A student evaluation system based on student personality profiling and answer verification

Abstract

The present invention relates to the field of enrollment assessment, and specifically provides a student assessment system based on student character profiling and answer verification. Block supervision module: used to build a blockchain supervision network and verify the answers of students during the re-examination process; difference balance module: used to build student character profiles based on student information, and determine the score difference through the relative deviation ratio, and determine the score data after eliminating the difference; student source assessment module: used to import the student score table after answer verification and the score data after eliminating the difference into a fuzzy cloud model to generate student admission assessment results; secretary module: used to generate a student score table based on the imported student information and student admission assessment results during the student re-examination process.

Description

Technical Field

[0001] The present invention relates to the technical field of graduate student admission, and in particular to a student evaluation system based on student personality profiling and question answer verification. Background Art

[0002] Currently in the student re-examination process, the re-examination department has a strong degree of autonomy and is more difficult to control. The re-examination process is generally difficult to enforce the rules. This is mainly because the existing re-examination process and methods are basically based on manual operations, and the overall process is difficult to fully supervise. It is also difficult to disclose the process data of the overall process.

[0003] Compared with the initial test, the assessment format, standards and content of the re-examination are more complicated. If it still relies on manual work, mistakes are likely to occur.

[0004] In summary, during the examination and admission process, first of all, the statistics of students' scores are prone to manual statistical errors and need to be improved. A more complete supervision network needs to be built to achieve full-process supervision. At the same time, the content of the re-examination cannot be adjusted for different subjects. Summary of the invention

[0005] The present invention provides a student assessment system based on student personality profiling and answer verification to solve the problems that student performance statistics are prone to manual statistical errors and need to be improved. A more complete supervision network needs to be built to achieve full-process supervision. Different subjects cannot adjust the re-examination content, and the like.

[0006] In a first aspect, the present invention provides a student evaluation system based on student character profiling and answer verification, comprising:

[0007] Blockchain supervision module: used to build a blockchain supervision network and check students' answers during the re-examination process;

[0008] Difference equalization module: used to build a student character portrait based on student information, determine the score difference through the relative deviation ratio, and determine the score data after eliminating the difference;

[0009] Student source assessment module: used to import the student score table after answer verification and the score data after eliminating differences into the fuzzy cloud model to generate student admission assessment results;

[0010] Secretary module: used to generate student score tables based on the imported student information and student admission evaluation results during the student re-examination process.

[0011] In combination with the first aspect, the block supervision network is composed of the school public chain and the college private chain; wherein,

[0012] The college has no less than 2 private chains;

[0013] The private chain of the college forms a star-shaped interaction model based on multi-layer information interaction;

[0014] The star-shaped interaction model at least includes: expert re-examination and assessment layer, subject re-examination and assessment layer, college re-examination and assessment layer and school re-examination and assessment layer.

[0015] In conjunction with the first aspect, the answer verification includes: rule verification, operation verification, scoring verification and reform verification; wherein,

[0016] The rule verification is carried out by the blockchain network to build process information nodes for execution verification. Each node unit of the process information node accesses data through the blockchain and is configured with an independent scoring mechanism;

[0017] Operation verification is used to automatically configure and verify the score calculation results and student information during the student re-examination process;

[0018] High score verification is used for differentiated verification through an adaptive algorithm, which determines the students' differential scores to be eliminated based on the relative deviation ratio of the students' personality profiles;

[0019] Reform verification is used to verify reforms through a preset student selection prediction model.

[0020] In combination with the first aspect, the student character portrait includes:

[0021] According to the student information, obtain the student file data, interview data and student performance table, and parse the interview expression data representing the personality of the target student from the interview data;

[0022] The interview expression data and the student score table are temporally and spatially correlated to obtain the test personality portrait of the target student;

[0023] Analyze the behavioral event data used to characterize the target student's personality from the student file data to generate a specialty portrait of the target student;

[0024] Establish a fusion matrix based on the fusion of test personality profile and specialty profile;

[0025] The fusion matrix is ​​weighted, and a hierarchical model of student personality portrait is built to output the student personality portrait.

[0026] In combination with the first aspect, determining the score difference by the relative deviation ratio and determining the score data after eliminating the difference includes:

[0027] Based on the blockchain supervision network, the subject mean and group mean in the student re-examination process are determined, and based on the ratio of the subject mean to the group mean, the first relative deviation ratio is determined;

[0028] Associating the preset scoring index with the first relative deviation ratio, and determining the second relative deviation ratio of each scoring index;

[0029] Based on the student's personality profile, determine the difference between each scoring indicator and each student's score;

[0030] According to the score difference point and the second relative deviation ratio, the corresponding student scores are judged to determine the subjective difference in the scores, and the differences are eliminated according to the degree of subjective differences to determine the score data after eliminating the differences.

[0031] In combination with the first aspect, the fuzzy cloud model includes the following construction steps:

[0032] Establish an initial fuzzy cloud model, establish a big data training set, and determine the evaluation value of each student;

[0033] The big data training set is used to divide historical enrollment information into horizontal attribute divisions and vertical evaluation divisions;

[0034] Horizontal attribute division includes: basic situation, professional foundation, awards, initial test situation, and re-examination situation;

[0035] The vertical assessment classification includes: excellent, good, good, acceptable, and poor;

[0036] One-hot encode the evaluation values ​​to generate the student evaluation matrix;

[0037] Input the student evaluation behavior into the preset convolutional neural network to determine the student admission evaluation value;

[0038] The output layer of the convolutional neural network uses cross entropy as the loss function and uses the Adam optimizer for adaptive iterative optimization.

[0039] In combination with the first aspect, the steps for obtaining the student admission evaluation results are as follows:

[0040] Preset the admission evaluation parameters for students participating in the re-examination, and determine the membership type of the admission evaluation parameters through the Gaussian combination membership function;

[0041] The predicted subjects that students will be admitted to;

[0042] Adopting "if-then" rules to build a fuzzy rule base for admission prediction that evaluates the corresponding relationship between the membership of evaluation parameters and the predicted disciplines of students to be admitted;

[0043] Input the predicted values ​​of the admission evaluation parameters into the Gaussian combination membership function to determine the membership degree of the evaluation parameters;

[0044] By comparing the admission prediction fuzzy rule base, the student admission subject corresponding to the evaluation parameter membership is determined, and the admission evaluation parameters are determined through the evaluation indicators after fuzzy processing.

[0045] In combination with the first aspect, the fuzzy cloud model performs student admission evaluation, further comprising:

[0046] Determine the correlation coefficient between the evaluation indicators of student admission evaluation and the probability of student admission, and build a multi-layer hierarchical structure model;

[0047] Establish an evaluation comparison matrix based on a multi-layer hierarchical structure model and student information;

[0048] According to the evaluation comparison matrix, determine the weights of different evaluation indicators in the multi-layer hierarchical structure model;

[0049] Obtain the standard membership of different evaluation indicators;

[0050] The promotion and inhibition values ​​of evaluation indicators on student admission evaluation are calculated based on weights and standard membership.

[0051] In combination with the first aspect, the fuzzy cloud model is further configured with an adaptive update mechanism; wherein the adaptive update mechanism is executed in the following steps:

[0052] When there is a change in the student's personality profile, determine the characteristics of the change;

[0053] According to the membership type of the fuzzy cloud model, the evaluation parameters of the corresponding students are determined, and the compensation parameters of the score difference weighting corresponding to the change characteristics are calculated based on the evaluation parameters;

[0054] The compensation parameters are imported into the fuzzy cloud model to update the fuzzy cloud model.

[0055] In combination with the first aspect, the student score table is automatically counted by the information integration mechanism and the parameter configuration mechanism; wherein:

[0056] The information integration mechanism is used to read student information and connect to the Excel spreadsheet document to set the batch grouping information and initial and compound score items of the student score table;

[0057] The parameter configuration mechanism is used to set the weight of each grade item and the automatic calculation formula of the student grade table.

[0058] The beneficial effects of the present invention are:

[0059] The student evaluation system based on student character profiling and answer verification proposed by the present invention can comprehensively and objectively evaluate the comprehensive quality of students and improve the accuracy of evaluation results.

[0060] By building a blockchain supervision network, the present invention can ensure the authenticity and reliability of student scores and avoid data errors.

[0061] By adopting the method of weighting the scoring differences, the present invention can fully consider the personality characteristics of the students and make the evaluation results more comprehensive and objective.

[0062] By conducting student admission evaluation through a fuzzy cloud model, the present invention can provide students with appropriate admission recommendations and improve the allocation efficiency of educational resources.

[0063] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0064] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0066] Figure 1 This is a module composition diagram of a student evaluation system based on student personality profiling and answer verification in an embodiment of the present invention;

[0067] Figure 2 It is an evaluation dependency suppression flow chart of the fuzzy cloud model for student admission evaluation in an embodiment of the present invention;

[0068] Figure 3 This is a flowchart of updating the fuzzy cloud model in an embodiment of the present invention. DETAILED DESCRIPTION

[0069] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0070] In the existing student evaluation system, students’ grades are often only emphasized, while their personality characteristics are ignored. In addition, the existing student evaluation system has certain limitations in answer verification and cannot guarantee the accuracy of the evaluation results. Figure 1 As shown, the present invention proposes a student evaluation system based on student character profiling and answer verification, comprising:

[0071] Block supervision module: used to build a blockchain supervision network and check the answers of students during the re-examination. In the specific implementation, firstly, a supervision network based on blockchain is established during the students' re-examination to realize comprehensive supervision of the entire re-examination process. The block supervision module builds a blockchain supervision network to check the answers of students' scores to ensure the authenticity and reliability of the scores. In the process of verification, it is necessary not only to check the errors in students' scores, but also to ensure the transparency of the re-examination process, and to supervise whether there are operational errors, incomplete supervision, and excessive subjectivity in the expert scoring of students during the re-examination. At the same time, it is judged whether there will be calculation errors, same-name errors, and subjective differences in the scoring process.

[0072] Difference balance module: used to build a student character portrait based on student information, determine the score difference through the relative deviation ratio, and determine the score data after eliminating the difference; in the specific implementation process, the student character portrait is built through the difference balance module, including the student's personality characteristics, interests, hobbies, specialties, etc. The score difference of each student is judged based on the personality portrait and the relative deviation ratio, and whether there is a subjective difference in the score, so as to eliminate the subjective difference and obtain the score data after eliminating the subjective difference.

[0073] Student source assessment module: used to import the student score table after answer verification and the score data after eliminating differences into the fuzzy cloud model to generate student admission assessment results; in the specific implementation process, the student score table after answer verification and the score difference weighting are imported into the fuzzy cloud model to conduct student admission assessment. According to the assessment results, provide appropriate student admission opinions to the scoring teachers.

[0074] Secretary module: used to generate student score tables based on the imported student information and student admission evaluation results during the student re-examination process.

[0075] The principle of the above technical solution is:

[0076] The block supervision module of the present invention:

[0077] This module mainly uses blockchain technology to establish a decentralized supervision network. It can monitor the students' answers and the scores of the expert scoring group in real time during the re-examination process to ensure the fairness and credibility of the answers and scores. Specifically, the block supervision module can achieve the following functions: Real-time monitoring of answers: record students' answering time, answering order and other information to prevent cheating. Abnormal answer detection: analyze students' answering behavior, find possible cheating risks, and remind them. Data on the chain: write students' answer information and feedback results to the blockchain to ensure data security and decentralization. Real-time scoring monitoring, monitor the scoring of each member of the scoring group, determine the full process data disclosure of the scoring, judge whether there are subjective scoring differences in the scoring process, and at the same time judge whether there are scoring errors in the scoring process. The scoring errors are mainly in the scoring process. It is judged whether the scoring process is closed and whether there are external factors in the scoring process, resulting in the scoring process not meeting the predetermined scoring specifications.

[0078] The beneficial effects of the above technical solution are:

[0079] The student evaluation system based on student character profiling and answer verification proposed by the present invention can comprehensively and objectively evaluate the comprehensive quality of students and improve the accuracy of evaluation results.

[0080] By building a blockchain supervision network, the present invention can ensure the authenticity and reliability of student scores and prevent data from being tampered with.

[0081] The calculation of the evaluation deviation ratio can fully consider the students' personality characteristics and the differences in the expert group's scores, making the evaluation results more comprehensive and objective.

[0082] By using the fuzzy cloud model to evaluate student admission, the present invention can provide appropriate admission suggestions to the members of the scoring group and improve the allocation efficiency of educational resources.

[0083] In one embodiment, the block supervision network is composed of a school public chain and a college private chain; wherein,

[0084] The college has no less than 2 private chains;

[0085] The private chain of the college forms a star-shaped interaction model based on multi-layer information interaction;

[0086] The star-shaped interaction model at least includes: expert re-examination and assessment layer, subject re-examination and assessment layer, college re-examination and assessment layer and school re-examination and assessment layer.

[0087] The principle of the above technical solution is:

[0088] The blockchain supervision network of this application is built by the school platform as a transparent public chain platform, and then based on each college that recruits students, a separate private chain is built for each college. Each college has different majors, so the scoring standards are also different. The college's private chain can set its own nodes, and only authorized personnel can access and operate it.

[0089] The private chain can protect the personal privacy of students and the sensitive information of the college. The college's private chain includes multiple review links from expert re-examination to school re-examination. Each level has a dedicated node for review to ensure that the data of the entire review process can be made public without any abnormalities.

[0090] The school's public chain can be seen as a decentralized distributed ledger, where everyone can directly view and verify the entire scoring process, thereby improving the transparency of the scoring data. The college's private chain can flexibly add or delete nodes, and it can respond to changes in the college's internal structure or external environment at any time.

[0091] The beneficial effects of the above technical solution are:

[0092] The present invention realizes the interaction of different assessment layers by setting up public chains and private chains in a star-shaped interaction model, and ensures the accuracy and transparency of the scoring data, which can be made public online.

[0093] In one embodiment, the answer verification includes: rule verification, operation verification, score verification and reform verification; wherein,

[0094] The rule verification is carried out by the blockchain network to build process information nodes for execution verification. Each node unit of the process information node accesses data through the blockchain and is configured with an independent scoring mechanism;

[0095] Operation verification is used to automatically configure and verify the score calculation results and student information during the student re-examination process;

[0096] High score verification is used for differentiated verification through an adaptive algorithm, which determines the students' differential scores to be eliminated based on the relative deviation ratio of the students' personality profiles;

[0097] Reform verification is used to verify reforms through a preset student selection prediction model.

[0098] The principle of the above technical solution is:

[0099] The rule verification of the present invention adopts a blockchain network to build a process information node. Each node unit accesses data through the blockchain and is configured with an independent scoring mechanism. During the student evaluation process, the rule verification module will verify the student's answer to ensure the fairness and transparency of the evaluation process. The operation verification automatically configures and verifies the calculation results of the student's score table and the student information. The operation verification module will automatically check whether the calculation results in the student's score table are correct and whether the student information is accurate, thereby improving the accuracy of the evaluation process. The high score verification is differentiated through an adaptive algorithm. The adaptive algorithm judges the score difference based on the student's personality portrait and relative deviation ratio, judges the reason for the student's difference score, and then eliminates the score difference to achieve personalized evaluation of the student. The high score verification module will verify the student's answer to ensure the objectivity and comprehensiveness of the evaluation results. The reform verification is to conduct a reform verification through a preset student selection prediction model. The reform verification module will verify the student's answer to predict the student's future development trend.

[0100] The beneficial effects of the above technical solution are:

[0101] The present invention can improve the accuracy of student evaluation: through multi-dimensional verification such as rule verification, operation verification, high score verification and reform verification, the objectivity and comprehensiveness of the evaluation results are ensured, thereby improving the accuracy of student evaluation.

[0102] Improve the transparency and fairness of the evaluation process: Use blockchain networks to build process information nodes to ensure the transparency and fairness of the evaluation process.

[0103] Achieve personalized assessment: Perform differentiated verification through adaptive algorithms, and make adaptive replacements based on the relative deviation ratio of students' personality portraits to achieve personalized assessment of students.

[0104] Provide support for education reform: Verify reforms through preset student selection prediction models, predict students' future development trends, and determine the direction of education improvement.

[0105] Improve assessment efficiency: Use technical means such as automated configuration verification and adaptive algorithms to improve the efficiency of the assessment process and reduce the workload of teachers.

[0106] In one embodiment, the student character portrait includes:

[0107] According to the student information, obtain the student file data, interview data and student performance table, and parse the interview expression data representing the personality of the target student from the interview data;

[0108] The interview expression data and the student score table are temporally and spatially correlated to obtain the test personality portrait of the target student;

[0109] Analyze the behavioral event data used to characterize the target student's personality from the student file data to generate a specialty portrait of the target student;

[0110] Establish a fusion matrix based on the fusion of test personality profile and specialty profile;

[0111] The fusion matrix is ​​weighted, and a hierarchical model of student personality portrait is built to output the student personality portrait.

[0112] The principle of the above technical solution is:

[0113] First, based on the student information, the present invention can obtain the basic information of the student from the student file data, such as name, gender, age, usual scores and disciplinary violations, etc. This information will become the basic data of the student's personality portrait. Then, the interview expression data representing the personality of the target student is parsed from the interview data. The emotional data can reflect the emotional state of the student during the interview process, and judge the accuracy of the student's answers and whether the answers are easy during the interview process. The interview expression data and the student's score table are temporally and spatially associated to form a test personality portrait of the target student.

[0114] Linking students' behavioral performance with their academic performance can provide a comprehensive and dynamic image of students. Extracting behavioral event data that represents the target student's personality from the student profile data, such as daily behavior records, participation in extracurricular activities, etc., can enrich the understanding of the student's personality. Using the multi-source data collected above, establish a fusion matrix based on the fusion of test personality portraits and specialty portraits. The fusion matrix will be based on personality empowerment, and then build a hierarchical model of student personality portraits, and finally output student personality portraits. Student personality portraits visualize student personality portraits so that people can intuitively see the student's personality characteristics and strengths.

[0115] The beneficial effects of the above technical solution are:

[0116] The present invention can reflect the personality characteristics and learning advantages of students, so as to evaluate the students' future scientific research capabilities during the scoring process.

[0117] In one embodiment, determining the score difference by the relative deviation ratio and determining the score data after eliminating the difference comprises:

[0118] Based on the blockchain supervision network, the subject mean and group mean in the student re-examination process are determined, and based on the ratio of the subject mean to the group mean, the first relative deviation ratio is determined;

[0119] Associating the preset scoring index with the first relative deviation ratio, and determining the second relative deviation ratio of each scoring index;

[0120] Based on the student's personality profile, determine the difference between each scoring indicator and each student's score;

[0121] According to the score difference point and the second relative deviation ratio, the corresponding student scores are judged to determine the subjective difference in the scores, and the differences are eliminated according to the degree of subjective differences to determine the score data after eliminating the differences.

[0122] The principle of the above technical solution is:

[0123] The present invention is based on a blockchain supervision network. Blockchain technology is a decentralized distributed ledger technology. Each node can view all transaction records, and the transaction records cannot be tampered with or deleted. Therefore, a student evaluation system is established on the blockchain. Each node can access the evaluation data of all students, and these data are public and transparent, and no one can modify or delete them. Therefore, the first relative deviation ratio can be determined by the subject mean and group mean during the student reexamination process. Secondly, the smart contract in the blockchain technology can be used to realize automated scoring calculation and difference judgment. For example, a smart contract can be written. When a student completes a question, the smart contract will automatically upload the student's answer data to the blockchain, and calculate the first relative deviation ratio and the second relative deviation ratio of each scoring indicator of the student on the blockchain.

[0124] Then, based on each student's personality profile, calculate the difference between each scoring indicator and each student's score. This process can be done using machine learning algorithms, such as neural networks, random forests and other deep learning models. By training these models, we can get the predicted value of each student's score difference for different scoring indicators. Finally, based on these predicted values, combined with the score difference point and the second relative deviation ratio, make a difference judgment on each student's score, and then provide the difference point to the scoring review team. If a student's score is found to be wrong, the review team can correct the error by adjusting the corresponding score.

[0125] The beneficial effects of the above technical solution are:

[0126] The present invention can calculate the scoring differences among students during the scoring process, and then eliminate the scoring differences based on the relative deviation ratio, that is, the reasons for the scoring differences, to prevent unfair or abnormal scoring.

[0127] In one embodiment, the fuzzy cloud model includes the following construction steps:

[0128] Establish an initial fuzzy cloud model, establish a big data training set, and determine the evaluation value of each student;

[0129] The big data training set is used to divide historical enrollment information into horizontal attribute divisions and vertical evaluation divisions;

[0130] Horizontal attribute division includes: basic situation, professional foundation, awards, initial test situation, and re-examination situation;

[0131] The vertical assessment classification includes: excellent, good, good, acceptable, and poor;

[0132] One-hot encode the evaluation values ​​to generate the student evaluation matrix;

[0133] Input the student evaluation behavior into the preset convolutional neural network to determine the student admission evaluation value;

[0134] The output layer of the convolutional neural network uses cross entropy as the loss function and uses the Adam optimizer for adaptive iterative optimization.

[0135] The principle of the above technical solution is:

[0136] First, the present invention will establish an initial fuzzy cloud model, which is composed of some basic rules or thresholds combined with an open source fuzzy cloud model. It can be used as a preliminary evaluation tool to quickly distinguish the levels of students.

[0137] Then, the big data training set is used to further optimize and improve the model. This training set may contain a lot of detailed information about students, such as their grades, personality, interests and hobbies, etc. This information will be used to adjust the various parameters of the model to make it more suitable for new tasks. After the model is trained, it is applied to a new student evaluation matrix. In this matrix, the evaluation value of each student is uniquely encoded, which is an effective way to convert text or sequential data into numerical data. Finally, a preset convolutional neural network is used to determine the final admission evaluation value of each student. The output layer of this network uses cross entropy as the loss function, which is a very effective loss function that can help the network find the best weight setting faster. In addition, the Adam optimizer is used for adaptive iterative optimization, which can ensure that the model can find the optimal solution in various situations. The present invention combines traditional fuzzy evaluation methods with advanced deep learning technology, which can fully utilize the powerful computing power of cloud computing and realize the advantages of online learning and deployment.

[0138] The beneficial effects of the above technical solution are:

[0139] The present invention can train the school's historical enrollment data through the fuzzy cloud model, evaluate the current students through the historical enrollment data, judge the students' strengths, and continuously perform adaptive iterative optimization on the fuzzy cloud model.

[0140] In one embodiment, the fuzzy cloud model for student admission assessment includes:

[0141] Preset the admission evaluation parameters for students participating in the re-examination, and determine the membership type of the admission evaluation parameters through the Gaussian combination membership function;

[0142] The predicted subjects that students will be admitted to;

[0143] Adopting "if-then" rules to build a fuzzy rule base for admission prediction that evaluates the corresponding relationship between the membership of evaluation parameters and the predicted disciplines of students to be admitted;

[0144] Input the predicted values ​​of the admission evaluation parameters into the Gaussian combination membership function to determine the membership degree of the evaluation parameters;

[0145] By comparing the admission prediction fuzzy rule base, the student admission subject corresponding to the evaluation parameter membership is determined, and the admission evaluation parameters are determined through the evaluation indicators after fuzzy processing.

[0146] The principle of the above technical solution is:

[0147] The present invention presets the evaluation parameters involved in predicting the student admission evaluation, and these parameters may be known or may need to be inferred based on other information. These parameters may be any factors that can affect the student admission, such as grades, hobbies, personality traits, etc. Then, the membership discipline of the evaluation parameters is determined by the Gaussian combination membership function. The Gaussian combination membership function is a commonly used function type, which can map the input value to a specific interval and generate a corresponding membership within this interval. The predicted disciplines for student admission are preset. These categories may have been set in advance or may need to be inferred based on actual conditions. The admission prediction comparison admission prediction fuzzy rule base of the corresponding relationship between the evaluation parameter membership and the predicted category of student admission is constructed using the "if-then" rule to determine the student admission discipline corresponding to the evaluation parameter membership. This process is to find the corresponding rule based on the membership value obtained previously, and obtain a value representing the student admission category. The admission evaluation parameters are determined by the evaluation index after fuzzy processing. This process is to obtain an evaluation value representing the admission evaluation parameter after comprehensively considering the previous evaluation parameter prediction value and the final admission discipline.

[0148] The beneficial effects of the above technical solution are:

[0149] The present invention can predict the subjects of students, determine which subjects the students are more suitable for, and provide corresponding evaluation indicators.

[0150] In one embodiment, the fuzzy cloud model performs student admission assessment, further comprising:

[0151] Determine the correlation coefficient between the evaluation indicators of student admission evaluation and the probability of student admission, and build a multi-layer hierarchical structure model;

[0152] Establish an evaluation comparison matrix based on a multi-layer hierarchical structure model and student information;

[0153] According to the evaluation comparison matrix, determine the weights of different evaluation indicators in the multi-layer hierarchical structure model;

[0154] Obtain the standard membership of different evaluation indicators;

[0155] The promotion and inhibition values ​​of evaluation indicators on student admission evaluation are calculated based on weights and standard membership.

[0156] The principle of the above technical solution is:

[0157] like Figure 2 As shown, first, we need to identify the key factors that evaluate the probability of student admission. These factors may include academic performance, non-academic performance, behavioral performance, social skills, etc. We can use machine learning methods such as correlation analysis or decision trees to determine these factors and convert them into digital codes for use in the model. Then, a multi-layer hierarchical structure model (such as an artificial neural network) is used to calculate the impact of each factor on the probability of student admission. The input of this model will include the student's personal data, the value of each factor, and their relationship with each other, and the output will be the weight of each factor, that is, its importance in the evaluation. Next, these weights are matched with the student's personal data and student behavior to create an evaluation comparison matrix. This matrix will show each student's score on each factor, as well as their total score on all factors. Then, different methods (such as simple average, weighted average, etc.) are used to calculate the weight of each student on each factor. This will give the proportion of weight each factor has in the evaluation. Finally, these weights are applied to each student's evaluation results to calculate their impact on student admission. This is called "the promotion value and inhibition value of evaluation indicator dependence on student admission evaluation." The promotion value indicates the positive impact of the factor on the student's admission, while the inhibition value indicates the negative impact of the factor on the student's admission. The entire evaluation process will be continuously iterated and improved to ensure that our evaluation results are the most accurate.

[0158] The beneficial effects of the above technical solution are:

[0159] In the process of evaluating student admission, the present invention can build a multi-layer hierarchical structure model for continuing student evaluation, conduct a comprehensive evaluation of students based on multiple evaluation indicators with increasing difficulty, and determine the promotion value and inhibition value of student admission.

[0160] In one embodiment, the fuzzy cloud model is further configured with an adaptive update mechanism; wherein the adaptive update mechanism is executed in the following steps:

[0161] When there is a change in the student's personality profile, determine the characteristics of the change;

[0162] According to the membership type of the fuzzy cloud model, the evaluation parameters of the corresponding students are determined, and the compensation parameters of the score difference weighting corresponding to the change characteristics are calculated based on the evaluation parameters;

[0163] The compensation parameters are imported into the fuzzy cloud model to update the fuzzy cloud model.

[0164] The principle of the above technical solution is:

[0165] like Figure 3 As shown in Figure 1, first, obtain the changing image of the student's personality portrait during the scoring process. This part of the image may come from a variety of sources, such as teachers' observation reports, classmates' feedback, classroom records, enrollment data of other schools, reform policies, etc. Next, determine the evaluation parameters of the corresponding students according to the membership type of the fuzzy cloud model. This step requires an understanding of the model structure and data, as well as an analysis of the student's personality portrait. The evaluation parameters can be inputs or outputs in the model, or other intermediate variables. Then, based on the evaluation parameters, calculate the compensation parameters for the weighted score differences corresponding to the changing image. This step involves the calculation of the improvement or reduction of model performance, which requires comprehensive consideration of factors such as model error and training objectives. Finally, import the calculated compensation parameters into the fuzzy cloud model to update the model. This part may affect the accuracy of the model when predicting new data, so the updated model needs to be re-evaluated and optimized.

[0166] The beneficial effects of the above technical solution are:

[0167] The present invention can continuously compensate and update the fuzzy cloud model, thereby preventing excessive errors from occurring when evaluating students.

[0168] In one embodiment, the student score table is automatically counted by an information integration mechanism and a parameter configuration mechanism; wherein:

[0169] The information integration mechanism is used to read student information and connect to the Excel spreadsheet document to set the batch grouping information and initial and compound score items of the student score table;

[0170] The parameter configuration mechanism is used to set the weight of each grade item and the automatic calculation formula of the student grade table.

[0171] The principle of the above technical solution is:

[0172] The information integration mechanism of this application can easily modify or expand the information integration mechanism and parameter configuration mechanism of the student score form according to the needs of different schools or classes to adapt to different evaluation scenarios. When the evaluation criteria or scoring rules change, only the corresponding Excel document needs to be updated without changing the code of the entire system, reducing the maintenance cost of the system. The automatic statistics function can realize intelligent grade management and reduce the possibility of errors through the synergy of the information integration mechanism and the parameter configuration mechanism. All information collation and calculation processes of the student score form can be recorded in an Excel document, which is convenient for later query and audit. All classes or schools use the same system for evaluation, which ensures the consistency of the evaluation results and avoids evaluation deviations caused by inconsistent evaluation standards.

[0173] The beneficial effects of the above technical solution are:

[0174] The present invention can automatically score the results according to the pre-set standards, thus avoiding the subjective influence of human beings and improving the objectivity of the evaluation.

[0175] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A student assessment system based on student character profiling and answer verification, characterized in that: include: Blockchain supervision module: used to build a blockchain supervision network and check students' answers during the re-examination process; Difference equalization module: used to build a student character portrait based on student information, determine the score difference through the relative deviation ratio, and determine the score data after eliminating the difference; Student source assessment module: used to import the student score table after answer verification and the score data after eliminating differences into the fuzzy cloud model to generate student admission assessment results; Secretary module: used to generate student score tables based on the imported student information and student admission evaluation results during the student reexamination process; The student character profile includes: According to the student information, obtain the student file data, interview data and student performance table, and parse the interview expression data representing the personality of the target student from the interview data; The interview expression data and the student score table are temporally and spatially correlated to obtain the test personality portrait of the target student; Analyze the behavioral event data used to characterize the target student's personality from the student file data to generate a special profile of the target student; Establish a fusion matrix based on the fusion of test personality profile and specialty profile; The fusion matrix is ​​weighted, and a hierarchical model of student personality portrait is constructed to output the student personality portrait; Determining the score difference by the relative deviation ratio and determining the score data after eliminating the difference includes: Based on the blockchain supervision network, the subject mean and group mean in the student re-examination process are determined, and based on the ratio of the subject mean to the group mean, the first relative deviation ratio is determined; A smart contract is set up in the blockchain supervision network in advance, and the student's answer data is uploaded to the blockchain supervision network through the smart contract, and the second relative deviation ratio of the student's answer data relative to each scoring indicator is calculated; Based on the student's personality profile, calculate the difference points of each scoring indicator relative to each student; Determine the predicted value of the difference in scores for each student for different scoring indicators; Combine the predicted value of the scoring difference point and the second relative deviation ratio to make a difference judgment on each student's score, then provide the difference point to the scoring review team to determine the subjective difference of the score, and eliminate the difference according to the degree of subjective difference to determine the scoring data after eliminating the difference; The student score table after answer verification and the score data after eliminating differences are imported into the fuzzy cloud model, including: Preset the admission evaluation parameters for students participating in the retest, and determine the membership type of the admission evaluation parameters through the Gaussian combination membership function; The predicted subjects that students will be admitted to; Adopt "if-then" rules to build a fuzzy rule base for admission prediction that evaluates the corresponding relationship between the membership of the evaluation parameters and the predicted disciplines for student admission; Input the predicted values ​​of the admission evaluation parameters into the Gaussian combination membership function to determine the membership degree of the evaluation parameters; By comparing the admission prediction fuzzy rule base, determine the student admission discipline corresponding to the evaluation parameter membership, and determine the admission evaluation parameters through the evaluation indicators after fuzzy processing; Determine the correlation coefficient between the evaluation indicators of student admission evaluation and the probability of student admission, and build a multi-layer hierarchical structure model; Establish an evaluation comparison matrix based on a multi-layer hierarchical structure model and student information; According to the evaluation comparison matrix, determine the weights of different evaluation indicators in the multi-layer hierarchical structure model; Obtain the standard membership of different evaluation indicators; Calculate the promotion and inhibition values ​​of evaluation index dependence on student admission evaluation according to weights and standard membership; Establish an initial fuzzy cloud model, establish a big data training set, and determine the evaluation value of each student; The big data training set is used to divide historical enrollment information into horizontal attribute divisions and vertical evaluation divisions; Horizontal attribute division includes: basic situation, professional foundation, awards, initial test situation, and re-examination situation; The vertical assessment classification includes: excellent, good, good, acceptable, and poor; One-hot encode the evaluation values ​​to generate the student evaluation matrix; The student evaluation matrix is ​​input into the preset convolutional neural network to determine the student admission evaluation value; The output layer of the convolutional neural network uses cross entropy as the loss function and uses the Adam optimizer for adaptive iterative optimization; The student score table is automatically counted by the information integration mechanism and parameter configuration mechanism; The information integration mechanism is used to read student information and connect to the Excel spreadsheet document to set the batch grouping information of the student score table and the initial and re-examination score items; The parameter configuration mechanism is used to set the weight of each grade item and the automatic calculation formula of the student grade table.

2. A student evaluation system based on student character profiling and answer verification as claimed in claim 1, characterized in that: The blockchain supervision network is composed of the school public chain and the college private chain; The college has no less than 2 private chains; The private chain of the college forms a star-shaped interaction model based on multi-layer information interaction; The star-shaped interaction model at least includes: expert re-examination and assessment layer, subject re-examination and assessment layer, college re-examination and assessment layer and school re-examination and assessment layer.

3. A student evaluation system based on student character profiling and answer verification as claimed in claim 1, characterized in that: The answer verification includes: rule verification, operation verification, scoring verification and reform verification; among which, The rule verification is carried out by the blockchain network to build process information nodes for execution verification. Each node unit of the process information node accesses data through the blockchain and is configured with an independent scoring mechanism; Operation verification is used to automatically configure and verify the score calculation results and student information during the student re-examination process; Score verification is used to conduct differentiated verification through an adaptive algorithm, which determines the student's differential scores to be eliminated based on the relative deviation ratio of the student's personality profile; Reform verification is used to verify reforms through a preset student selection prediction model.

4. A student evaluation system based on student character profiling and answer verification as claimed in claim 1, characterized in that: The fuzzy cloud model is also configured with an adaptive update mechanism; wherein the execution steps of the adaptive update mechanism are as follows: When there is a change in the student's personality profile, determine the characteristics of the change; According to the membership type of the fuzzy cloud model, the evaluation parameters of the corresponding students are determined, and the compensation parameters of the score difference weighting corresponding to the change characteristics are calculated based on the evaluation parameters; The compensation parameters are imported into the fuzzy cloud model to update the fuzzy cloud model.

Images