Entrepreneurship practical training platform optimization method

Abstract

The invention provides an entrepreneurship practical training platform optimization method. The entrepreneurship practical training platform optimization method comprises the steps of a, collecting student data, including learning behavior data, social interaction data, practical training project data and examination score data of students, and performing data cleaning and preprocessing, noise data removal and missing data filling, and b, based on historical learning data and personal characteristics of the students, carrying out data processing on the data, so as to obtain the entrepreneurship practical training platform. And constructing a student learning model by using a machine learning algorithm, wherein the learning model can dynamically predict the learning progress, the course demand and the personalized path recommendation of the student. According to the entrepreneurship practical training platform optimization method, by introducing a multi-dimensional data acquisition and processing technology, the problem of inaccurate learning path recommendation caused by incomplete data or poor quality in a traditional platform can be effectively solved. And through data cleaning and preprocessing, a noise data detection and missing data filling method is adopted to ensure the quality of the data, so that the accuracy of a recommendation algorithm is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of entrepreneurship training, and specifically to an optimization method for an entrepreneurship training platform. Background Art

[0002] The basic structure of an optimization method for an entrepreneurship training platform includes three core modules: a user management module, a course and training module, and a data analysis and feedback module. The user management module is responsible for user registration, login, and management of personal information, and supports permission management for different roles such as students, teachers, and administrators. The course and training module provides entrepreneurship-related courses, training content, and project displays, and supports online learning, case discussions, and project practical operations. The data analysis and feedback module collects students' learning data on the platform, including learning progress, exam scores, and project submission situations, conducts data analysis, and generates feedback reports to provide personalized learning guidance and advancement suggestions for students, and at the same time provides teaching effect evaluation for teachers.

[0003] Although the platform has strong personalized recommendation and data feedback capabilities, there are still some deficiencies. The platform relies on big data analysis to achieve optimization. However, for some students with incomplete or poor-quality data, the system's recommendation algorithm may be affected, resulting in inaccurate learning path planning. Although the platform can adjust according to data feedback, for the cultivation of some entrepreneurial qualities and practical abilities that are difficult to quantify, the system's support and optimization measures are still insufficient. Summary of the Invention

[0004] In view of the deficiencies of the prior art, the present invention provides an optimization method for an entrepreneurship training platform, which solves the problems that for some students with incomplete or poor-quality data, the system's recommendation algorithm may be affected, resulting in inaccurate learning path planning; and for the cultivation of some entrepreneurial qualities and practical abilities that are difficult to quantify, the system's support and optimization measures are still insufficient.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: An optimization method for an entrepreneurship training platform, including:

[0006] a. Collect students' data, including students' learning behavior data, social interaction data, training project data, and exam score data, and perform data cleaning and preprocessing to remove noise data and fill in missing data. The following formula is used for the detection and supplementation of noise data in the data cleaning and preprocessing:

[0007] D clean =D raw -Δ noise +Δ fill

[0008] where D cleanFor the data after cleaning, D raw For the original data, Δ noise For the noise data, Δ fill Is the part for filling in the missing data;

[0009] b. Based on the historical learning data and personal characteristics of the students, use machine learning algorithms to construct a student learning model. The learning model can dynamically predict the learning progress of students, course requirements, and personalized path recommendations. The relationship between the learning progress and path recommendations of students is expressed by the following formula for the learning model and the recommendation algorithm:

[0010]

[0011] Among them, f path Is the personalized learning path of the student, α i Is the weight coefficient of the course or project recommendation item, P i Is the recommendation item, and n is the number of recommendation items;

[0012] c. According to the learning model of the student and the preset course library and training project library of the platform, generate personalized learning path recommendations. The path recommendations include course content, project arrangements, and training tasks, and are dynamically adjusted according to the feedback and learning progress of the student;

[0013] d. During the learning process of the student, collect the learning progress and performance of the student in real time, and adjust the course content and project arrangements in the path recommendations to ensure that the recommended learning path can effectively match the abilities and needs of the student;

[0014] e. Based on the data analysis of the student, adjust the recommendation weight of the course content to ensure that it adapts to the needs of different students. The recommended adjustment model used is:

[0015]

[0016] Among them, W adjusted Is the adjusted course recommendation weight, W original Is the original course weight, λ is the adjustment factor, Δ performance Is the change in the learning performance of the student;

[0017] f. Introduce a social interaction analysis module. By analyzing the social interactions of students within the platform, evaluate the teamwork ability, communication ability, and social adaptation ability of students, and adjust the learning path and project arrangements of students in combination with these abilities;

[0018] g. Use the project training data, combined with expert scoring and peer evaluation, to evaluate the actual performance of students in entrepreneurial projects, and adjust the learning plan of students by analyzing these data. The practical ability evaluation formula is:

[0019]

[0020] Among them, A practical is the entrepreneurial practice ability of the students, F j is the feedback score for each project, and m is the number of feedback items;

[0021] h. Regularly evaluate the students' abilities through the entrepreneurial literacy and practice ability evaluation module, and comprehensively analyze the evaluation results with the students' learning progress, course feedback, and practice data to generate a personalized phased learning feedback report to guide the students to further optimize their learning plans;

[0022] i. Introduce an emotion analysis module to analyze the emotional fluctuations of the students during the learning process. Through the emotion analysis of the students' interaction content, provide personalized learning support for the students. The emotion analysis model is:

[0023]

[0024] Among them, S emotional is the student's emotional state score, T i is the emotional score for each interaction, θ i is the emotion weight, and k is the number of interactions;

[0025] j. Dynamically adjust the difficulty of the learning content, teaching strategies, and interaction forms according to the students' emotional feedback and social interaction data to enhance the students' learning motivation and participation;

[0026] k. Based on the comprehensive data of the students, including academic performance, practice ability, and emotional state, generate a feedback report on the development of comprehensive ability. The feedback report combines the students' academic ability, practice ability, and emotional state to provide optimization suggestions for the platform managers. The report formula is:

[0027] R comprehensive =α 1 ·A academic +α 2 ·A practical +α 3 ·S emotional

[0028] Among them, R comprehensive is the comprehensive report, A academic is the student's academic ability score, A practical is the practice ability score, S emotional is the emotional state score, α 1 ,α 2 ,α 3 are the weight coefficients.

[0029] Preferably, for data cleaning and preprocessing, a model-based filling method is used to fill in missing data, and interpolation or regression prediction methods are used for interpolation according to data characteristics.

[0030] Preferably, the entrepreneurship literacy and practical ability assessment module analyzes the entrepreneurship logs and reflection records of students through natural language processing technology, and combines sentiment analysis and text mining to judge the innovation ability and leadership performance of students.

[0031] Preferably, the sentiment analysis module can dynamically evaluate the emotional changes of students at different learning stages, and adjust task allocation and interaction methods according to the emotional state of students to ensure that students maintain a positive attitude during the learning process.

[0032] Preferably, when generating the path recommendation, it will combine the interests, emotional feedback and historical behavior data of students, and adopt an adaptive weight adjustment method to ensure the high personalization of the path recommendation.

[0033] Preferably, the social interaction analysis module builds a social network graph based on user behavior data, and further analyzes the social activity and teamwork ability of students.

[0034] Preferably, the feedback report generates dynamic charts based on the learning progress, emotional fluctuations, and practical feedback data of students, and presents them through visualization technology to help students and platform managers monitor and adjust the growth trajectory of students in real time.

[0035] Preferably, the machine learning algorithm is a learning model based on convolutional neural network and long short-term memory network, and the learning model can dynamically adjust the learning path of students according to historical behavior and real-time data.

[0036] The present invention provides an optimization method for an entrepreneurship training platform. It has the following beneficial effects:

[0037] This optimization method for the entrepreneurship training platform can effectively solve the problem of inaccurate learning path recommendation caused by incomplete or poor-quality data in traditional platforms by introducing multi-dimensional data collection and processing technology. Through data cleaning and preprocessing, noise data detection and missing data filling methods are used to ensure the quality of data, thereby ensuring the accuracy of the recommendation algorithm. On this basis, the student learning model constructed based on machine learning algorithms can predict the learning progress and needs of students in real time, and combine the preset course library and training project library of the platform to provide personalized learning path recommendations for students.

[0038] In addition to precise learning path recommendations, the present invention also comprehensively supports the improvement of the entrepreneurial qualities and practical abilities of trainees by introducing diversified ability assessment and sentiment analysis mechanisms. The system can not only comprehensively evaluate the entrepreneurial practice abilities of trainees through regular ability assessments and project training feedback, but also dynamically track the emotional fluctuations of trainees during the learning process through the sentiment analysis module, providing personalized learning support and emotional guidance for trainees. By combining the social interaction data of trainees, the system can evaluate their teamwork ability, communication ability, and social adaptability, and dynamically adjust the learning paths and task arrangements of trainees according to changes in these abilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a schematic diagram of the development trend of the comprehensive abilities of the trainees of the present invention;

[0040] Figure 2 It is a schematic diagram of the distribution of the learning progress and emotional state of the trainees of the present invention;

[0041] Figure 3 It is a schematic diagram of the distribution of the abilities of the trainees of the present invention in different dimensions. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0042] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0043] Embodiment 1

[0044] As Figures 1 - 3 shown, the embodiment of the present invention provides an optimization method for an entrepreneurial training platform, including: a. Collecting trainee data, including the learning behavior data, social interaction data, training project data, and exam score data of trainees, and performing data cleaning and preprocessing to remove noise data and fill in missing data. The following formula is used for the detection and supplementation of noise data in the data cleaning and preprocessing of path recommendations:

[0045] D clean = D raw - Δ noise + Δ fill

[0046] Wherein, D clean is the data after cleaning, D raw is the original data, Δ noise is the noise data, and Δ fillTo fill in the missing data, the path recommendation data cleaning and preprocessing uses a model-based filling method to fill in the missing data, and interpolation or regression prediction methods are used for interpolation according to the data characteristics.

[0047] b. Based on the historical learning data and personal characteristics of the trainees, a machine learning algorithm is used to construct a trainee learning model. The path recommendation learning model can dynamically predict the learning progress, course requirements, and personalized path recommendations of the trainees. The relationship between the trainee's learning progress and path recommendation is expressed by the following formula for the path recommendation learning model and recommendation algorithm:

[0048]

[0049] Among them, f path is the personalized learning path of the trainee, α i is the weight coefficient of the course or project recommendation item, P i is the recommendation item, n is the number of recommendation items. The path recommendation machine learning algorithm is a learning model based on convolutional neural network and long short-term memory network. The path recommendation learning model can dynamically adjust the trainee's learning path according to historical behavior and real-time data.

[0050] c. According to the trainee's learning model and the preset course library and training project library on the platform, generate personalized learning path recommendations. The path recommendation includes course content, project arrangement, and training tasks, and is dynamically adjusted according to the trainee's feedback and learning progress. When generating the path recommendation, it will combine the trainee's interests, emotional feedback, and historical behavior data, and adopt an adaptive weight adjustment method to ensure the high personalization of the path recommendation.

[0051] d. During the trainee's learning process, collect the trainee's learning progress and performance in real time, and adjust the course content and project arrangement in the path recommendation to ensure that the recommended learning path can effectively match the trainee's abilities and needs.

[0052] e. Based on the data analysis of the trainees, adjust the recommendation weight of the course content to ensure that it adapts to the needs of different trainees. The recommended adjustment model used is:

[0053]

[0054] Among them, W adjusted is the adjusted course recommendation weight, W original is the original course weight, λ is the adjustment factor, and Δ performance is the change in the trainee's learning performance.

[0055] f. Introduce a social interaction analysis module. By analyzing the social interactions of students within the platform, evaluate the students' teamwork ability, communication ability, and social adaptability, and adjust the students' learning paths and project arrangements based on these abilities. The path recommendation social interaction analysis module builds a social network graph based on user behavior data to further analyze the social activity and teamwork ability of students.

[0056] g. Utilize project training data, combine expert ratings and peer evaluations of classmates to evaluate the actual performance of students in entrepreneurial projects, and adjust the learning plans of students by analyzing these data. The practical ability evaluation formula is:

[0057]

[0058] Among them, A practical is the entrepreneurial practice ability of the student, F j is the score of each project feedback, and m is the number of feedback items.

[0059] h. Regularly evaluate the abilities of students through the entrepreneurial literacy and practical ability evaluation module, and comprehensively analyze the evaluation results with the students' learning progress, course feedback, and practical data to generate a personalized phased learning feedback report to guide the students to further optimize their learning plans. The path recommendation entrepreneurial literacy and practical ability evaluation module analyzes the entrepreneurial logs and reflection records of students through natural language processing technology, and combines sentiment analysis and text mining to judge the innovation ability and leadership performance of students.

[0060] i. Introduce a sentiment analysis module to analyze the emotional fluctuations of students during the learning process. Through the sentiment analysis of the interaction content of students, provide personalized learning support for students. The sentiment analysis model is:

[0061]

[0062] Among them, S emotional is the student's emotional state score, T i is the emotional score of each interaction, θ i is the emotional weight, and k is the number of interactions. The path recommendation sentiment analysis module can dynamically evaluate the emotional changes of students at different learning stages and adjust the task assignment and interaction methods according to the emotional state of students to ensure that students maintain a positive attitude during the learning process.

[0063] j. Dynamically adjust the difficulty of learning content, teaching strategies, and interaction forms according to the emotional feedback and social interaction data of students to enhance the learning motivation and participation of students.

[0064] k. Generate a feedback report on the comprehensive development of capabilities based on the comprehensive data of the trainees, including academic performance, practical ability, and emotional state. The path recommendation feedback report combines the academic ability, practical ability, and emotional state of the trainees to provide optimization suggestions for the platform administrators. The report formula is as follows:

[0065] R comprehensive =α 1 ·A academic +α 2 ·A practical +α 3 ·S emotional

[0066] Wherein, R comprehensive is the comprehensive report, A academic is the academic ability score of the trainee, A practical is the practical ability score, S emotional is the emotional state score, α 1 , α 2 , α 3 are the weight coefficients. The path recommendation feedback report generates a dynamic chart based on the learning progress, emotional fluctuations, and practical feedback data of the trainees, and presents it through visualization technology to help the trainees and platform administrators monitor and adjust the growth trajectory of the trainees in real time.

[0067] Example Two

[0068] This example optimizes the learning behavior, social interaction, and practical ability of the trainees based on the actual data of a university entrepreneurship training platform, and finally generates a personalized learning path and a comprehensive ability feedback report.

[0069] a. Data collection and cleaning

[0070] In a batch of entrepreneurship training courses, the data of 30 trainees were collected, including the following dimensions:

[0071]

[0072] After data collection, the following cleaning steps were carried out:

[0073] Noise detection: The recorded number of social interactions of trainee 1 was 0, which was significantly abnormal. After statistical analysis, it was supplemented to 10.

[0074] Missing data filling: The exam score of trainee 4 was missing. Using the regression prediction method, it was estimated to be 82 points based on the learning progress and project scores.

[0075] The formula for the cleaned data is as follows:

[0076] D clean =D raw -D noise +Dfill

[0077] b. Construction of personalized learning model

[0078] Based on the historical learning data of the trainees, such as learning progress, social interaction, and exam scores, construct a personalized recommendation model, and use the following formula to generate a personalized learning path:

[0079]

[0080] Where:

[0081] L path : The personalized learning path of the trainee;

[0082] α i : The recommendation weight, dynamically adjusted in combination with the trainee's interests and abilities;

[0083] P i : The recommended item or course content.

[0084] Example data (Trainee 1):

[0085] Historical data input: Learning progress 85%, social interaction 10 times, exam score 90 points;

[0086] The recommendation weights generated by the model:

[0087] Recommendation weight of basic courses α 1 = 0.3;

[0088] Recommendation weight of practical training tasks α 2 = 0.4;

[0089] Recommendation weight of teamwork α 3 = 0.3.

[0090] Generated path:

[0091] L path = 0.3·P 基础课程 + 0.4·P 实训任务 + 0.3·P 团队协作

[0092] c. Generation and optimization of dynamic learning path

[0093] The initial learning path of Trainee 2 is:

[0094] Basic courses: Entrepreneurial financial planning, market analysis;

[0095] Practical training tasks: Writing a project plan;

[0096] Teamwork: Team division of labor and task execution.

[0097] Dynamic adjusted path: During the learning process, trainee 2 showed a greater interest in the market analysis course, but had a lower completion rate of the project planning task. Therefore, the system redistributed the weights:

[0098] W 市场分析 = 0.5 (increase learning time);

[0099] W 项目策划 = 0.2 (reduce task difficulty).

[0100] Optimized learning path:

[0101] Key courses: Market analysis;

[0102] Practical tasks: The planning task is decomposed into simple modules;

[0103] Additional courses: Case analysis and report writing.

[0104] d. Data-driven optimization of course weights

[0105] Combined with the trainee performance data, adjust the course recommendation weights, the optimization formula:

[0106]

[0107] Where:

[0108] W adjusted : Adjusted weight;

[0109] W original : Original weight;

[0110] Δ performance : Change in trainee performance.

[0111] Example data:

[0112] The learning progress of trainee 3 increased from 60% to 75% (+15%), and the course recommendation weights were adjusted as follows:

[0113] Initial weight W 课程1 = 0.3;

[0114] Adjusted weight W 课程1 = 0.3·(1 + 0.15) = 0.345.

[0115] e. Team collaboration and social analysis

[0116] Construct a trainee social network diagram to analyze trainee team collaboration and social activity:

[0117] Student ID Teamwork Score Social Activity Level Student 1 88 High Student 2 70 Medium Student 3 95 High Student 4 65 Low

[0118] Based on data analysis, the system provides the following optimization measures for students with relatively weak teamwork abilities (such as student 4):

[0119] 1. Increase teamwork courses;

[0120] 2. Assign students to complete tasks jointly with classmates with high collaboration abilities.

[0121] f. Comprehensive ability feedback report

[0122] Generate a comprehensive ability feedback report based on students' academic performance, practical abilities, and emotional states. The formula is as follows:

[0123] R comprehensive = α 1 ·A academic + α 2 ·A practical + α 3 ·S emotional

[0124] Where:

[0125] R comprehensive : Comprehensive ability score;

[0126] A academic = 90 (academic ability score);

[0127] A practical = 85 (practical ability score);

[0128] S emotional = 80 (emotional state score);

[0129] Weight α 1 = 0.4, α 2 = 0.4, α 3 = 0.2.

[0130] Example report (student 1):

[0131] R comprehensive = 0.4·90 + 0.4·85 + 0.2·80 = 86

[0132] The comprehensive ability score of student 1 is 86 points. The platform recommends that they further improve their teamwork abilities.

[0133] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for optimizing an entrepreneurial training platform, characterized in that: include: a. Collect student data, including student learning behavior data, social interaction data, practical training project data and test score data, and perform data cleaning and preprocessing to remove noise data and fill in missing data. The data cleaning and preprocessing uses the following formula to detect and supplement noise data: D clean =D raw -D noise +D fill Among them, D clean is the cleaned data, D raw is the original data, Δ noise is the noise data, Δ fill To fill in the missing data; b. Based on the students’ historical learning data and personal characteristics, a machine learning algorithm is used to build a student learning model, which can dynamically predict the students’ learning progress, course requirements and personalized path recommendations. The learning model and recommendation algorithm express the relationship between the students’ learning progress and path recommendations through the following formula: Among them, f path Personalized learning path for students, i is the weight coefficient of the course or project recommendation, P i is the recommended item, n is the number of recommended items; c. Generate personalized learning path recommendations based on the student's learning model and the platform's preset course library and training project library. The path recommendations include course content, project arrangements, and training tasks, and are dynamically adjusted based on the student's feedback and learning progress; d. During the learning process, students’ learning progress and performance are collected in real time, and the course content and project arrangements in the path recommendation are adjusted to ensure that the recommended learning path can effectively match the students’ abilities and needs; e. Based on the data analysis of students, adjust the recommended weight of course content to ensure that it meets the needs of different students. The recommended adjustment model used is: Among them, W adjusted is the adjusted course recommendation weight, W original is the original course weight, λ is the adjustment factor, Δ performance Changes in students’ learning performance; f. Introduce a social interaction analysis module to evaluate students’ teamwork, communication and social adaptability by analyzing their social interactions on the platform, and adjust students’ learning paths and project arrangements based on these abilities; g. Use project training data, combined with expert ratings and peer evaluation, to evaluate the actual performance of students in entrepreneurial projects, and adjust the students' learning plans by analyzing these data. The practical ability evaluation formula is: Among them, A practical To enhance students’ entrepreneurial practice ability, j is the feedback score for each item, and m is the number of feedback items; h. Conduct regular ability assessments on students through the entrepreneurial literacy and practical ability assessment modules, and conduct comprehensive analysis of the assessment results with the students’ learning progress, course feedback, and practical data to generate personalized periodic learning feedback reports to guide students to further optimize their learning plans; i. Introduce sentiment analysis module to analyze students’ emotional fluctuations during the learning process. Through sentiment analysis of students’ interactive content, provide students with personalized learning support. The sentiment analysis model is: Among them, S emotional Score the students’ emotional state, T i is the sentiment score of each interaction, θ i is the sentiment weight, k is the number of interactions; j. Dynamically adjust the difficulty of learning content, teaching strategies and interaction forms based on students’ emotional feedback and social interaction data to enhance students’ learning motivation and participation; k. Based on the comprehensive data of students, including academic performance, practical ability, and emotional state, generate a feedback report on comprehensive ability development. The feedback report combines the students' academic ability, practical ability, and emotional state to provide optimization suggestions for platform managers. The report formula is: R comprehensive =α1·A academic +α2·A practical +α3·S emotional Among them, R comprehensive For comprehensive reporting, A academic The score for the student's academic ability is A. practical S is the score for practical ability. emotional is the emotional state score, and α1, α2, α3 are weight coefficients.

2. The method for optimizing an entrepreneurial training platform according to claim 1, characterized in that: The data cleaning and preprocessing adopts a model-based filling method to fill in the missing data, and adopts an interpolation method or a regression prediction method to interpolate according to data characteristics.

3. The method for optimizing an entrepreneurial training platform according to claim 1, characterized in that: The entrepreneurial literacy and practical ability assessment module analyzes the students' entrepreneurial diaries and reflection records through natural language processing technology, and combines sentiment analysis and text mining to judge the students' innovation ability and leadership performance.

4. The method for optimizing an entrepreneurial training platform according to claim 1, characterized in that: The sentiment analysis module can dynamically evaluate the emotional changes of students at different learning stages, and adjust task allocation and interaction methods according to the emotional state of the students, ensuring that the students maintain a positive attitude during the learning process.

5. The method for optimizing an entrepreneurial training platform according to claim 1, characterized in that: When generating the path recommendation, the student's interests, emotional feedback, and historical behavior data are combined, and an adaptive weight adjustment method is used to ensure that the path recommendation is highly personalized.

6. The method for optimizing an entrepreneurial training platform according to claim 1, characterized in that: The social interaction analysis module establishes a social network diagram based on user behavior data to further analyze the students' social activity and teamwork ability.

7. The method for optimizing an entrepreneurial training platform according to claim 1, characterized in that: The feedback report generates dynamic charts based on the students' learning progress, emotional fluctuations, and practical feedback data, and presents them through visualization technology to help students and platform managers monitor and adjust the students' growth trajectory in real time.

8. The method for optimizing an entrepreneurial training platform according to claim 1, characterized in that: The machine learning algorithm is a learning model based on convolutional neural networks and long short-term memory networks. The learning model can dynamically adjust the student's learning path based on historical behavior and real-time data.

Images