Higher vocational college student financial literacy evaluation system based on big data analysis

Abstract

A higher vocational college student financial literacy evaluation system based on big data analysis. The system comprises a data collection module, a feature extraction module, a dynamic weight allocation module, and a comprehensive evaluation module. By means of collecting student behavior data and environmental data, establishing a feature mapping model, generating comprehensive evaluation factors, and performing dynamic weight adjustment, the present system solves the problem of significant deviations in evaluation results caused by static weights in conventional evaluation methods, such that the evaluation results are more accurate and comprehensive. Moreover, predictive analysis is performed on financial literacy by means of a deep learning model, optimization suggestions are dynamically adjusted, personalized education requirements are met, the evaluation efficiency is improved, resource waste and evaluation lag are prevented, and the development of the financial literacy of students is anticipated to guarantee the achievement of educational objectives.

Description

A Financial Literacy Assessment System for Vocational College Students Based on Big Data Analysis Technical Field

[0001] This invention belongs to the field of educational technology and big data analysis technology, specifically a financial literacy assessment system for vocational college students based on big data analysis. Background Technology

[0002] With the rapid development of big data and artificial intelligence technologies, the application of financial literacy assessment systems for vocational college students is receiving increasing attention from the education sector. However, existing assessment systems still have some shortcomings in data collection, processing, and analysis, leading to room for improvement in the accuracy and comprehensiveness of the assessment results.

[0003] A search revealed a patent, CN118841178B, concerning a method and system for assessing students' psychological literacy, published on December 24, 2024. This patent relates to the field of health data analysis technology. Through a health data aggregation module, a health assessment module, a data storage and analysis module, and a statistical adjustment module, it aims to more comprehensively understand and support the psychological health needs of students at different age levels. However, this technical solution primarily focuses on students' mental health; the assessment indicators and algorithm design fail to cover financial literacy, thus failing to comprehensively assess the financial literacy of vocational college students. Furthermore, the system is relatively simplistic in its data collection and processing, lacking the ability to comprehensively analyze multi-source data, resulting in low accuracy and reliability of the assessment results.

[0004] A search revealed a method and system for assessing students' comprehensive literacy, with publication number CN118840020B, published on February 11, 2025. This patent relates to the field of education system technology and, through modules for goal setting, preliminary assessment, adjustment and checking, comprehensive assessment, and comparison and calibration, can dynamically assess students' comprehensive literacy. However, this technical solution primarily focuses on students' academic performance and communication skills, failing to cover the assessment of financial literacy and thus unable to meet the specific needs of assessing the financial literacy of vocational college students. Furthermore, the system lacks full utilization of big data technology in data processing and analysis, resulting in low data processing and analysis efficiency and making it difficult to cope with the assessment needs of a large number of vocational college students. Technical issues

[0005] The aforementioned problems indicate that existing student assessment systems still have certain shortcomings in terms of the comprehensiveness of assessment content and data processing and analysis capabilities. Therefore, this invention provides a financial literacy assessment system for vocational college students based on big data analysis. It aims to optimize assessment indicators and algorithms through the comprehensive collection and analysis of multi-source data, thereby improving the accuracy and comprehensiveness of assessment results and meeting the needs of vocational colleges for an efficient and accurate financial literacy assessment system. Technical solutions

[0006] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a financial literacy assessment system for vocational college students based on big data analysis, which solves the problems mentioned in the background art through multi-source data fusion and dynamic weight adjustment mechanisms.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a financial literacy assessment system for vocational college students based on big data analysis, comprising a data acquisition module, a feature extraction module, a dynamic weight allocation module, and a comprehensive assessment module; and signal connections between the modules.

[0008] The data acquisition module is used to collect student behavior data and environmental data. Through data processing, it obtains consumption behavior frequency, financial knowledge mastery, social influence factor, and learning resource utilization rate, and sends the student behavior data and environmental data to the feature extraction module.

[0009] The feature extraction module is used to receive student behavior data and environmental data, establish a feature mapping model, generate comprehensive evaluation factors, and send the comprehensive evaluation factors to the dynamic weight allocation module.

[0010] The dynamic weight allocation module is used to obtain the comprehensive evaluation factors for each time period, and dynamically adjust the weights of each factor according to the changing trend within the time period to obtain the optimized weight distribution, and then send the optimized weight distribution to the comprehensive evaluation module.

[0011] The comprehensive assessment module receives the optimized weight distribution, combines it with real-time collected student behavior data, and uses a deep learning model to comprehensively assess students' financial literacy, generate an assessment report, and output optimization suggestions.

[0012] In a preferred embodiment, student behavior data includes the frequency of consumption behavior and the degree of financial literacy, while environmental data includes social influence factors and the utilization rate of learning resources.

[0013] The data processing includes: defining a time window, statistically analyzing students' consumption records within the time window, calculating the ratio of consumption frequency to the time window, and obtaining the consumption behavior frequency; where i is the i-th time window; the i-th time window corresponds to the i-th evaluation period;

[0014] By using a set financial knowledge test question bank, the system automatically calculates the students' correct answer rate and compares it with the total number of questions to obtain the degree of mastery of financial knowledge.

[0015] The system automatically identifies students' social network relationships, determines their influence in their social circles, collects the amount of financial literacy-related information disseminated in their social circles, and calculates the ratio with time windows to obtain the social influence factor.

[0016] The utilization rate of learning resources is calculated by statistically analyzing the duration and frequency of students' access to learning resources. Specifically, it is the average of the ratio of access duration to total learning duration plus the ratio of access frequency to total learning frequency.

[0017] In a preferred embodiment, student behavior data and environmental data are acquired, a feature mapping model is established, and a comprehensive evaluation factor is generated, specifically using the formula: C i =α·F i +β·K i +γ·S i +δ·R i

[0018] In the formula, C i To comprehensively evaluate factors, F i K i S i and R i These are preset proportional coefficients for consumption behavior frequency, financial knowledge mastery, social influence factor, and learning resource utilization rate, respectively, and α, β, γ, and δ are all greater than 0.

[0019] In a preferred embodiment, the comprehensive evaluation factors for each time period are obtained, resulting in C1,C1,…,C n The comprehensive evaluation factors for each time period are compared and calculated, and the calculation logic is ΔC. g =C g+1 -C g The differences of each comprehensive evaluation factor are obtained, where g is the g-th comparison;

[0020] Differences in comprehensive evaluation factors greater than 0 are defined as an upward trend, while differences in comprehensive evaluation factors less than 0 are defined as a downward trend.

[0021] In a preferred embodiment, the differences among the comprehensive evaluation factors are weighted and summed, and the calculation logic is as follows: Obtain the overall trend of financial literacy and compare it with the assessment threshold;

[0022] If the overall trend of financial literacy is greater than or equal to the assessment threshold, the current student's financial literacy level will be rated as advanced, and an incentive signal will be generated.

[0023] If the overall trend of financial literacy is less than the assessment threshold, the current student's financial literacy level will be rated as intermediate or low, and an improvement signal will be generated.

[0024] In a preferred embodiment, the optimization suggestion includes the evaluation results of generating incentive signals and the evaluation results of generating improvement signals, and financial literacy prediction analysis is performed through a deep learning model to determine the optimization adjustment strategy;

[0025] The system identifies the individual needs of students, automatically matches suitable learning resources, and counts the number of students currently using the recommended resources to obtain the resource utilization activity level.

[0026] In a preferred embodiment, the deep learning model used is the LSTM-Attention model. The specific steps for financial literacy prediction and analysis using the deep learning model are as follows:

[0027] Step B1: Obtain training data;

[0028] Step B2, construct the LSTM-Attention model;

[0029] Step B3: Train the model parameters using the Adam optimization algorithm;

[0030] Step B4: Verify the model's performance by checking its generalization ability using cross-validation.

[0031] Step B5: Use the trained model to predict the trend of financial literacy changes over a future period of time, and take the maximum value of the prediction result as the key node for the improvement of financial literacy in the future period of time.

[0032] In a preferred embodiment, the LSTM-Attention model formula is: H t =σ(W h· [H t-1 ,X t ]+b h A t =softmax(W a ·H t +b a Y t =W y ·(A t ·H t )+b y

[0033] In the formula, H t Let W be the hidden state at the current time step, σ be the activation function, and W be the hidden state at the current time step. h and b h These are the hidden layer weight matrix and bias term, respectively, X t Given the input feature vector, A t For attention weights, W a and b a These are the attention layer weight matrix and bias term, respectively. t To output the predicted value, W y and b y These are the output layer weight matrix and the bias term, respectively.

[0034] In a preferred embodiment, in step B3, W h W a W y Obtained through iterative updates using the Adam optimization algorithm. Beneficial effects

[0035] This invention establishes a feature mapping model by collecting student behavior data and environmental data, generates comprehensive evaluation factors, and dynamically adjusts their weights. This solves the problem of large deviations in evaluation results caused by static weights in traditional evaluation methods, making the evaluation results more accurate and comprehensive.

[0036] This invention receives the optimized weight distribution, evaluates the generated improved signal, monitors student behavior data in real time, and uses a deep learning model to predict and analyze financial literacy to obtain future trends. It then dynamically adjusts and optimizes suggestions to meet personalized education needs while improving evaluation efficiency, avoiding resource waste and evaluation lag. At the same time, it anticipates students' financial literacy development to ensure the achievement of educational goals. Attached Figure Description

[0037] Figure 1 is a schematic diagram of the modules of a financial literacy assessment system for higher vocational students based on big data analysis according to the present invention.

[0038] Figure 2 illustrates the computational logic and parameter update process of the LSTM-Attention model in the comprehensive evaluation module of this invention. Embodiments of the present invention

[0039] This invention provides a financial literacy assessment system for vocational college students based on big data analysis. Its core lies in achieving accurate assessment of students' financial literacy through multi-source data fusion and a dynamic weight adjustment mechanism. The specific embodiments of this invention are described in detail below with reference to Figures 1 to 6.

[0040] First, as shown in Figure 1, this system includes a data acquisition module, a feature extraction module, a dynamic weight allocation module, and a comprehensive evaluation module. These modules are interconnected via signals to form a complete evaluation process. The data acquisition module is the foundation of the entire system. Its main function is to collect student behavior data and environmental data, and then process this data to generate key indicators such as consumption behavior frequency, financial literacy, social influence factor, and learning resource utilization. In practical applications, the data acquisition module defines a time window to statistically analyze students' consumption records and calculates the ratio of consumption frequency to the time window, thus obtaining the consumption behavior frequency. For example, in a specific embodiment, a one-month time window is set, and a student's consumption records within that month are analyzed. If the number of consumptions is 30, the consumption behavior frequency is 30 / 30 = 1. Furthermore, the data acquisition module automatically calculates students' correct answer rate using a pre-defined financial literacy test question bank and compares it to the total number of questions to obtain the financial literacy level. For example, if a student answers 80 questions correctly in a test out of a total of 100 questions, the financial literacy level is 80 / 100 = 0.8. The social influence factor is calculated by automatically identifying students' social network relationships, determining their influence within their social circles, and collecting the amount of financial literacy-related information disseminated within those circles, then calculating the ratio between this amount and a time window. For example, if a student disseminates 50 pieces of financial literacy-related information within a month, the social influence factor is approximately 50 / 30 ≈ 1.67. The learning resource utilization rate is calculated by statistically analyzing the duration and frequency of students' access to learning resources. The ratio of access duration to total learning time is added to the ratio of access frequency to total learning frequency, and the average is taken. For example, if a student's total access time to learning resources is 20 hours, total learning time is 40 hours, access frequency is 10 times, and total learning frequency is 20 times within a month, then the learning resource utilization rate is (20 / 40 + 10 / 20) / 2 = 0.75.

[0041] Next, the feature extraction module receives student behavior data and environmental data from the data acquisition module, and establishes a feature mapping model to generate comprehensive evaluation factors. The core formula of the feature extraction module is C. i =α·F i +β·K i +γ·S i +δ·R i C i To comprehensively evaluate factors, F i K i S i and R iThese are preset proportional coefficients for consumption behavior frequency, financial literacy, social influence factor, and learning resource utilization, with α, β, γ, and δ all greater than 0. In practical applications, these proportional coefficients can be adjusted according to specific needs. For example, in one embodiment, setting α = 0.3, β = 0.3, γ = 0.2, and δ = 0.2 allows the calculation results of the comprehensive evaluation factors to reflect the relative importance of each indicator. The feature extraction module sends the generated comprehensive evaluation factors to the dynamic weight allocation module.

[0042] The dynamic weight allocation module obtains the comprehensive evaluation factors for each time period and dynamically adjusts the weights of each factor based on the changing trends within the time period, thereby obtaining the optimized weight distribution. The dynamic weight allocation module first obtains the comprehensive evaluation factors C1, C1, ..., C1 for each time period. n Then, the difference ΔC between each comprehensive evaluation factor is obtained through comparative calculation. g =C g+1 -C g For example, if a student's comprehensive evaluation factor is 0.8 in the first evaluation period and 0.9 in the second, the difference in comprehensive evaluation factor is 0.9 - 0.8 = 0.1, indicating an upward trend. Conversely, if the difference in comprehensive evaluation factor is less than 0, it indicates a downward trend. The dynamic weight allocation module further calculates the weighted sum of the differences in each comprehensive evaluation factor, and its calculation logic is as follows: This yields the overall trend of financial literacy and compares it with an assessment threshold. For example, setting the assessment threshold to 0.5, if the overall trend of financial literacy is greater than or equal to 0.5, the student's financial literacy level is rated as advanced, and an incentive signal is generated; if the overall trend is less than 0.5, the student's financial literacy level is rated as intermediate or low, and an improvement signal is generated. The dynamic weight allocation module sends the optimized weight distribution to the comprehensive assessment module.

[0043] The comprehensive assessment module is the core of the entire system. Its main function is to receive the optimized weight distribution, combine it with real-time collected student behavior data, and use a deep learning model to comprehensively assess students' financial literacy, generating an assessment report and optimization suggestions. The comprehensive assessment module uses an LSTM-Attention model for financial literacy prediction and analysis. The specific steps are as follows: Step B1, obtain training data; Step B2, construct the LSTM-Attention model; Step B3, train the model parameters using the Adam optimization algorithm; Step B4, verify the model's performance by checking its generalization ability through cross-validation; Step B5, use the trained model to predict the trend of financial literacy changes over a future period, and use the maximum value of the prediction result as the key node for improvement in financial literacy over the future period. The core formula of the LSTM-Attention model is H... t =σ(W h ·[H t-1 ,X t ]+b h A) t =softmax(W a ·H t +b a ), Y t =W y ·(A t ·H t )+b y H t Let W be the hidden state at the current time step, σ be the activation function, and W be the hidden state at the current time step. h and b h These are the hidden layer weight matrix and bias term, respectively, X t Given the input feature vector, A t For attention weights, W a and b a These are the attention layer weight matrix and bias term, respectively. t To output the predicted value, W y and b y These are the output layer weight matrix and the bias term, respectively. In step B3, W... h W a W y The results are obtained through iterative updates using the Adam optimization algorithm. The comprehensive evaluation module generates optimization suggestions based on the predictions of the deep learning model, including evaluation results for generating incentive signals and evaluation results for generating improvement signals. For the evaluation results of the improvement signals, the system identifies the current student's individual needs, automatically matches suitable learning resources, and counts the number of students currently using the recommended resources to obtain resource utilization activity.

[0044] In summary, this invention achieves a comprehensive assessment of students' financial literacy through the collaborative work of a data acquisition module, a feature extraction module, a dynamic weight allocation module, and a comprehensive evaluation module. The data acquisition module obtains student behavioral and environmental data by defining time windows and setting up a test question bank. The feature extraction module generates comprehensive evaluation factors through a feature mapping model. The dynamic weight allocation module optimizes the evaluation results by dynamically adjusting the weight distribution. The comprehensive evaluation module uses a deep learning model to predict and analyze financial literacy and generate optimization suggestions. The technical advantage of this invention lies in solving the problem of large deviations in evaluation results caused by static weights in traditional evaluation methods, making the evaluation results more accurate and comprehensive. It also meets the needs of personalized education, improves evaluation efficiency, avoids resource waste and evaluation lag, and anticipates the development of students' financial literacy to ensure the achievement of educational goals.

Claims

1. A financial literacy assessment system for vocational college students based on big data analysis, characterized in that: It includes a data acquisition module (1), a feature extraction module (2), a dynamic weight allocation module (3), and a comprehensive evaluation module (4); the modules are connected by signals. The data acquisition module (1) is used to collect student behavior data and environmental data. Through data processing, it obtains the frequency of consumption behavior, the degree of financial knowledge mastery, social influence factors and the utilization rate of learning resources, and sends the student behavior data and environmental data to the feature extraction module (2). The feature extraction module (2) is used to receive student behavior data and environmental data, establish a feature mapping model, generate comprehensive evaluation factors, and send the comprehensive evaluation factors to the dynamic weight allocation module (3); The dynamic weight allocation module (3) is used to obtain the comprehensive evaluation factors for each time period, and dynamically adjust the weight of each factor according to the changing trend within the time period to obtain the optimized weight distribution, and send the optimized weight distribution to the comprehensive evaluation module (4). The comprehensive evaluation module (4) is used to receive the optimized weight distribution, combine it with the real-time collected student behavior data, and conduct a comprehensive evaluation of students' financial literacy through a deep learning model, generate an evaluation report and output optimization suggestions.

2. The financial literacy assessment system for vocational college students based on big data analysis according to claim 1, characterized in that: Student behavior data includes the frequency of consumption behavior and the level of financial literacy; environmental data includes social influence factors and the utilization rate of learning resources. The data processing includes: defining a time window, statistically analyzing students' consumption records within the time window, calculating the ratio of consumption frequency to the time window, and obtaining the frequency of consumption behavior; where i is the i-th time window; the i-th time window corresponds to the i-th evaluation period; By using a set financial knowledge test question bank, the system automatically calculates the students' correct answer rate and compares it with the total number of questions to obtain the degree of mastery of financial knowledge. The system automatically identifies students' social network relationships, determines their influence in their social circles, collects the amount of financial literacy-related information disseminated in their social circles, and calculates the ratio with time windows to obtain the social influence factor. The utilization rate of learning resources is calculated by statistically analyzing the duration and frequency of students' access to learning resources. Specifically, it is the average of the ratio of access duration to total learning duration plus the ratio of access frequency to total learning frequency.

3. The financial literacy assessment system for vocational college students based on big data analysis according to claim 1, characterized in that: By acquiring student behavior data and environmental data, a feature mapping model is established to generate a comprehensive evaluation factor. The specific formula is: C i =α·F i +β·K i +γ·S i +δ·R i In the formula, C i To comprehensively evaluate factors, F i K i S i and R i These are preset proportional coefficients for consumption behavior frequency, financial knowledge mastery, social influence factor, and learning resource utilization rate, respectively, and α, β, γ, and δ are all greater than 0.

4. The financial literacy assessment system for vocational college students based on big data analysis according to claim 1, characterized in that: Obtain the comprehensive evaluation factors for each time period, and get C1, C1, ..., C n The comprehensive evaluation factors for each time period are compared and calculated, and the calculation logic is ΔC. g =C g+1 -C g The differences of each comprehensive evaluation factor are obtained, where g is the g-th comparison; Differences in comprehensive evaluation factors greater than 0 are defined as an upward trend, while differences in comprehensive evaluation factors less than 0 are defined as a downward trend.

5. The financial literacy assessment system for vocational college students based on big data analysis according to claim 4, characterized in that: The differences among the comprehensive evaluation factors are weighted and summed; the calculation logic is as follows: Obtain the overall trend of financial literacy and compare it with the assessment threshold; If the overall trend of financial literacy is greater than or equal to the assessment threshold, the current student's financial literacy level will be rated as advanced, and an incentive signal will be generated. If the overall trend of financial literacy is less than the assessment threshold, the current student's financial literacy level will be rated as intermediate or low, and an improvement signal will be generated.

6. The financial literacy assessment system for vocational college students based on big data analysis according to claim 1, characterized in that: The optimization suggestions include the evaluation results of generating incentive signals and the evaluation results of generating improvement signals. Financial literacy prediction analysis is performed through a deep learning model to determine optimization and adjustment strategies. The system identifies the individual needs of students, automatically matches suitable learning resources, and counts the number of students currently using the recommended resources to obtain the resource utilization activity level.

7. The financial literacy assessment system for vocational college students based on big data analysis according to claim 1, characterized in that: The deep learning model uses the LSTM-Attention model. The specific steps for predicting and analyzing financial literacy using the deep learning model are as follows: Obtain training data; Construct an LSTM-Attention model; The model parameters are trained using the Adam optimization algorithm; To verify the model's performance, cross-validation is used to check its generalization ability. The trained model is used to predict the trend of financial literacy changes over a future period, and the maximum value of the prediction result is taken as the key node for the improvement of financial literacy in the future period.

8. The financial literacy assessment system for vocational college students based on big data analysis according to claim 7, characterized in that: The formula for the LSTM-Attention model is: H t =σ(W h ·[H t-1 ,X t ]+b h A t =softmax(W a ·H t +b a Y t =W y ·(A t ·H t )+b y In the formula, H t Let W be the hidden state at the current time step, σ be the activation function, and W be the hidden state at the current time step. h and b h These are the hidden layer weight matrix and bias term, respectively, X t Given the input feature vector, A t For attention weights, W a and b a These are the attention layer weight matrix and bias term, respectively. t To output the predicted value, W y and b y These are the output layer weight matrix and the bias term, respectively.

9. A financial literacy assessment system for vocational college students based on big data analysis according to claim 8, characterized in that: During the training of model parameters using the Adam optimization algorithm, W h W a W y Obtained through iterative updates.

10. The financial literacy assessment system for vocational college students based on big data analysis according to claim 1, characterized in that: The data acquisition module (1) acquires student behavior data and environmental data by defining time windows and setting test question banks. The feature extraction module (2) generates comprehensive evaluation factors through feature mapping models. The dynamic weight allocation module (3) optimizes the evaluation results by dynamically adjusting the weight distribution. The comprehensive evaluation module (4) performs financial literacy prediction analysis and generates optimization suggestions through deep learning models.

Images