Big data-driven scientific research whole process-oriented college student innovation ability evaluation method

Abstract

The invention provides a big data-driven college student innovation ability evaluation method oriented to the whole scientific research process. The big data-driven scientific research whole process-oriented college student innovation ability evaluation method comprises the following steps of: a, acquiring daily scientific research learning behavior data of college students by utilizing attendance checking and card punching equipment, a center console, a school database and monitoring equipment of a laboratory self-study room, including laboratory attendance checking data, self-study time records and activity tracks; according to the big data-driven college student innovation ability evaluation method oriented to the whole scientific research process, a set of complete evaluation system is formed from data acquisition to analysis, prediction and feedback. By introducing the attendance card punching equipment, the center console, the school database and the monitoring equipment of the laboratory study room, the daily scientific research and learning behavior data of the college students are comprehensively collected, and the data are cleaned and preprocessed by using the big data technology, so that the comprehensive monitoring and safety of the data in the whole scientific research process are ensured.

Description

Technical Field

[0001] The present invention relates to the field of big data technology, and in particular to a big data-driven method for evaluating the innovation ability of college students in the entire scientific research process. Background Art

[0002] The evaluation methods of college students' innovation ability mainly adopt a combination of qualitative and quantitative methods. From a qualitative perspective, teacher evaluation, peer evaluation and expert review are common means, focusing on the subjective evaluation of students' innovation performance and project contribution. From a quantitative perspective, questionnaire surveys, the number of patents or papers on innovative achievements, and competition awards provide data support. In addition, comprehensive quality files and process evaluations (such as innovation process records) are gradually being used to comprehensively reflect the level of students' innovation ability. These methods focus on both process and results, and strive to scientifically and fairly evaluate college students' innovation ability.

[0003] Although the system is efficient and intelligent, it still has some shortcomings. The data of the entire scientific research process cannot be fully monitored and the security cannot be guaranteed. The relationship between scientific research behavior and learning behavior cannot be modeled, and the quantitative evaluation criteria for innovation ability are unclear. Summary of the invention

[0004] In view of the shortcomings of the existing technology, the present invention provides a big data-driven method for evaluating college students' innovation ability in the entire scientific research process, which solves the problems that data in the entire scientific research process cannot be fully monitored and the security cannot be guaranteed, the relationship between scientific research behavior and learning behavior cannot be modeled, and the quantitative evaluation criteria for innovation ability are unclear.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a big data-driven method for evaluating the innovation ability of college students in the entire scientific research process, comprising:

[0006] a. Use attendance punching equipment, central control console, school database and monitoring equipment in laboratory study rooms to collect college students' daily scientific research and learning behavior data, including laboratory attendance data, study time records and activity tracks;

[0007] b. Use big data processing technology to clean and preprocess the collected raw behavior data to form a standard data set;

[0008] c. Monitor and shoot the entire scientific research process, collect panoramic images of students' scientific research process, and process the images in combination with intelligent behavior analysis models to extract scientific research behavior characteristics;

[0009] The behavior analysis model is:

[0010]

[0011] Among them, F action is the characteristic value of scientific research behavior, a i is the i-th action feature, w i is the weight of the action feature;

[0012] d. Jointly analyze standard data sets and scientific research behavior characteristics, generate daily learning behavior patterns, and build a behavior-result association model;

[0013] The behavior-outcome association model is:

[0014] R=cosine(V behavior ,V outcome )

[0015] Among them, R is the correlation, V behavior and V outcome They represent the behavior feature vector and the outcome feature vector respectively;

[0016] e. Based on the behavior and achievement association model, generate multi-dimensional semantic association feature vectors to quantify the innovation representation value in students' scientific research behavior;

[0017] The calculation formula of innovation characterization value is:

[0018]

[0019] Among them, I is the innovation capability characterization value, f j is the jth behavior characteristic value, w j is the feature weight;

[0020] f. Using the innovation representation value and historical academic data, establish an innovation capability prediction model; the innovation capability prediction formula is:

[0021]

[0022] Among them, Y is the estimated value of innovation capability, X k is the input feature, β k is the regression coefficient, ∈ is the error;

[0023] g. Dynamically track students’ scientific research and learning process and update the estimated value of innovation ability in real time;

[0024] h. Generate visual reports based on the estimated value of innovation ability and provide optimization suggestions to help students improve their scientific research and learning performance.

[0025] Preferably, the attendance punching device collects laboratory entry and exit records through student cards or biometric devices and uploads them to the central control console in real time. The school database is used to store students' personal information, course schedules and historical academic data.

[0026] Preferably, the monitoring device can record students' self-study time, equipment usage and learning content, and the behavior analysis model is based on a multimodal learning method of big data to assign dynamic weights to action features.

[0027] Preferably, the behavior and achievement association model determines the relationship between the behavior characteristics and the scientific research achievements by calculating the similarity of high-dimensional vectors, and the innovation characterization value is adjusted in real time through a dynamic weight mechanism to reflect the changes in the importance of the behavior characteristics.

[0028] Preferably, the innovation ability prediction model adopts a multi-task learning framework, combines behavioral characteristics and academic achievements for joint prediction, and the dynamic tracking is achieved through sensors and real-time computing platforms to ensure the timeliness of the data.

[0029] Preferably, the visualization report displays the dynamic changes of scientific research behavior and innovation ability in the form of radar charts and trend charts, and the optimization suggestions are generated by identifying weaknesses in behavioral characteristics, including learning behavior optimization, scientific research habit adjustment, and innovation training plans.

[0030] Preferably, the method can dynamically adjust the weights according to the behavioral patterns of students at different stages to improve prediction accuracy.

[0031] Preferably, the behavior analysis model combines image features and time series features to achieve a comprehensive evaluation of the dynamic scientific research process.

[0032] The present invention provides a big data-driven method for evaluating the innovation ability of college students in the entire scientific research process. It has the following beneficial effects:

[0033] This big data-driven evaluation method for college students' innovation ability in the entire scientific research process has formed a complete evaluation system from data collection to analysis, prediction and feedback. By introducing attendance punch-in equipment, central control console, school database and monitoring equipment in laboratory study rooms, the daily scientific research and learning behavior data of college students are comprehensively collected, and the data is cleaned and preprocessed using big data technology to ensure the comprehensive monitoring and security of the data in the entire scientific research process. Based on the intelligent behavior analysis model, this method can extract key behavior features from the images of the entire scientific research process, build a correlation model between behavior and results, and generate multi-dimensional semantic correlation feature vectors, realizing the accurate modeling of the relationship between behavior features and scientific research results. In addition, the innovation ability is characterized by a dynamic weight adjustment mechanism, and a prediction model is established in combination with historical academic data, which solves the problem of unclear quantitative evaluation standards for innovation ability in existing technologies.

[0034] Through dynamic tracking and real-time updating, the timeliness of students' scientific research learning process is guaranteed; by using multi-task learning framework and joint prediction of behavioral characteristics, the evaluation of innovation ability is more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a schematic diagram of the structure of the present invention. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] like Figure 1 As shown, an embodiment of the present invention provides a big data-driven method for evaluating the innovation ability of college students in the entire scientific research process, including: a. using attendance punching equipment, a central control console, a school database and monitoring equipment in a laboratory study room to collect college students' daily scientific research and learning behavior data, including laboratory attendance data, self-study time records and activity trajectories. The attendance punching equipment collects laboratory entry and exit records through student cards or biometric devices and uploads them to the central control console in real time. The school database is used to store students' personal information, course schedules and historical academic data. The monitoring equipment can record students' self-study time, equipment usage and learning content. The behavior analysis model is based on a multimodal learning method of big data, and dynamic weights are assigned to action features.

[0038] b. Through big data processing technology, the collected raw behavioral data is cleaned and preprocessed to form a standard data set. During the data cleaning process, the data is processed through an outlier detection algorithm. Three methods are combined: the triple standard deviation method based on statistics, the isolation forest algorithm based on machine learning, and the sliding window detection based on time series analysis. Multiple algorithms are combined to ensure the accuracy and completeness of the data and improve the quality of the generated standard data set.

[0039] c. Monitor and film the entire scientific research process, collect panoramic images of students' scientific research process, and process the images in combination with the intelligent behavior analysis model to extract scientific research behavior characteristics.

[0040] The behavior analysis model is:

[0041]

[0042] Among them, F action is the characteristic value of scientific research behavior, a i is the i-th action feature, w i is the weight of the action feature.

[0043] d. Jointly analyze standard data sets and scientific research behavior characteristics, generate daily learning behavior patterns, and build a behavior-achievement correlation model.

[0044] The behavior-outcome association model is:

[0045] R=cosine9V behavior ,V outcome 0

[0046] Among them, R is the correlation, V behavior and V outcome They represent the behavior feature vector and the achievement feature vector respectively. The behavior and achievement association model determines the relationship between behavior features and scientific research achievements by calculating the similarity of high-dimensional vectors. The innovation representation value is adjusted in real time through a dynamic weight mechanism to reflect the changes in the importance of behavior features. The behavior analysis model combines image features and time series features to achieve a comprehensive evaluation of the dynamic scientific research process. The behavior and achievement association model introduces a personalized weight adaptability mechanism. Based on the historical behavior data of individual students, the weight is dynamically adjusted through the following formula:

[0047]

[0048] Among them, w i is the initial weight, δ i is the individual performance deviation value of the ith behavioral characteristic, w′ i is the adjusted weight. This mechanism can optimize the performance of the association model according to individual differences of students.

[0049] e. Based on the behavior and achievement association model, a multi-dimensional semantic association feature vector is generated to quantify the innovation representation value in students' scientific research behavior.

[0050] The calculation formula of innovation characterization value is:

[0051]

[0052] Among them, I is the innovation capability characterization value, f j is the jth behavior characteristic value, w j The innovation capability prediction model adopts a multi-task learning framework to combine behavioral characteristics and academic achievements for joint prediction. Dynamic tracking is achieved through sensors and real-time computing platforms to ensure the timeliness of data. The innovation representation value is dynamically monitored through a time window mechanism. The length of the time window is calculated by the following formula:

[0053]

[0054] Where T is the time window length, N activities is the total number of scientific research activities, α is the adjustment coefficient, L semesterFor the length of the semester, through dynamic monitoring, the trend changes of students' scientific research behavior can be captured in real time, and the timeliness and accuracy of innovation ability evaluation can be improved.

[0055] f. Use the innovation representation value and historical academic data to establish an innovation capability prediction model. The innovation capability prediction formula is:

[0056]

[0057] Among them, Y is the estimated value of innovation capability, X k is the input feature, β k is the regression coefficient, ∈ is the error, and the innovation capability prediction model introduces a robustness verification mechanism during the construction process, and the stability and prediction accuracy of the model are evaluated by the cross-validation method. Noise data is introduced for perturbation testing to verify the model's tolerance to data anomalies, and the robustness of the model is quantified through model evaluation indicators such as mean square error and F1-score.

[0058] g. Dynamically track the scientific research and learning process of students, and update the estimated value of innovation ability in real time. The dynamic tracking part integrates data visualization components to generate trend charts and distribution charts in real time to show the changes in scientific research behavior, academic achievements and estimated values ​​of innovation ability. By showing the cyclical fluctuations of scientific research and learning behavior through intuitive charts, it provides a basis for dynamic monitoring and precise intervention for scientific research managers and students.

[0059] h. Generate a visualization report based on the estimated value of innovation ability and provide optimization suggestions to help students improve their scientific research and learning performance. The visualization report uses radar charts and trend charts to show the dynamic changes in scientific research behavior and innovation ability. Optimization suggestions are generated by identifying weaknesses in behavioral characteristics, including learning behavior optimization, scientific research habit adjustment, and innovation training plans. The method can dynamically adjust weights according to students' behavioral patterns at different stages to improve prediction accuracy.

[0060] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A big data-driven method for evaluating college students' innovation ability in the whole scientific research process, characterized by: include: a. Use attendance punching equipment, central control console, school database and monitoring equipment in laboratory study rooms to collect college students' daily scientific research and learning behavior data, including laboratory attendance data, study time records and activity tracks; b. Use big data processing technology to clean and preprocess the collected raw behavior data to form a standard data set; c. Monitor and shoot the entire scientific research process, collect panoramic images of students' scientific research process, and process the images in combination with intelligent behavior analysis models to extract scientific research behavior characteristics; The behavior analysis model is: Among them, F action is the characteristic value of scientific research behavior, a i is the i-th action feature, w i is the weight of the action feature; d. Jointly analyze standard data sets and scientific research behavior characteristics, generate daily learning behavior patterns, and build a behavior-result association model; The behavior-outcome association model is: R=cosine(V behavior ,V outcome ) Among them, R is the correlation, V behavior and V outcome They represent the behavior feature vector and the outcome feature vector respectively; e. Based on the behavior and achievement association model, generate multi-dimensional semantic association feature vectors to quantify the innovation representation value in students' scientific research behavior; The calculation formula of innovation characterization value is: Among them, I is the innovation capability characterization value, f j is the jth behavior characteristic value, w j is the feature weight; f. Using the innovation representation value and historical academic data, establish an innovation capability prediction model; the innovation capability prediction formula is: Among them, Y is the estimated value of innovation capability, X k is the input feature, β k is the regression coefficient, ∈ is the error; g. Dynamically track students’ scientific research and learning process and update the estimated value of innovation ability in real time; h. Generate visual reports based on the estimated value of innovation ability and provide optimization suggestions to help students improve their scientific research and learning performance.

2. According to claim 1, a big data-driven method for evaluating the innovation ability of college students in the whole scientific research process is characterized by: The attendance punching device collects laboratory entry and exit records through student cards or biometric devices and uploads them to the central control console in real time. The school database is used to store students' personal information, course schedules and historical academic data.

3. According to claim 1, a big data-driven method for evaluating the innovation ability of college students in the whole scientific research process is characterized by: The monitoring equipment can record students' self-study time, equipment usage and learning content. The behavior analysis model is based on a multimodal learning method of big data and assigns dynamic weights to action features.

4. According to claim 1, a big data-driven method for evaluating the innovation ability of college students in the whole scientific research process is characterized by: The behavior and achievement association model determines the relationship between behavior characteristics and scientific research achievements by calculating the similarity of high-dimensional vectors, and the innovation representation value is adjusted in real time through a dynamic weight mechanism to reflect the changes in the importance of behavior characteristics.

5. According to claim 1, a big data-driven method for evaluating the innovation ability of college students in the whole scientific research process is characterized by: The innovation ability prediction model adopts a multi-task learning framework and combines behavioral characteristics and academic achievements for joint prediction. The dynamic tracking is achieved through sensors and real-time computing platforms to ensure the timeliness of the data.

6. According to claim 1, a big data-driven method for evaluating the innovation ability of college students in the whole scientific research process is characterized by: The visualization report uses radar charts and trend charts to display the dynamic changes in scientific research behavior and innovation capabilities. The optimization suggestions are generated by identifying weaknesses in behavioral characteristics, including learning behavior optimization, scientific research habit adjustment, and innovation training plans.

7. According to claim 1, a big data-driven method for evaluating the innovation ability of college students in the whole scientific research process is characterized by: The method can dynamically adjust the weights according to the behavior patterns of students at different stages to improve the prediction accuracy.

8. According to claim 1, a big data-driven method for evaluating the innovation ability of college students in the whole scientific research process is characterized by: The behavior analysis model combines image features and time series features to achieve a comprehensive evaluation of the dynamic scientific research process.

Images