Skill learning data analysis system based on artificial intelligence

Abstract

The invention relates to the technical field of big data, and particularly discloses a skill learning data analysis system based on artificial intelligence, and the system comprises a video learning module which obtains the watching frequency and the single learning duration of a target course, and calculates the mean value and the learning coefficient of the single learning duration; the live broadcast module obtains the live broadcast times of the target course and the live broadcast watching times of the user, and calculates a live broadcast coefficient; the score module is used for acquiring user score data and calculating a comprehensive score based on the score data; and the correction module is used for correcting the comprehensive score based on the learning coefficient and the live broadcast coefficient, obtaining a corrected score, obtaining a user's check-out score, calculating a skill level coefficient of the user based on the check-out score and the corrected score, carrying out sorting according to the skill level coefficient, and analyzing the learning condition of the user. The invention provides a skill learning data analysis system based on artificial intelligence. The learning effect of a platform user is analyzed according to the learning condition of the platform user.

Description

Technical Field

[0001] The present invention relates to the field of big data technology, and in particular to an artificial intelligence-based skill learning data analysis system. Background Art

[0002] Due to the continuous development of technology and changing job requirements, many job seekers may lack the skills required for certain roles, causing difficulties in their job search. Conversely, many employers are also seeking candidates with specific skills, but due to fierce competition or difficulty in recruiting, they may find it difficult to find suitable candidates. This results in some job vacancies remaining unfilled, impacting the operation and development of enterprises.

[0003] Therefore, the Skills Learning Platform emerged as a new online vocational education and training platform in the mobile internet era. Leveraging portals and WeChat, the Skills Learning Platform expands the reach of online vocational skills training, targeting employees on hold, returning to work, and currently employed, as well as unemployed university graduates and key employment groups. Students can choose the appropriate learning device based on their specific needs. The system records their learning progress based on their account, allowing them to confidently learn online vocational training courses.

[0004] In terms of current regulation, skill learning platforms are unable to provide multiple regulatory measures and lack visibility into the learning progress of platform students, making it impossible to ensure effective and rigorous training supervision. Because skill learning utilizes online learning methods, it is impossible to understand the user's specific learning status and lacks knowledge of the student's situation, making it uncertain about the student's learning situation. Therefore, in order to fully understand the learning situation of each user, an AI-based skill learning data analysis system is needed to analyze the learning outcomes of platform users based on their learning situation. Summary of the Invention

[0005] The purpose of this invention is to provide an artificial intelligence-based skill learning data analysis system to solve the above technical problems.

[0006] The purpose of the present invention can be achieved through the following technical solutions: An artificial intelligence-based skill learning data analysis system, comprising: Video learning module: Get the courses that the user has registered for and record them as target courses, get the number of times the user watches the target course N and the single learning time T, and calculate the average single learning time T ave , calculate the learning coefficient , where T n represents the nth single learning time, and γ represents the preset learning time threshold; Live broadcast module: Get the number of live broadcasts of the target course M and the number of times users watch the live broadcast m, and calculate the live broadcast coefficient , where C k represents the total communication time between the user and the lecturer during the kth live broadcast, TC k Represents the total duration of the k-th live broadcast; Grade module: obtains the user's grade data, including test grade G and homework grade H, and calculates the comprehensive grade , where I and J represent the number of test scores and the number of homework scores respectively, and G all , H all分别 Represents full marks for tests and assignments, G i represents the i-th test score, H j represents the j-th homework score, λ1 and λ2 represent the first and second preset coefficients respectively, and λ1>λ2>1; Correction module: Corrects the comprehensive score Z based on the learning coefficient X and the live broadcast coefficient ZB, and records the corrected comprehensive score as the corrected score XZ; Obtain the user's completion score JY, calculate the user's skill level coefficient SK based on the completion score JY and the revised score XZ, sort the users from high to low according to their skill level coefficients, and analyze the user's learning status based on the sorting.

[0007] As a further solution of the present invention: in the video learning module, a learning time threshold T is preset. min , if the single learning time T≤T min , this learning is recorded as invalid learning, and invalid learning is not counted in the number N of times the user has learned the target course.

[0008] As a further solution of the present invention: in the video learning module, the calculation method of the learning time threshold γ includes: Calculate the learning time difference ΔT=|TT ave |, get the maximum value of the learning time difference ΔT max , let the learning time threshold γ = (ΔT max -T ave ) 2 .

[0009] As a further solution of the present invention: in the correction module, if there are users with equal skill level coefficients SK, the user with a higher completion score JY is placed at a higher position in the ranking.

[0010] As a further solution of the present invention: in the live broadcast module, if the time a user spends watching a target course is less than a preset standard time, this viewing behavior will not be counted in the number of live broadcasts M.

[0011] As a further solution of the present invention: In the correction module, the method for correcting the comprehensive score Z based on the learning coefficient X and the live broadcast coefficient ZB includes: Calculate corrected learning effect , where X ave Represents the mean of the learning coefficients.

[0012] As a further solution of the present invention: In the correction module, the method for calculating the user's skill level coefficient SK based on the completion score JY and the corrected score XZ includes: The skill level coefficient SK is calculated by the formula, which is specifically: ; Here, η1 represents a preset first weight value, η2 represents a preset second weight value, η1>0, η2>0 and η1+η2=1.

[0013] As a further solution of the present invention: in the correction module, the method of sorting users from high to low according to their skill level coefficients and analyzing the user learning status based on the sorting includes: The top 30% of users in the ranking are recorded as excellent users, and the average of the excellent users' completion scores is calculated. ave and the mean of the corrected scores XZ ave , if JY≥JY ave , XZ≥XZ ave , it means that the user is in good learning condition and will be awarded a skill certificate; If JY<JY ave , XZ≥XZ ave , it means that the user's actual operation ability is unqualified and the skill certificate will not be granted; If JY≥JY ave , XZ<XZ ave , it means that the user's theoretical knowledge is not qualified and the skill certificate will not be granted; If JY<JY ave , XZ<XZ ave , they will be recorded as key users and the platform staff will be prompted to communicate with them.

[0014] The beneficial effects of the present invention are as follows: first, a preliminary judgment is made on the user's learning effect based on the length of time the user watches the target course online. According to objective laws, the user's learning effect is related to the number of times the user watches the target course and the length of a single learning session. The average length of a single viewing session is calculated to determine the user's optimal learning session. According to the formula for calculating the learning coefficient, it can be seen that when a single learning session is too long or too short, it will affect the user's learning effect. The larger the learning coefficient, the better the learning effect.

[0015] It should be noted that live broadcast activities will be carried out for target courses on this platform to help users answer questions and resolve doubts. Therefore, based on the number of times users participate in live broadcasts and the length of time they watch, we can also make a rough judgment on their learning situation. The number of times users participate in live broadcasts is a factor affecting the live broadcast coefficient, and the length of time users communicate with instructors is also an important parameter affecting the live broadcast coefficient, and both are proportional to the live broadcast coefficient. The more times users discuss with instructors and the longer the time, the better the effect on users' skill learning.

[0016] In addition, the platform will regularly test and assign homework to users. Based on the test scores and homework scores, the user's learning status can be further judged, and then a comprehensive score can be calculated. The comprehensive score is used to reflect the user's learning status of the target course. The higher the comprehensive score, the better the user's learning effect on the target course. In the formula, the proportion of test scores is greater than that of homework scores, because the difficulty of the tests assigned by the platform is greater than the difficulty of homework, so it is in line with the actual situation.

[0017] Finally, the comprehensive score needs to be corrected based on the learning coefficient and the live broadcast coefficient to obtain the corrected score. The corrected score represents the user's mastery of the theoretical level of the target course. The higher the corrected score, the more proficient the theoretical mastery of the target course. The final score is the score made by the platform based on the user's actual operation ability. In the final analysis process, it is necessary to comprehensively consider the situation between the two and analyze them to determine the user's mastery of the target course and analyze their learning situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The present invention will be further described below with reference to the accompanying drawings.

[0019] Figure 1 It is a structural diagram of an artificial intelligence-based skill learning data analysis system of the present invention. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making any creative efforts shall fall within the scope of protection of the present invention.

[0021] See also Figure 1 As shown, the present invention is a skill learning data analysis system based on artificial intelligence, comprising: Video learning module: Get the courses that the user has registered for and record them as target courses, get the number of times the user watches the target course N and the single learning time T, and calculate the average single learning time T ave , calculate the learning coefficient , where T n represents the nth single learning time, and γ represents the preset learning time threshold; Live broadcast module: Get the number of live broadcasts of the target course M and the number of times users watch the live broadcast m, and calculate the live broadcast coefficient , where C k represents the total communication time between the user and the lecturer during the kth live broadcast, TC k Represents the total duration of the k-th live broadcast; Grade module: obtains the user's grade data, including test grade G and homework grade H, and calculates the comprehensive grade , where I and J represent the number of test scores and the number of homework scores respectively, and G all , H all分别 Represents full marks for tests and assignments, G i represents the i-th test score, H j represents the j-th homework score, λ1 and λ2 represent the first and second preset coefficients respectively, and λ1>λ2>1; Correction module: Corrects the comprehensive score Z based on the learning coefficient X and the live broadcast coefficient ZB, and records the corrected comprehensive score as the corrected score XZ; Obtain the user's completion score JY, calculate the user's skill level coefficient SK based on the completion score JY and the revised score XZ, sort the users from high to low according to their skill level coefficients, and analyze the user's learning status based on the sorting.

[0022] It's important to note that when initially evaluating a user's learning outcomes based on the amount of time they spend watching the target course online, it's important to consider multiple objective factors. Research shows that learning outcomes are primarily influenced by two key factors: the total number of times a user views the target course (learning frequency), and the duration of each session (learning intensity). To more accurately assess learning outcomes, we first need to calculate the average duration of each session, a metric that reflects a user's learning habits and attention span.

[0023] Further analysis reveals that there is an optimal range for learning duration, which can be quantified by calculating the learning coefficient. The learning coefficient calculation formula comprehensively considers the appropriateness of a single learning session, and its curve exhibits an inverted U-shaped characteristic: when a single learning session is too short, users struggle to deeply understand the course content; whereas, when a single learning session is too long, excessive cognitive load leads to decreased learning efficiency. Both scenarios negatively impact the ultimate learning outcome.

[0024] Specifically, the calculation of the learning coefficient assigns a higher weight to learning behaviors within a medium-length interval, as this interval best conforms to the objective laws of human cognition. Big data analysis shows that when a user's learning time falls within this optimal range, their knowledge absorption rate, memory retention, and depth of understanding all reach optimal levels. Therefore, as a quantitative indicator, the larger the learning coefficient, the more reasonable the user's learning strategy and the better the ultimate learning effect. This coefficient can not only be used to evaluate individual learning outcomes, but also provide course designers with a scientific basis for optimizing course duration and content distribution.

[0025] In today's digital learning landscape, this platform is committed to providing users with a comprehensive and high-quality learning experience. We've launched a series of livestreaming events for specific target courses. As users progress through their target courses, they inevitably encounter various questions and challenges. Livestreaming is crucial. Through livestreaming, users can raise questions about their learning and practice with instructors in real time. Whether it's about understanding course content or navigating practical application scenarios, they can receive prompt and professional answers. This instant, interactive communication allows users to gain a deeper understanding of the course content and resolve any challenges promptly.

[0026] For this reason, user participation is crucial for evaluating learning outcomes. Specifically, based on the number of times a user participates in a live broadcast and the duration of their viewing, we can provide a relatively accurate and comprehensive assessment of their learning progress.

[0027] The number of times a user participates in a live broadcast is a key factor influencing the live broadcast coefficient. Each participation in a live broadcast signifies that the user is actively engaged in learning, eager to acquire more knowledge and skills. Frequent participation not only allows users to promptly address current learning issues but also exposes them to more expansive knowledge and case studies, broadening their knowledge and perspectives. For example, some users may only be able to participate in a live broadcast occasionally due to busy work or personal commitments, resulting in a relatively limited understanding and grasp of the course content. However, users who frequently participate in live broadcasts, due to the constant exposure to new knowledge and interaction with the instructor, will have a more thorough understanding of the course and a more comprehensive knowledge base.

[0028] The duration of user-instructor communication is also a significant factor influencing the livestream coefficient, with both being directly proportional to the coefficient. During livestreams, communication and interaction between users and instructors is a crucial aspect of learning. More frequent and longer discussions with instructors provide more opportunities for users to delve deeper into difficult and key issues within the course. They can engage in thorough discussions with instructors on specific knowledge points, exploring and analyzing them from diverse perspectives and ultimately gaining deeper insights and understanding. This in-depth communication not only helps users resolve current learning challenges but also cultivates their independent thinking and problem-solving skills.

[0029] For example, in complex programming courses, users may encounter various coding errors and logic problems during practical operation. Through lengthy discussions with instructors, they can guide users through step-by-step analysis of the root causes of the problems and help them find the right solutions. In this process, users not only learn how to solve the current problems but also master methods and techniques for debugging and optimizing code, which will be of great benefit to their future learning and practice.

[0030] Furthermore, the longer the discussion time with the instructor, the more opportunities users have to understand the instructor's way of thinking and problem-solving approaches, allowing them to better absorb and draw on the instructor's experience and wisdom. This subtle influence is crucial to the effectiveness of user skill learning. It helps users gradually develop their own learning methods and styles, improving learning efficiency and quality, and ultimately achieving the transformation and improvement of knowledge into skills. The number of times users participate in live broadcasts and the length of time they communicate with the instructor are important indicators for evaluating user learning progress and the live broadcast coefficient, and they have a direct positive impact on the user's skill learning results.

[0031] To understand users' progress in their target courses, the platform regularly organizes quizzes and assigns homework. This periodic assessment is an important means for the platform to verify user learning outcomes and a key step in understanding their learning status.

[0032] The quiz platform systematically assesses users' mastery of course content, their ability to construct a knowledge system, and their ability to apply acquired knowledge to solve real-world problems. Quizzes cover key and difficult areas of the course, aiming to comprehensively assess users' overall learning performance. Assignments focus on consolidating and applying classroom knowledge. Through diverse question types and task requirements, users are encouraged to deepen their understanding through practical application.

[0033] Based on test and assignment scores, the platform can further accurately assess a user's learning status. These two scores act like two mirrors, clearly reflecting a user's strengths and weaknesses in their learning process, providing valuable feedback to both the platform and the user. Based on this, the platform uses scientific methods to calculate each user's overall score. This overall score isn't simply the sum of test and assignment scores; it's the result of careful consideration and rigorous weighing, providing a more objective and comprehensive reflection of the user's overall learning progress in the target course.

[0034] In the formula for calculating the overall score, test scores account for a greater proportion than homework scores. This designation is based on extensive teaching practice and a deep understanding of learning patterns, and is fully consistent with reality. The tests assigned by the platform are more difficult than homework assignments in terms of question difficulty, depth of knowledge, and comprehensive ability assessment. Test questions often place greater emphasis on the flexible application of knowledge points and comprehensive analysis, requiring users to possess advanced thinking skills and problem-solving techniques.

[0035] Precisely because the difficulty of tests is greater than that of assignments, test scores have a greater weight in the overall score and can better reflect the user's true level and ability in the learning process. The higher the overall score, the better the user's learning effect on the target course, indicating that they have performed well in mastering and applying knowledge and improving their comprehensive literacy. On the contrary, users with lower overall scores need to conduct targeted learning and improvement on their weak links to improve their learning effects. This comprehensive evaluation system based on test scores and assignment scores provides the platform and users with a scientific and objective learning assessment tool, which helps to promote users' continuous progress and improve the quality of learning.

[0036] During the final analysis, it's necessary to comprehensively consider both the revised and final grades for a comprehensive and in-depth analysis. This comprehensive analysis provides users with a comprehensive, multi-faceted learning assessment, accurately assessing their overall mastery of the target course. If a user's revised grade is high but their final grade is unsatisfactory, this may mean that they excel in theoretical knowledge but have shortcomings in applying theory to practice, perhaps due to a lack of practical experience or a need to improve their problem-solving skills. Conversely, if the final grade is strong but the revised grade is relatively low, this suggests that the user has a certain talent and ability in practical application, but needs to further strengthen their theoretical knowledge to better support their practical application.

[0037] The platform's course resources are developed in accordance with national vocational skill standards, specialized professional competency specifications, and vocational training curriculum specifications, conforming to the principles of vocational skill training and instruction. Furthermore, the Skills Pass platform has developed a live streaming feature to adapt to the live streaming era. This platform regularly organizes live streams focusing on classes to explain relevant course content. These live streams, featuring interactive teacher-student and student-student interactions, closely replicate a real classroom environment. To better serve users, the platform has designed a user-friendly and multifunctional personal center to help students learn efficiently.

[0038] In another preferred embodiment of the present invention, a learning time threshold T is preset. min , if the single learning time T≤T min , this learning is recorded as invalid learning, and invalid learning is not counted in the number N of times the user has learned the target course.

[0039] It is worth noting that when the single learning time is less than the learning time threshold, the user may enter the target course due to accidental touch. In order to ensure the accuracy of the final analysis results, the invalid learning records will be eliminated and will not be involved in subsequent calculations.

[0040] In another preferred embodiment of the present invention, the method for calculating the learning time threshold γ includes: Calculate the learning time difference ΔT=|TT ave |, get the maximum value of the learning time difference ΔT max , let the learning time threshold γ = (ΔT max -T ave ) 2 .

[0041] It is understandable that the purpose of determining the learning threshold is to reduce analysis errors caused by the occurrence of abnormal data, thereby further improving the accuracy of the final analysis results.

[0042] In another preferred embodiment of the present invention, if there are users with the same skill level coefficient SK, the user with the higher completion score JY is placed at a higher position in the ranking.

[0043] It should be noted that when there are users with equal skill level coefficients SK, users with higher graduation scores JY are placed at the top of the ranking, because hands-on ability is more important than theoretical ability in actual practice, and this further improves the system's ability to resist extreme situations.

[0044] In another preferred embodiment of the present invention, if the time a user spends watching the target course is less than the preset standard time, this viewing behavior will not be counted in the number of live broadcasts M.

[0045] It should be noted that when a user's viewing time falls short of the preset threshold, it means they may have simply skimmed the course content without engaging in in-depth, systematic learning and reflection. In this case, counting their viewing behavior as part of the live broadcast count clearly fails to truly reflect the user's actual learning progress and violates the platform's requirements for the accuracy and validity of learning data.

[0046] In another preferred embodiment of the present invention, the method for correcting the comprehensive score Z based on the learning coefficient X and the live broadcast coefficient ZB includes: Calculate corrected learning effect , where X ave Represents the mean of the learning coefficients.

[0047] It is understandable that the learning effect is corrected when the learning coefficient X is greater than X ave At this time, the user's learning intensity is higher than the average, so multiplying the two will increase the value of the comprehensive score.

[0048] In another preferred embodiment of the present invention, the method for calculating the user's skill level coefficient SK based on the final score JY and the revised score XZ includes: The skill level coefficient SK is calculated by the formula, which is specifically: ; Here, η1 represents a preset first weight value, η2 represents a preset second weight value, η1>0, η2>0 and η1+η2=1.

[0049] It is worth noting that the skill level coefficient is calculated through a weighted formula, and the sum of the two weighted coefficients is equal to 1. Note that in actual situations, the first weight value and the second weight value should be set according to the importance of theory and actual operation.

[0050] In another preferred embodiment of the present invention, the method of sorting users from high to low according to their skill level coefficients and analyzing the user learning status based on the sorting includes: The top 30% of users in the ranking are recorded as excellent users, and the average of the excellent users' completion scores is calculated. ave and the mean of the corrected scores XZ ave , if JY≥JY ave , XZ≥XZ ave , it means that the user is in good learning condition and will be awarded a skill certificate; If JY<JY ave , XZ≥XZ ave , it means that the user's actual operation ability is unqualified and the skill certificate will not be granted; If JY≥JY ave , XZ<XZ ave, it means that the user's theoretical knowledge is not qualified and the skill certificate will not be granted; If JY<JY ave , XZ<XZ ave , they will be recorded as key users and the platform staff will be prompted to communicate with them.

[0051] The above is a detailed description of an embodiment of the present invention. However, the content is only a preferred embodiment of the present invention and should not be considered to limit the scope of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A skill learning data analysis system based on artificial intelligence, characterized in that: include: Video learning module: Get the courses that the user has registered for and record them as target courses, get the number of times the user watches the target course N and the single learning time T, and calculate the average single learning time T ave , calculate the learning coefficient , where T n represents the nth single learning time, and γ represents the preset learning time threshold; Live broadcast module: Get the number of live broadcasts of the target course M and the number of times users watch the live broadcast m, and calculate the live broadcast coefficient , where C k represents the total communication time between the user and the lecturer during the kth live broadcast, TC k Represents the total duration of the k-th live broadcast; Grade module: obtains the user's grade data, including test grade G and homework grade H, and calculates the comprehensive grade , where I and J represent the number of test scores and the number of homework scores respectively, and G all , H all分别 Represents full marks for tests and assignments, G i represents the i-th test score, H j represents the j-th homework score, λ1 and λ2 represent the first and second preset coefficients respectively, and λ1>λ2>1; Correction module: Corrects the comprehensive score Z based on the learning coefficient X and the live broadcast coefficient ZB, and records the corrected comprehensive score as the corrected score XZ; Obtain the user's completion score JY, calculate the user's skill level coefficient SK based on the completion score JY and the revised score XZ, sort the users from high to low according to their skill level coefficients, and analyze the user's learning status based on the sorting.

2. The skill learning data analysis system based on artificial intelligence according to claim 1, characterized in that: In the video learning module, a learning time threshold T is preset. min , if the single learning time T≤T min , this learning is recorded as invalid learning, and invalid learning is not counted in the number N of times the user has learned the target course.

3. The skill learning data analysis system based on artificial intelligence according to claim 1, characterized in that: In the video learning module, the calculation method of the learning time threshold γ includes: Calculate the learning time difference ΔT=|TT ave |, get the maximum value of the learning time difference ΔT max , let the learning time threshold γ = (ΔT max -T ave ) 2 .

4. The skill learning data analysis system based on artificial intelligence according to claim 1, characterized in that: In the correction module, if there are users with the same skill level coefficient SK, the user with the higher completion score JY is placed at the top of the ranking.

5. The skill learning data analysis system based on artificial intelligence according to claim 1 is characterized in that: In the live broadcast module, if the time a user spends watching the target course is less than the preset standard time, this viewing behavior will not be counted in the live broadcast times M.

6. The skill learning data analysis system based on artificial intelligence according to claim 1, characterized in that: In the correction module, the method for correcting the comprehensive score Z based on the learning coefficient X and the live broadcast coefficient ZB includes: Calculate corrected learning effect , where X ave Represents the mean of the learning coefficients.

7. The skill learning data analysis system based on artificial intelligence according to claim 1, characterized in that: In the correction module, the method for calculating the user's skill level coefficient SK based on the completion score JY and the correction score XZ includes: The skill level coefficient SK is calculated by the formula, which is specifically: ； Here, η1 represents a preset first weight value, η2 represents a preset second weight value, η1>0, η2>0 and η1+η2=1.

8. The skill learning data analysis system based on artificial intelligence according to claim 1, characterized in that: In the correction module, the method of sorting users from high to low according to their skill level coefficients and analyzing the user learning status based on the sorting includes: The top 30% of users in the ranking are recorded as excellent users, and the average of the excellent users' completion scores is calculated. ave and the mean of the corrected scores XZ ave , if JY≥JY ave , XZ≥XZ ave , it means that the user is in good learning condition and will be awarded a skill certificate; If JY<JY ave , XZ≥XZ ave , it means that the user's actual operation ability is unqualified and the skill certificate will not be granted; If JY≥JY ave , XZ<XZ ave , it means that the user's theoretical knowledge is not qualified and the skill certificate will not be granted; If JY<JY ave , XZ<XZ ave , they will be recorded as key users and the platform staff will be prompted to communicate with them.

Images