002] In the current education assistance system, in order to ensure that students with financial difficulties receive appropriate assistance, traditional methods for assessing students with financial difficulties mainly rely on direct data such as family income and family assets. However, these methods are difficult to fully reflect the actual financial difficulties of students. Furthermore, due to the strong subjectivity in the assessment of students with financial difficulties, they are easily affected by factors such as inaccurate data and inconsistent assessment standards, leading to unfairness in the assistance provided to students with financial difficulties. However, with the rapid development of big data and data mining technologies, more and more fields are beginning to try to apply these technologies to the assessment of students with financial difficulties in order to more accurately identify and assess students' financial difficulties.

[0003] Existing technologies, such as the patent application with publication number CN113902055A, disclose a method and system for certifying college students based on multi-dimensional evaluation. This method includes: acquiring information about students to be certified; inputting the acquired information into a student certification prediction model to obtain certification results; classifying students to be certified according to the certification results to complete the certification process. This invention collects more comprehensive information about those to be certified, including not only the applicant's personal consumption, learning attitude, daily behavior, work-study activities, and student loan applications, but also the applicant's basic family situation, making the aid system more accurate and the assessment more comprehensive.

[0004] Based on the above findings, the limitations of existing technologies include at least the following problems: Existing technologies are prone to insufficient information completeness and authenticity risks during the evaluation process; they struggle to effectively identify hidden economic difficulties caused by objective factors such as sudden family changes, resulting in unreliable evaluation results; and they are unable to dynamically capture and continuously track and analyze students' on-campus consumption behavior over time, easily overlooking changes in students' daily consumption. This is especially true for students on the verge of economic hardship, whose economic difficulties may not depend on short-term changes in family income, but rather on long-term accumulated consumption constraints and life pressures. Existing technologies cannot perform timely time-series analysis, easily leading to evaluation results that deviate from reality. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a data mining-based method and system for assessing economically disadvantaged students, solving the problem that existing technologies, lacking dynamic time-series analysis capabilities, struggle to accurately identify students' economic hardship status.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a data mining-based method for assessing economically disadvantaged students, comprising the following steps: acquiring time-series data of consumption behavior and cost constraints of the student's place of origin for several students to be assessed within a designated class; analyzing the comprehensive consumption constraint characteristic value of each student based on a pre-trained consumption evolution mapping model and the time-series data of consumption behavior of each student to be assessed within the designated class; analyzing the educational support need characteristic value of each student based on the cost constraints of the student's place of origin within the designated class, and analyzing the comprehensive hardship assessment characteristic value of each student based on the comprehensive hardship assessment characteristic value; and conducting economically disadvantaged student assessment processing for each student to be assessed within the designated class based on the comprehensive hardship assessment characteristic value.

[0007] Furthermore, the consumption behavior time-series data includes the number of times dining in the canteen, the amount spent in the canteen, the number of times shopping in the supermarket, the amount spent in the supermarket, the number of times receiving express deliveries, the expenditure on personal care products, the consumption time constraint, the consumption trajectory expansion, the low-price dependence, and the micro-expenditure activity value for each time period. The consumption evolution mapping model includes a consumption input layer, a consumption evolution layer, a correlation mapping layer, and an output layer.

[0008] Furthermore, the specific steps for analyzing and setting the comprehensive consumption restriction characteristic value for each student to be evaluated in the class are as follows: Input the time series data of the consumption behavior of each student to be evaluated in the class into the pre-trained consumption evolution mapping model, and analyze the corresponding consumption pattern mapping characteristic set of the student to be evaluated, including the consumption aggregation intensity characteristic value, the low-price expenditure level tendency characteristic value, and the consumption diffusion amplitude characteristic value; Based on the consumption pattern mapping characteristic set of each student to be evaluated in the class, analyze the corresponding comprehensive consumption restriction characteristic value of the student to be evaluated.

[0009] Furthermore, the specific steps for analyzing the consumption pattern mapping feature set of each student to be evaluated within the set class are as follows: In the consumption input layer of the consumption evolution mapping model, the consumption behavior time series data of each student to be evaluated within the set class is received and preprocessed; in the consumption evolution layer of the consumption evolution mapping model, the preprocessed consumption behavior time series data of each student to be evaluated within the set class is subjected to time series analysis to extract the corresponding consumption evolution feature vector of the student to be evaluated; in the association mapping layer of the consumption evolution mapping model, the consumption evolution feature vector of each student to be evaluated within the set class is fused and mapped to extract the corresponding consumption state feature vector of the student to be evaluated; in the output layer of the consumption evolution mapping model, based on the consumption state feature vector of each student to be evaluated within the set class, the corresponding consumption pattern mapping feature set of the student to be evaluated is output.

[0010] Furthermore, the source cost constraint data includes residents' Engel coefficient, urban-rural income ratio, employment rate, CPI year-on-year change rate, minimum living allowance, industrial transformation intensity, and labor productivity. The specific steps for analyzing the educational support needs characteristic values ​​of the students to be evaluated are as follows: Based on the source cost constraint data of each student to be evaluated in the set class, analyze the corresponding source assessment characteristic set of the students to be evaluated, including the source prosperity characteristic value and the source living pressure characteristic value; obtain the family burden characteristic value of each student to be evaluated in the set class, and combine it with the source assessment characteristic set to analyze the corresponding educational support needs characteristic value of the students to be evaluated.

[0011] Furthermore, the specific steps for analyzing and setting the evaluation characteristic set of each student's place of origin within the class are as follows: Based on the urban-rural income ratio, employment rate, industrial transformation intensity, and labor productivity of each student in the class, analyze the corresponding prosperity characteristic value of their place of origin; based on the Engel coefficient, minimum living allowance, and CPI year-on-year change rate of each student in the class, analyze the corresponding living pressure characteristic value of their place of origin.

[0012] Furthermore, the specific steps for obtaining the family burden characteristic value of each student to be assessed in the designated class are as follows: obtain the family status data of each student to be assessed in the designated class and perform normalization processing; based on the normalized family status data of each student to be assessed in the designated class, analyze the family burden characteristic value of the students to be assessed.

[0013] Furthermore, the specific formula for calculating the comprehensive difficulty assessment characteristic value of a student to be assessed within a given class is as follows: ;in, , , The criteria are as follows: setting comprehensive hardship assessment characteristic values, comprehensive consumption restriction characteristic values, and educational support need characteristic values ​​for a specific student in the class to be assessed. , , The coefficients stored in the database are, in order: consumption restriction adjustment coefficient, education support adjustment coefficient, and interaction adjustment coefficient. .

[0014] Furthermore, the specific steps for assessing economic hardship for each student in a designated class based on comprehensive hardship assessment feature values ​​are as follows: the comprehensive hardship assessment feature values ​​of each student in a designated class are compared with preset comprehensive hardship assessment feature thresholds; and economic hardship is identified for each student in a designated class based on the results of the comparison.

[0015] A data mining-based assessment system for economically disadvantaged students includes: a data acquisition module for acquiring time-series data on consumption behavior and cost constraints from the student's place of origin for several students to be assessed within a designated class; a consumption constraint analysis module for analyzing the comprehensive consumption constraint characteristic value of each student to be assessed based on a pre-trained consumption evolution mapping model and the time-series data on consumption behavior; a comprehensive hardship assessment module for analyzing the educational support need characteristic value of each student to be assessed based on the cost constraints from the student's place of origin, and combining the comprehensive consumption constraint characteristic value to analyze the comprehensive hardship assessment characteristic value; and an economically disadvantaged student assessment feedback module for conducting economically disadvantaged student assessment processing for each student to be assessed within the designated class based on the comprehensive hardship assessment characteristic value.

[0016] The present invention has the following beneficial effects:

[0017] (1) This data mining-based assessment method for economically disadvantaged students introduces time-series data of consumption behavior and cost constraints of the student's place of origin, and performs time-series processing on the consumption behavior time-series data based on a pre-trained consumption evolution mapping model to extract a set of consumption pattern mapping features. This allows for in-depth analysis of the evolution of students' consumption behavior, captures changes in their consumption, and provides a dynamic assessment of economic hardship. Furthermore, the method deeply mines the cost constraints of the student's place of origin to obtain the characteristic values ​​of the student's need for educational support, forming a comprehensive hardship assessment characteristic value. This comprehensively reflects the actual economic hardship of students, thereby improving the accuracy of the assessment results and effectively enhancing the fairness of the assessment of economically disadvantaged students.

[0018] (2) This data mining-based assessment method for economically disadvantaged students introduces a pre-trained consumption evolution mapping model and uses it to perform in-depth analysis of consumption behavior time series data, thereby extracting features related to the evolution of students' consumption behavior, capturing changes in students' consumption in different time periods, and on this basis, integrating and generating comprehensive consumption restriction feature values, so as to accurately reflect the economic difficulties of students in real life, thereby reducing the bias of human intervention, and providing a fairer identification of economically disadvantaged students, thus effectively improving the accuracy of the assessment of economically disadvantaged students.

[0019] (3) This data mining-based assessment method for economically disadvantaged students generates characteristic values ​​of their educational support needs by conducting in-depth analysis of the cost constraints of the students' place of origin. This provides more accurate background support for the identification of economically disadvantaged students, reveals the multiple economic pressures that students may face in the process of pursuing their studies, and can dynamically reflect the degree of economic hardship in the students' place of origin and its impact on their lives. This allows the assessment of economically disadvantaged students to be more comprehensive through the economic and family conditions of the students' place of origin, thereby accurately analyzing each student's actual educational support needs and improving the reliability of the economic hardship results.

[0020] (4) This data mining-based assessment system for economically disadvantaged students improves the objectivity of assessment results through collaborative analysis between modules. The data collection module can accurately collect the time-series data of consumption behavior and the cost constraint data of the student's place of origin for each student to be assessed in the set class, ensuring the comprehensiveness of the assessment data. The consumption restriction analysis module combines the pre-trained consumption evolution mapping model to automatically extract features related to students' consumption behavior, comprehensively analyze the students' consumption restriction status, and improve the reliability of the assessment of economically disadvantaged students. The comprehensive difficulty assessment module combines the cost constraint data of the student's place of origin to extract the feature value of the need for school support, and integrates it with the above features to generate the feature value of the comprehensive difficulty assessment, thereby improving the accuracy of the assessment of economically disadvantaged students.

[0021] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0022] Figure 1 This is a flowchart of a data mining-based assessment method for economically disadvantaged students according to the present invention.

[0023] Figure 2 This is a schematic diagram illustrating the data of the consumption pattern mapping feature set of students to be assessed within a class in the data mining-based assessment method for economically disadvantaged students of the present invention.

[0024] Figure 3 This is a flowchart illustrating the specific steps involved in analyzing and setting the consumption pattern mapping feature set for each student to be assessed within a class in a data mining-based assessment method for economically disadvantaged students according to the present invention.

[0025] Figure 4 This is a block diagram of an assessment system for economically disadvantaged students based on data mining, according to the present invention. Detailed Implementation

[0026] Please see Figure 1This invention provides a technical solution: a data mining-based method for assessing economically disadvantaged students, comprising the following steps: acquiring time-series data of consumption behavior and cost constraints of the student's place of origin for several students to be assessed within a set period (e.g., one semester); analyzing the comprehensive consumption constraint characteristic value of each student based on a pre-trained consumption evolution mapping model and the time-series data of consumption behavior of each student to be assessed within the set class; analyzing the educational support need characteristic value of each student based on the cost constraints of the student's place of origin within the set class, and analyzing the comprehensive difficulty assessment characteristic value of each student based on the comprehensive difficulty assessment characteristic value; and performing economically disadvantaged student assessment processing on each student to be assessed within the set class based on the comprehensive difficulty assessment characteristic value.

[0027] The specific formula for calculating the comprehensive difficulty assessment characteristic value of a student to be assessed within a given class is as follows: ;in, To set a comprehensive difficulty assessment characteristic value for a student in the class to be assessed, To set a comprehensive consumption restriction characteristic value for a student in the class to be evaluated, The consumption restriction adjustment coefficient is stored in the database. To define the educational support needs characteristic value for a student in the class who is to be assessed, Adjustment coefficients for school attendance stored in the database. , These are the interaction adjustment coefficients stored in the database.

[0028] It needs to be explained that the consumption-restricted adjustment coefficient stored in the database Education support adjustment coefficient The steps are as follows: Read the comprehensive consumption restriction characteristic value and the educational support need characteristic value for each student to be assessed in the designated class. Extract the mean of the comprehensive consumption restriction characteristic value and the mean of the educational support need characteristic value, sum them up to obtain the economic hardship sum value. Ratio the mean of the comprehensive consumption restriction characteristic value and the mean of the educational support need characteristic value to the economic hardship sum value, and use the corresponding results as the consumption restriction adjustment coefficient. Education support adjustment coefficient .

[0029] Interaction adjustment coefficients stored in the database The steps are as follows: Read the comprehensive consumption limitation characteristic value and the school support need characteristic value for each student to be assessed in the designated class, and analyze the correlation between the two (taking the absolute value) using the Pearson correlation coefficient method as the interaction moderating coefficient. .

[0030] The specific steps for assessing economic hardship for each student in a designated class based on comprehensive hardship assessment characteristic values ​​are as follows: The comprehensive hardship assessment characteristic values ​​of each student in the designated class are compared with a preset comprehensive hardship assessment characteristic threshold. Based on the comparison results, each student in the designated class is identified as economically disadvantaged. Specifically: if the comprehensive hardship assessment characteristic value of each student in the designated class is higher than the preset comprehensive hardship assessment characteristic threshold, then the student is identified as economically disadvantaged; if the comprehensive hardship assessment characteristic value of each student in the designated class is lower than or equal to the preset comprehensive hardship assessment characteristic threshold, then the student is not identified as economically disadvantaged.

[0031] Specifically, the time-series data of consumer behavior includes the number of times dining in the canteen, the amount spent in the canteen, the number of times shopping in the supermarket, the amount spent in the supermarket, the number of times receiving express deliveries, the spending on personal care products, the consumption time constraint, the consumption trajectory expansion, the low-price dependence, and the micro-spending activity for each time period (such as a week). The consumption evolution mapping model includes a consumption input layer, a consumption evolution layer, a correlation mapping layer, and an output layer.

[0032] Among them, the number of times students dine in the canteen is the frequency of their consumption in the campus canteen within a set period (such as one week). This can be obtained by filtering the transaction records of the campus card system, selecting the canteen window as the merchant category, and counting the number of transactions by student ID.

[0033] The cafeteria meal amount is the average expenditure level of students in cafeteria-type consumption venues. It can be obtained by filtering the consumption logs of merchants with cafeteria windows as the merchant category through the transaction records of the campus card system, counting the number of consumptions and the amount of each consumption by student ID, and taking the average as the cafeteria meal amount.

[0034] The supermarket purchase frequency value is the frequency of purchases made by students at the campus supermarket or convenience store within a set period. It can be obtained by filtering transaction records in the transaction log of the campus card system, filtering by merchant category (supermarket or convenience store), counting the number of transactions within the set period by student ID, and obtaining the supermarket purchase frequency value.

[0035] The supermarket spending amount is the average expenditure level of students in a single transaction at the campus supermarket or convenience store. It can be obtained by filtering transaction records in the transaction log of the campus card system that are identified as supermarkets or convenience stores, counting the number of transactions and the corresponding transaction amount within the set period by student ID, and taking the average as the supermarket spending amount.

[0036] The number of times a student picks up a package is a value representing the frequency of their off-campus spending. This value can be obtained by recording and counting each package pickup event through an identity verification record connected to the campus logistics pickup data interface. The package pickup records within the set period are clustered and counted using the student ID as an index, and the result is the number of times a student picks up a package.

[0037] The laundry and personal care expenditure value represents the student's basic living expenses level within a set period. It can be obtained by using the student ID as an index to extract all bathing and washing machine usage billing records stored in the database. The two types of amount fields are summed separately and then the total is obtained. The result is the student's laundry and personal care expenditure value. When this value is low, it indicates that the student spends less on basic living expenses and exhibits a frugal consumption pattern.

[0038] The consumption time constraint value is the degree to which students' consumption behavior is concentrated in a specific time period (such as dining hours in the cafeteria) within a set period. It represents the regularity of their consumption time and the characteristics of their economic expenditure constraints. It can be achieved by reading the transaction time field in the student card consumption log, dividing the 24 hours of each day into several preset time periods (e.g., breakfast period: 6:00–9:00, lunch period: 11:00–13:30, dinner period: 17:00–19:30, and non-meal period: the rest of the time). The number of consumption transactions in each time period is counted by student ID, and the proportion of consumption transactions during meal times to the total number of consumption transactions is used as the consumption time constraint value. When students' consumption is mainly concentrated in fixed meal times and there is almost no consumption behavior in non-meal times, the value is high, indicating that their consumption time is restricted and their economic expenditure elasticity is low.

[0039] The consumption trajectory expansion value is the range of students' consumption activities within a set period. It can divide the campus consumption terminals according to functional areas, establish a correspondence between merchant numbers and functional area numbers, and read all student consumption records in the campus card consumption log to count the number of consumptions in each area, calculate the proportion of consumptions in each area to the total number of consumptions, and analyze the above proportions based on information entropy. The result is used as the consumption trajectory expansion value. When this value is low, it indicates that students' consumption behavior is mainly concentrated in a few fixed areas and their consumption activity range is limited.

[0040] The low-price dependency value is the degree of concentration of students' spending at low-price windows in the cafeteria within a set period. It can be obtained by extracting all students' historical transaction records at each window in the cafeteria from the student card consumption log, calculating the historical average single transaction amount for each window within a preset statistical period (e.g., the most recent month), determining the low-price window threshold based on the historical average single transaction amount (e.g., the historical average single transaction amount is less than or equal to 6 yuan), and establishing a set of low-price window numbers accordingly. Then, it reads all the student's cafeteria transaction records within the set period, identifies the window number corresponding to each transaction, counts the number of transactions belonging to the low-price window number set and compares it with the total number of cafeteria transactions, and uses the result as the low-price dependency value.

[0041] The micro-spending activity value is the frequency of a student's low-amount spending within a set period. It can be determined by extracting all transaction records of the student within the set period from the card consumption log, reading all single-transaction amounts, sorting all single-transaction amounts from largest to smallest, taking the upper limit of the last 40% of the amount sequence as the student's micro-spending threshold, counting the number of transactions with single-transaction amounts below the threshold, calculating the proportion of these transactions to the total number of transactions, and using the result as the micro-spending activity value.

[0042] The specific steps for analyzing the comprehensive consumption restriction characteristic value of each student to be evaluated in the set class are as follows: Input the time series data of the consumption behavior of each student to be evaluated in the set class into the pre-trained consumption evolution mapping model, and analyze the corresponding consumption pattern mapping characteristic set of the student to be evaluated, including the consumption aggregation intensity characteristic value, the low-price expenditure level tendency characteristic value, and the consumption diffusion amplitude characteristic value; Based on the consumption pattern mapping characteristic set of each student to be evaluated in the set class, analyze the corresponding comprehensive consumption restriction characteristic value of the student to be evaluated (used to characterize the student's consumption capacity restriction level within the set period; the larger the value, the greater the economic pressure of the student, the higher the degree of expenditure restriction, the more economical the overall consumption behavior, and the more obvious the tendency of economic hardship).

[0043] The specific formula for calculating the comprehensive consumption restriction characteristic value of a student to be evaluated within a given class is as follows: ;in, To set a comprehensive consumption restriction characteristic value for a student in the class to be evaluated, To define a characteristic value for the consumption clustering intensity of a specific student within a class, The consumption aggregation adjustment coefficient is stored in the database. To set a low-price spending tier propensity characteristic value for a specific student in the class to be evaluated, The low-price expenditure adjustment coefficient is stored in the database. To define a characteristic value for the consumption diffusion range of a specific student to be evaluated within a class, The consumption diffusion moderating coefficient is stored in the database. .

[0044] It needs to be explained that the consumption aggregation adjustment coefficient stored in the database Low-price expenditure adjustment coefficient Consumption diffusion moderating coefficient The acquisition steps are as follows: Read the consumption clustering intensity characteristic value, low-price expenditure level tendency characteristic value, and consumption diffusion amplitude characteristic value for each student to be evaluated in the designated class. Extract the mean values ​​of the consumption clustering intensity characteristic, low-price expenditure level tendency characteristic, and consumption diffusion amplitude characteristic, sum them, and obtain a restricted sum value. Ratio the mean values ​​of the consumption clustering intensity characteristic, low-price expenditure level tendency characteristic, and consumption diffusion amplitude characteristic to the restricted sum value, and use the corresponding results as the consumption clustering adjustment coefficient. Low-price expenditure adjustment coefficient Consumption diffusion moderating coefficient .

[0045] The following is a specific implementation example of calculating the comprehensive consumption restriction characteristic value of a student to be assessed within a set class. The available data includes the consumption clustering intensity characteristic value, low-price expenditure level tendency characteristic value, and consumption diffusion amplitude characteristic value of 5 randomly selected students to be assessed within the class, as detailed in Table 1 and... Figure 2 As shown:

[0046] Table 1. Example of a data set mapping the consumption patterns of students to be evaluated within a class.

[0047] Consumption Aggregation Intensity Characteristic Value Low-price spending hierarchy propensity characteristic value Consumption diffusion amplitude characteristic value Student 1 to be evaluated 0.684 0.628 0.342 2 students awaiting evaluation 0.656 0.726 0.283 3 students awaiting evaluation 0.834 0.758 0.243 4 students awaiting evaluation 0.765 0.674 0.367 5 students pending evaluation 0.782 0.826 0.184

[0048] Consumption clustering adjustment coefficients stored in the database Approximately 0.341;

[0049] Low-price expenditure adjustment coefficient stored in the database Approximately 0.402;

[0050] Consumption diffusion moderating coefficients stored in the database Approximately 0.257;

[0051] Substituting the data from Table 1 and the aforementioned coefficients into the specific formula for calculating the comprehensive consumption restriction characteristic value of a student to be assessed within the class, we obtain:

[0052] Set the comprehensive consumption restriction characteristic value of the first student to be evaluated in the class as 0.341×0.684+0.402×√0.628+0.257×exp(-0.342)≈0.502;

[0053] Set the comprehensive consumption restriction characteristic value of the second student to be evaluated in the class as 0.341×0.656+0.402×√0.726+0.257×exp(-0.283)≈0.524;

[0054] Set the comprehensive consumption restriction characteristic value of the third student to be evaluated in the class as 0.341×0.834+0.402×√0.758+0.257×exp(-0.243)≈0.601;

[0055] Set the comprehensive consumption restriction characteristic value of the fourth student in the class to be evaluated as 0.341×0.765+0.402×√0.674+0.257×exp(-0.367)≈0.531;

[0056] Set the comprehensive consumption restriction characteristic value of the fifth student to be evaluated in the class as 0.341×0.782+0.402×√0.826+0.257×exp(-0.184)≈0.615.

[0057] like Figure 3 As shown, the specific steps for analyzing and setting the consumption pattern mapping feature set of each student to be evaluated in the class are as follows: In the consumption input layer of the consumption evolution mapping model, the consumption behavior time series data of each student to be evaluated in the class is received and preprocessed, such as normalizing the consumption behavior time series data of the students to be evaluated and mapping its specific values ​​to between 0 and 1.

[0058] In the consumption evolution layer of the consumption evolution mapping model, time series analysis is performed on the preprocessed consumption behavior time series data of each student to be evaluated in the set class to extract the corresponding consumption evolution feature vector of the student to be evaluated. Specifically, this layer uses LSTM. The LSTM network models the time dependency of the input consumption behavior time series data. By learning the consumption change relationship between adjacent time steps, it captures the consumption change pattern and expenditure rhythm characteristics of students in different periods. Through the mechanism of recurrent neural network (RNN), it updates its hidden state at each time step to retain the contextual information of historical consumption behavior. At the same time, combined with the nonlinear control mechanism of gating units (input gate, forget gate, and output gate), it selectively memorizes and updates historical expenditure characteristics, thereby strengthening the capture of key consumption patterns (such as periodic consumption, sudden expenditure, or continuous low consumption). After LSTM processing, the output hidden state sequence is the consumption evolution feature vector that can characterize the consumption change trend and behavioral stability of the student to be evaluated within the set period, such as:

[0059] For each time period, the average number of meals in the canteen and the rate of change of the number of meals in the canteen are extracted (e.g., the number of meals in the canteen in the first time period / the number of meals in the canteen in the second time period, and the average is taken). Normalization is then performed. Based on the normalized average number of meals in the canteen and the rate of change of the number of meals in the canteen, a weighted average is calculated. In this weighted average, the rate of change of the number of meals in the canteen is taken as its reciprocal, which is expressed as 1 / (1 + rate of change of the number of meals in the canteen) to extract the concentrated features of on-campus dining, which is used to characterize the degree of students' dependence on on-campus dining within a set period. The larger the value, the greater the degree of economic constraint.

[0060] For each time period, the average and standard deviation of the canteen meal expenses are extracted. Based on the skewness formula, the skewness feature of the meal expense distribution is extracted to characterize the students' tendency to spend on food and beverages at different levels within a set period. When the skewness is low, it indicates that the meal expenses are concentrated in the low-price range, and students maintain low-price dining behavior for a long time, and their spending is highly restricted.

[0061] For each time period, the average number of supermarket purchases and the rate of change of supermarket purchases are extracted and normalized. Based on the normalized average number of supermarket purchases and the rate of change of supermarket purchases, a weighted average is calculated. In this weighted average, the rate of change of supermarket purchases is taken as its reciprocal, expressed as 1 / (1+ rate of change of supermarket purchases), to extract the concentrated characteristics of supermarket consumption. This is used to characterize the stability of students' daily consumption within a set period. When the value is high, it indicates that students' supermarket consumption behavior is relatively concentrated and has small frequency fluctuations within the set period, and their economic affluence is relatively low.

[0062] For each time period, the supermarket spending amount is extracted as mean and standard deviation. Based on the skewness formula, the skewness feature of supermarket spending distribution is extracted to characterize the students' daily non-food spending level tendency within a set period. When it is low, it indicates that the supermarket spending amount is concentrated in the low price range, and students mainly buy low-priced or daily necessities, with limited consumption choices.

[0063] For the number of times a package is collected in each time period, a Fast Fourier Transform is performed to obtain several frequency components and corresponding package collection amplitudes. The maximum value of the package collection amplitude is extracted, squared, and then compared with the sum of the squares of the package collection amplitudes of all frequency components to extract the package collection activity characteristics. This is used to characterize the periodic regularity of students' package signing behavior and the degree of external consumption activity within a set period. The lower the value, the more irregular the students' package signing behavior is, the fewer the overall number of signings, the lower the external consumption activity, and the more limited their economic expenditure.

[0064] For each time period, the average value of laundry and personal care expenditure is calculated. The number of time periods with expenditures higher than the average and the number of time periods with expenditures lower than the average are counted separately, and the ratio is calculated to extract the laundry and personal care expenditure bias feature. This feature is used to characterize the students' hierarchical tendency in laundry and personal care expenditure. When the ratio is small, it indicates that the students maintain a low level of laundry and personal care expenditure in the long term, showing thrift or economic constraints.

[0065] For each time period's consumption time constraint value, the mean, maximum, and minimum consumption time constraints are extracted and then comprehensively processed. Specifically, the squared difference between the maximum and mean consumption time constraints is added to the squared difference between the mean and minimum consumption time constraints. The result is then divided by the difference between the maximum and minimum values ​​plus 1 to extract the consumption time extension feature, which characterizes the degree of relaxation in the consumption time distribution within a set period. When it is low, it indicates that students' consumption is mainly concentrated in fixed mealtimes, and the consumption time distribution is concentrated and regular, indicating that their consumption is subject to both time and budget constraints, and their economic freedom is low.

[0066] For the consumption trajectory expansion value of each time period, linear fitting is performed (such as using the least squares method to perform straight line fitting) to extract the fitting slope value, which is used as the trajectory diffusion trend feature to characterize the changing trend of students' consumption activity range within a set period. When it is small, it indicates that the students' consumption activity range has shrunk and has been concentrated in a fixed area for a long time, reflecting their economic constraints and shrinking living space.

[0067] For the low-price dependency value of each time period, the low-price dependency value of adjacent time periods is differentially processed to obtain the low-price dependency change sequence. The variance of the sequence is extracted, and its reciprocal is normalized to extract the low-price dependency stability feature, which is used to characterize the persistence of students' low-price consumption dependency within a set period. When it is high, it indicates that students' low-price dependency behavior is stable and fluctuates little, that is, it is fixed in the low-price window for a long time, showing a persistent economic constraint.

[0068] For the activity value of micro-spending in each time period, the change value of micro-spending activity in adjacent time periods is extracted in turn. The mean change value and the maximum change value of micro-spending activity are calculated and the ratio is processed to extract the micro-spending proportion offset feature, which is used to characterize the fluctuation range of students' low-spending behavior within a set period. When it is low, it indicates that the change of students' low-spending proportion is small and stable, which shows that they continue to maintain a frugal consumption habit.

[0069] The following features were combined to form a consumption evolution feature vector: concentrated characteristics of on-campus dining, skewness of dining amount distribution, concentrated characteristics of supermarket consumption, skewness of supermarket consumption distribution, active characteristics of express delivery collection, skewness of personal care expenditure, expansion of consumption time, trend of trajectory diffusion, stability of low-price dependence, and deviation of micro-expenditure proportion.

[0070] In the association mapping layer of the consumption evolution mapping model, the consumption evolution feature vector of each student to be evaluated in the set class is fused and mapped to extract the corresponding consumption status feature vector of the student to be evaluated. Specifically, the on-campus dining concentration feature, supermarket consumption concentration feature, and consumption time extension feature in the consumption evolution feature vector are weighted. In this weighting process, the consumption time extension feature is taken in its reciprocal form, which is expressed as 1 / (1+consumption time extension feature), to extract the consumption clustering intensity feature, which is used to characterize the overall consumption concentration of students within a set period. The higher the value, the more concentrated the consumption behavior is on campus and in fixed time periods, reflecting that their economic freedom is low and their economic constraints are high.

[0071] The skewness of the dining expenditure distribution, the skewness of the supermarket consumption distribution, the bias of the personal care expenditure, and the stability of low-price dependence in the consumption evolution feature vector are weighted. In this weighting process, the skewness of the dining expenditure distribution, the skewness of the supermarket consumption distribution, and the bias of the personal care expenditure are all taken in their reciprocal form. Taking the skewness of the dining expenditure distribution as an example, its form is 1 / (1+skewness of the dining expenditure distribution) to extract the low-price expenditure level tendency feature, which is used to characterize the degree of low-price consumption expenditure level bias of students within a set period. The higher the value, the more likely they are to maintain a low-price consumption structure in the long term and have obvious economic constraints.

[0072] The characteristics of express delivery collection activity, trajectory diffusion trend, and micro-expenditure proportion offset in the consumption evolution feature vector are weighted to extract the consumption diffusion amplitude feature, which is used to characterize the degree of diffusion of students' consumption activities within a set period. The lower the value, the more limited the students' consumption activities, the less external expenditure, and the more concentrated the consumption pattern, reflecting a higher degree of economic constraint. The consumption aggregation intensity feature, low-price expenditure level tendency feature, and consumption diffusion amplitude feature are concatenated into a consumption status feature vector.

[0073] In the output layer of the consumption evolution mapping model, based on the consumption status feature vector of each student to be evaluated in the class, the corresponding consumption pattern mapping feature set of the student to be evaluated is output. Specifically, the consumption clustering intensity feature, low-price expenditure level tendency feature, and consumption diffusion amplitude feature in the consumption status feature vector are activated by the Sigmoid function to obtain consumption clustering intensity feature value, low-price expenditure level tendency feature value, and consumption diffusion amplitude feature value between 0 and 1.

[0074] The pre-training steps of the consumption evolution mapping model are as follows:

[0075] Obtain a labeled dataset, which consists of time-series data on the consumption behavior of several students and their corresponding economic status labels. The economic status labels include information on the level of economic hardship or economic restriction confirmed by the school. Each sample in the labeled dataset includes time-series data on the consumption behavior of the target student over multiple consecutive time periods. The time-series data on consumption behavior includes multi-dimensional parameters such as the number of times the student ate in the cafeteria, the amount spent in the cafeteria, the number of times the student shopped at the supermarket, the amount spent at the supermarket, the number of times the student received a package, the expenditure on personal care products, the consumption time constraint, the consumption trajectory expansion, the low-price dependence, and the micro-expenditure activity, as well as the corresponding ground truth labels on economic restriction.

[0076] In the data preprocessing stage, the time-series data of each consumption behavior are normalized, and the values ​​of each parameter are uniformly mapped to the interval of 0, 10, 10, 1. The data is then divided into sliding windows according to the time order, and each sample sequence contains data from multiple consecutive time periods to maintain time dependence. After normalization and window division, the labeled dataset is divided into training set, validation set and test set, with 80% of the samples used for model training, 10% of the samples used for validation and 10% of the samples used for testing, to ensure a balanced data distribution.

[0077] The consumption evolution mapping model is trained. Taking the consumption evolution layer as an example, the consumption behavior feature vector of each student in a continuous time period is input into the LSTM network structure. The LSTM layer models the consumption feature sequence on a temporal basis through its gating mechanism of input gate, forget gate and output gate, learns the long-term and short-term change patterns of consumption behavior between adjacent time steps, and captures the characteristics of periodic consumption patterns, sudden expenditure fluctuations and continuous low expenditure. The model uses the backpropagation algorithm (BPTT) to optimize the parameters in order to minimize the error (such as mean square error MSE) between the predicted output consumption comprehensive restricted feature value and the labeled economic restricted level.

[0078] During training, optimization algorithms (such as the Adam optimizer) are used to iteratively update weight parameters and adjust hyperparameters such as learning rate, number of hidden layer units, and time step length to improve the model's convergence speed and generalization performance. The model's reconstruction error and classification accuracy at different training stages are monitored through the validation set, and an early stopping mechanism is used to prevent overfitting.

[0079] Once training is complete, the model's generalization ability is evaluated using a test set to verify its accuracy in predicting consumption-restricted features on unseen samples. This ensures that the model can accurately extract consumption evolution features and consumption pattern features. Finally, a pre-trained consumption evolution mapping model is obtained, and the model parameters and weight files are saved for use in the online identification and comprehensive consumption-restricted feature value calculation stages of subsequent assessment methods for economically disadvantaged students.

[0080] This implementation plan, through detailed analysis of time-series data on student consumption behavior and combined with a pre-trained consumption evolution mapping model, comprehensively extracts features closely related to students' consumption status, thereby forming a comprehensive consumption constraint feature value. This allows for in-depth analysis of students' consumption patterns and behavioral evolution, identifying long-term consumption patterns and their adaptation to consumption constraints. Secondly, based on LSTM, through time-series processing of multiple parameters in the consumption behavior data, it can meticulously depict the degree of consumption aggregation, expenditure hierarchy tendency, and consumption diffusion range of students in different time periods, thus revealing students' actual consumption status. Finally, through weighted processing of consumption evolution feature vectors, it can accurately reflect students' consumption stability, low-price dependence, and consumption activity range during the school year, thereby providing a more reliable consumption behavior analysis, improving the accuracy of assessment results, avoiding subjective bias in assessment methods, and ensuring a more fair identification of economically disadvantaged students.

[0081] Specifically, the data on cost constraints related to the student's place of origin includes residents' Engel coefficient, urban-rural income ratio, employment rate, CPI year-on-year change rate, minimum living allowance, industrial transformation intensity, and labor productivity. The specific steps for analyzing the characteristics of the educational support needs of students to be assessed are as follows: Based on the cost constraint data of the student's place of origin for each student in the designated class, analyze the corresponding student's place of origin assessment characteristic set, including the prosperity characteristic value and the living pressure characteristic value of the student's place of origin; obtain the family burden characteristic value for each student in the designated class, and combine it with the student's place of origin assessment characteristic set to analyze their... The corresponding educational support needs characteristic value for students to be assessed is obtained by weighting the characteristic values ​​of the prosperity of the student's place of origin, the pressure of living in the student's place of origin, and the burden of family burden for each student in the set class. In this weighting process, the characteristic value of the prosperity of the student's place of origin is expressed in its reciprocal form, which is 1 / (1+the characteristic value of the prosperity of the student's place of origin), to obtain the educational support needs characteristic value. This value is used to characterize the objective degree of assistance that the student needs to maintain his or her studies, ensure his or her basic living standard, and bridge the gap between his or her family's economic background and the general cost of living in his or her place of study.

[0082] Among them, the Engel coefficient is the proportion of food consumption expenditure to total consumption expenditure of local residents. It reflects the living standards and economic difficulties of residents. The higher the coefficient, the lower the living standard and the more consumption is spent on food. It can be identified by the place of origin in the student files stored in the database, and can be directly obtained from relevant reports of local statistical bureaus (such as the resident consumption expenditure statistical report) stored in the database.

[0083] The urban-rural income ratio is the proportional relationship between the income of urban and rural residents. It usually reflects the degree of uneven economic development between urban and rural areas. The higher the value, the greater the income gap between urban and rural areas. It can be obtained by extracting the per capita disposable income of urban residents and the per capita disposable income of rural residents from the social and livelihood statistical reports or similar statistical reports stored in the database of the statistics bureau of the student's place of origin, and then performing ratio processing. The result is used as the urban-rural income ratio.

[0084] The employment rate is the proportion of the working-age population to the employed population. It is used to measure the employment situation in a region. A higher employment rate usually means better economic development. It can be obtained by taking the average of the monthly labor market-related statistical reports stored in the database of the local statistics bureau in the student's place of origin.

[0085] The CPI year-on-year change rate is the magnitude of the change in the Consumer Price Index (CPI, which represents the relative change in the prices of a basket of goods and services) year-on-year. It is usually used to measure the speed of price changes. It can be obtained by looking at the monthly reports (such as the monthly regional consumer price index report) stored in the database of the local statistics bureau in the student's place of origin. The CPI year-on-year value for each month can be processed by trend analysis, such as dividing the difference between the CPI year-on-year values ​​of the first and second consecutive months by the CPI year-on-year value of the second consecutive month, and then taking the average as the value.

[0086] The minimum subsistence allowance is the economic level of a region that guarantees the basic survival needs of its residents. It can be obtained by retrieving the relevant reports on the minimum subsistence allowance standards stored in the database of the local civil affairs bureau in the student's place of origin.

[0087] The industrial transformation intensity value is the degree to which a region transforms from traditional industries to emerging and high-tech industries. Generally speaking, the higher the proportion of output value of high-tech industries, the more it means that the region's economic structure has been transformed and that it has a higher capacity for sustainable development. It can be calculated by statistically analyzing the output value of various industries (high-tech industries) in the regional economic statistics reports stored in the database of the local statistics bureau of the student's place of origin, and then comparing it with the total output value. The result is used as the industrial transformation intensity value.

[0088] Labor productivity is the economic value created by workers, reflecting the efficiency of labor in economic activities. Higher labor productivity means that workers create more wealth and that the economy is more robust. It can be obtained by extracting regional GDP and labor force from labor productivity reports or regional GDP reports stored in the database of the local statistics bureau of the student's place of origin, performing ratio processing, and using the result as the labor productivity value.

[0089] Furthermore, it should be noted that in the weighted processing involved in this implementation example, the weighting coefficients corresponding to each parameter can be obtained using sample entropy weighting. Taking the weighted processing process for obtaining the characteristic value of school support needs as an example, the characteristic values ​​of the prosperity of the student's place of origin, the characteristic value of the student's place of origin living pressure, and the characteristic value of the family burden for each student to be evaluated in the set class are read and normalized respectively. The reciprocal of the information entropy value of each parameter is extracted and summed to obtain the information entropy sum value. The reciprocal of the information entropy value of each parameter is compared with the information entropy sum value to obtain the weighting coefficient corresponding to each parameter.

[0090] The specific steps for analyzing the evaluation feature set of the student's place of origin for each student in the set class are as follows: Based on the urban-rural income ratio, employment rate, industrial transformation intensity, and labor productivity of each student in the set class, the prosperity feature value of the corresponding student's place of origin is analyzed. Specifically, the urban-rural income ratio, employment rate, industrial transformation intensity, and labor productivity of each student in the set class are standardized. The standardized results are then weighted. In this weighting process, the standardized urban-rural income ratio is expressed in its reciprocal form as 1 / (1+standardized urban-rural income ratio) to obtain the prosperity feature value of the corresponding student's place of origin. This value is used to characterize the social prosperity of the student's place of origin. For students, those from these prosperous areas have a lower degree of poverty.

[0091] Based on the Engel coefficient, minimum living allowance, and CPI year-on-year change rate of each student to be assessed in a set class, the corresponding living pressure characteristic value of the student to be assessed is analyzed. Specifically, the Engel coefficient, minimum living allowance, and CPI year-on-year change rate of each student to be assessed in the set class are standardized. The standardized results are then weighted. In this weighting process, the standardized minimum living allowance is expressed as its reciprocal, which is expressed as 1 / (1+standardized minimum living allowance). This yields the living pressure characteristic value of the student's place of origin, which is used to characterize the economic affordability and living burden of the student's place of origin. The higher the value, the greater the cost of living pressure and the worse the residents' economic affordability.

[0092] The specific steps to obtain the family burden characteristic value of each student to be assessed in the set class are as follows: Obtain the family status data (including family dependency ratio, family income value, and student loan value) of each student to be assessed in the set class, and perform normalization processing (that is, normalize the family dependency ratio, family income value, and total student loan value, and map their corresponding values ​​to the range of 0-1).

[0093] Based on the normalized family status data of each student to be assessed in the designated class, the family burden characteristic value of the students to be assessed is analyzed. That is, the family dependency ratio, family income and total student loan amount of each student to be assessed in the designated class are weighted. In this weighting process, the normalized family income value is expressed in its reciprocal form, which is expressed as 1 / (1+normalized family income value), so as to obtain the corresponding family burden characteristic value of the student to be assessed.

[0094] The family dependency ratio is the ratio of dependent members (such as children, elderly parents, etc.) to the total population of the family. It reflects the economic burden of the family. The higher the dependency ratio, the more members in the family need to bear responsibility and the greater the economic pressure. It can be obtained by obtaining the total number of employed persons and the total population of the family from the student's hardship application form stored in the database, and then performing ratio processing to use the result as the family dependency ratio.

[0095] Family income is the total income of a family over a certain period of time, which can be obtained from the student's hardship assessment application form stored in the database.

[0096] Student loans are the amount of money that families borrow from banks or other institutions to support their children's education. They can be obtained from student loan application records stored in a database.

[0097] This implementation plan, through in-depth analysis of the cost constraints of the students' place of origin, can comprehensively reflect the gap between the students' family economic background and their needs for educational support. By standardizing and weighting the analysis of multiple key parameters (such as the urban-rural income ratio, Engel's coefficient, and employment rate), it can accurately extract the economic characteristics of the students' place of origin, thereby extracting their living pressure and educational support needs. This provides objective data for the assessment of economically disadvantaged students, ensuring that the assessment results fully consider the development level and cost of living in different regions, avoiding potential regional biases in the assessment. In addition, by accurately extracting the characteristic values ​​of family burden and combining them with the assessment characteristic set of the place of origin, it is possible to clarify the actual needs of each student, making the assessment process more comparable and fair, thereby improving the accuracy of the assessment.

[0098] Please see Figure 4 This invention provides a technical solution: a data mining-based assessment system for economically disadvantaged students, comprising: a data acquisition module for acquiring time-series data on consumption behavior and cost constraints of the student's place of origin for several students to be assessed within a set class; a consumption constraint analysis module for analyzing the comprehensive consumption constraint characteristic value of each student to be assessed based on a pre-trained consumption evolution mapping model and the time-series data on consumption behavior of each student to be assessed within the set class; a comprehensive hardship assessment module for analyzing the educational support need characteristic value of each student to be assessed based on the cost constraints of the student's place of origin within the set class, and analyzing the comprehensive hardship assessment characteristic value of each student to be assessed based on the comprehensive hardship assessment characteristic value; and an economically disadvantaged student assessment feedback module for conducting economically disadvantaged student assessment processing for each student to be assessed within the set class based on the comprehensive hardship assessment characteristic value.

[0099] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0100] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A data mining-based assessment method for economically disadvantaged students, characterized in that, Includes the following steps: Obtain time-series data on the consumption behavior of several students to be evaluated within a set class, as well as data on the cost constraints of the students' place of origin; Based on a pre-trained consumption evolution mapping model, and combined with the time series data of the consumption behavior of each student to be evaluated in a set class, the comprehensive consumption limitation characteristic value of the corresponding student to be evaluated is analyzed. Based on the cost constraint data of each student's place of origin in the class, we analyze the corresponding student's school support needs characteristic value, and combined with the comprehensive consumption restriction characteristic value, we analyze the comprehensive difficulty assessment characteristic value of the corresponding student. Based on the comprehensive hardship assessment feature values, each student in the designated class is assessed for economic hardship.

2. The data mining-based assessment method for economically disadvantaged students according to claim 1, characterized in that, The consumption behavior time-series data includes the number of times dining in the canteen, the amount spent in the canteen, the number of times shopping in the supermarket, the amount spent in the supermarket, the number of times receiving express deliveries, the expenditure on personal care products, the consumption time constraint, the consumption trajectory expansion, the low-price dependence, and the micro-spending activity value for each time period. The consumption evolution mapping model includes a consumption input layer, a consumption evolution layer, a correlation mapping layer, and an output layer.

3. The data mining-based assessment method for economically disadvantaged students according to claim 2, characterized in that, The specific steps for analyzing and setting the comprehensive consumption limitation characteristic value for each student to be evaluated in the class are as follows: The consumption behavior time series data of each student to be evaluated in the class is input into the pre-trained consumption evolution mapping model to analyze the corresponding consumption pattern mapping feature set of the students to be evaluated, including consumption clustering intensity feature value, low-price expenditure level tendency feature value, and consumption diffusion amplitude feature value. Based on the consumption pattern mapping feature set of each student to be evaluated in the class, the comprehensive consumption restriction feature value of the corresponding student to be evaluated is analyzed.

4. The data mining-based assessment method for economically disadvantaged students according to claim 3, characterized in that, The specific steps for analyzing and defining the consumption pattern mapping feature set for each student to be evaluated within the class are as follows: In the consumption input layer of the consumption evolution mapping model, the consumption behavior time series data of each student to be evaluated in the set class is received and preprocessed. In the consumption evolution layer of the consumption evolution mapping model, time series analysis is performed on the consumption behavior time series data of each student to be evaluated in the preprocessed set class, and the corresponding consumption evolution feature vector of the student to be evaluated is extracted. In the association mapping layer of the consumption evolution mapping model, the consumption evolution feature vector of each student to be evaluated in the set class is fused and mapped to extract the consumption status feature vector of the corresponding student to be evaluated. In the output layer of the consumption evolution mapping model, based on the consumption status feature vector of each student to be evaluated in the class, the corresponding consumption pattern mapping feature set of the student to be evaluated is output.

5. The data mining-based assessment method for economically disadvantaged students according to claim 1, characterized in that, The data on cost constraints related to the student's place of origin includes residents' Engel coefficient, urban-rural income ratio, employment rate, CPI year-on-year change rate, minimum living allowance, industrial transformation intensity, and labor productivity. The specific steps for analyzing the characteristics of the students' educational support needs are as follows: Based on the cost constraint data of each student's place of origin in the class, the corresponding evaluation feature set of the student's place of origin is analyzed, including the characteristic value of the prosperity of the place of origin and the characteristic value of the pressure of living in the place of origin. Obtain the family burden characteristic value of each student to be assessed in the set class, and combine it with the student's place of origin assessment characteristic set to analyze the corresponding student's school support need characteristic value.

6. The data mining-based assessment method for economically disadvantaged students according to claim 5, characterized in that, The specific steps for analyzing and setting the assessment characteristic set of each student's place of origin within the class are as follows: Based on the urban-rural income ratio, employment rate, industrial transformation intensity, and labor productivity of each student to be evaluated in the class, the corresponding prosperity characteristics of the student's place of origin are analyzed. Based on the set Engel coefficient, minimum living allowance, and CPI year-on-year change rate for each student to be assessed in the class, the corresponding life stress characteristics of the students to be assessed are analyzed.

7. The data mining-based assessment method for economically disadvantaged students according to claim 5, characterized in that, The specific steps to obtain the family burden characteristic value for each student to be assessed in a given class are as follows: Obtain the family status data of each student to be evaluated in the designated class and perform normalization processing; Based on the normalized family status data of each student in the designated class, the characteristic values ​​of the family burden of the students to be assessed are analyzed.

8. The data mining-based assessment method for economically disadvantaged students according to claim 1, characterized in that, The specific formula for calculating the comprehensive difficulty assessment characteristic value of a student to be assessed within a given class is as follows: ； in, , , The criteria are as follows: setting comprehensive difficulty assessment characteristic values, comprehensive consumption restriction characteristic values, and educational support need characteristic values ​​for a student in the class to be assessed. , , The coefficients stored in the database are, in order: consumption restriction adjustment coefficient, education support adjustment coefficient, and interaction adjustment coefficient. .

9. The data mining-based assessment method for economically disadvantaged students according to claim 1, characterized in that, The specific steps for assessing and processing economically disadvantaged students in a designated class based on comprehensive hardship assessment feature values ​​are as follows: The comprehensive difficulty assessment characteristic value of each student to be assessed in the class will be compared with the preset comprehensive difficulty assessment characteristic threshold. Based on the judgment and processing results, each student in the designated class is identified as having financial difficulties.

10. A data mining-based assessment system for economically disadvantaged students, employing the data mining-based assessment method for economically disadvantaged students as described in any one of claims 1-9, characterized in that, include: The data acquisition module is used to acquire time-series data on the consumption behavior of several students to be evaluated within a set class, as well as data on the cost constraints of the students' place of origin. The consumption restriction analysis module is used to analyze the comprehensive consumption restriction characteristic value of each student to be evaluated based on a pre-trained consumption evolution mapping model and combined with the time series data of the consumption behavior of each student to be evaluated in a set class. The comprehensive hardship assessment module is used to analyze the corresponding student support needs characteristic values ​​based on the cost constraints of the student's place of origin in a set class, and to analyze the comprehensive hardship assessment characteristic values ​​of the corresponding student based on the comprehensive consumption restriction characteristic values. The Economically Disadvantaged Students Assessment and Feedback Module is used to assess each student in a designated class based on comprehensive hardship assessment feature values.

Images