Adolescents' digital media literacy intervention system and method based on big data

Abstract

The present invention discloses a digital media literacy intervention system and method for adolescents based on big data, which relates to the field of learning intervention technology, recommends corresponding learning resources for students, collects learning status data to generate reminder values, issues learning reminders to the outside when the reminder values ​​are lower than expected, generates reminder density from the state data of received learning reminders, and when the reminder density is higher than expected, performs data analysis on the performance feedback data, generates target completion from the learning feedback values ​​before and after the intervention in the student learning process, and if the target completion does not exceed the preset completion threshold, obtains the influencing factors affecting the target completion by principal component analysis, optimizes the learning resource push strategy, executes instant incentive feedback measures or outputs optimization suggestions for optimizing the intervention path. Customize learning resources for students so that students can obtain more matching learning resources on the current learning path.

Description

Technical Field

[0001] The present invention relates to the technical field of learning intervention, and in particular to a system and method for intervening in digital media literacy for adolescents based on big data. Background Art

[0002] The online learning platform is a digital learning environment that integrates diverse educational resources and learning tools. It makes full use of the advantages of Internet technology and breaks the time and space limitations of traditional education. On this platform, learners can access rich course content from all over the world anytime and anywhere, including high-quality video lectures, lively and interesting interactive tutorials, real-time updated online tests and exercise books, etc. The online learning platform also provides convenient learning management tools, such as learning progress tracking, performance analysis, online discussions, etc., to promote communication and cooperation among learners and improve learning effects. Through the online learning platform, people can acquire knowledge and skills more flexibly and efficiently, and promote the widespread dissemination of knowledge and the development of lifelong learning.

[0003] Learning intervention is a method of deeply analyzing students' learning behavior data on online learning platforms, and then adopting personalized and precise teaching strategies. Through diversified means such as customized learning plans, intelligent recommendation of learning resources, targeted learning skills training, and real-time learning progress tracking and feedback, it aims to effectively identify and solve the difficulties and challenges encountered by students in the learning process, fully stimulate students' internal learning motivation, improve learning efficiency and quality, and promote students' all-round development and personalized growth.

[0004] In the Chinese invention patent with the authorization announcement number CN109635869B, an online learning intervention system is disclosed, which includes a data collection module, a group division module, and a strategy library formation module. The data collection module is used to collect several copies of historical data; the group division module is used to divide the historical data into multiple classification groups and generate classification models for each learning stage; the strategy library formation module is used to abstract the group characteristics and intervention measures rules of each classification group, and obtain the strategy set of each classification group. The strategy sets of all classification groups constitute the learner strategy library. The generated strategy library uses the experience of all learners in the same learning stage of the same course in the past, which is not easy to cause evaluation bias.

[0005] Combining the above applications with the contents of the prior art;

[0006] When students are in a self-study state or studying at home, they usually need to use online learning platforms. However, considering the lack of supervision in independent learning, students may find it difficult to achieve their expected learning goals. Therefore, it is necessary to collect learning feedback data regularly and evaluate the students' status based on the learning feedback data so as to adjust the current learning status in time.

[0007] In the existing intervention methods for digital media literacy of adolescents, big data and new media are usually introduced when students conduct independent learning through online learning platforms. New media and big data are used to recommend corresponding learning resources to students, and personalized customization of learning resources is achieved, so as to improve learning efficiency and improve learning feedback. However, when students are learning independently, in the existing intervention methods, the intervention path for students' learning is usually determined by the students' current learning feedback data. When there is a major change in the students' learning status, the pre-constructed intervention path and the change of intervention methods will have a large lag. When there is a deviation in the personalized customization strategy of learning resources and the pushed content, it cannot be adjusted quickly, which easily leads to the existing intervention methods failing to achieve the expected improvement effect.

[0008] To this end, the present invention provides a digital media literacy intervention system and method for teenagers based on big data. Summary of the invention

[0009] 1. Technical issues to be resolved

[0010] In view of the deficiencies of the prior art, the present invention provides a digital media literacy intervention system and method for adolescents based on big data, which recommends corresponding learning resources to students, collects learning status data to generate reminder values, issues learning reminders to the outside when the reminder values ​​are lower than expected, generates reminder density from the status data of received learning reminders, and performs data analysis on the performance feedback data when the reminder density is higher than expected, generates target completion from the learning feedback values ​​before and after the intervention in the student learning process, and if the target completion does not exceed the preset completion threshold, obtains the influencing factors affecting the target completion by principal component analysis, optimizes the learning resource push strategy, executes instant incentive feedback measures or outputs optimization suggestions for the optimization of the intervention path. Customized learning resources for students enable students to obtain more matching learning resources on the current learning path; thereby solving the technical problems recorded in the background technology.

[0011] (II) Technical solution

[0012] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0013] The intervention method for adolescent digital media literacy based on big data includes: collecting students' learning feedback data to generate students' learning feedback values, predicting and analyzing the trend of learning feedback values, and generating student learning files based on students' learning process data and learning goals if the change trend is inconsistent with expectations;

[0014] Generate learning intervention paths from student learning profiles, recommend corresponding learning resources to students based on learning resource push strategies, and collect learning status data to generate reminder values , in the reminder value Send out learning reminders to the outside world when it is lower than expected, and switch the reminder mode when it is not confirmed within the scheduled time;

[0015] Generate reminder density from the state data of receiving learning reminders , in the reminder density When the score is higher than expected, the performance feedback data will be analyzed and a learning feedback report will be generated;

[0016] Generate goal completion from learning feedback values ​​before and after intervention in students’ learning process If the goal completion The completion degree that does not exceed the preset completion threshold is obtained by principal component analysis. Optimize the learning resource push strategy based on the impact factor of the learning resource.

[0017] Guide students to create learning sharing content through new media channels and collect sharing feedback data, and generate sharing coefficients based on sharing feedback data , among which, the support rate , interaction rate and sharing rate Perform linear normalization as follows:

[0018] ;

[0019] Where: and is the weight coefficient, , ,and .

[0020] Based on sharing coefficient Execute immediate incentive feedback measures, or output optimization suggestions for intervention path optimization by learning the intervention path optimization knowledge graph.

[0021] Furthermore, student learning feedback data is collected, including study time distribution, test scores, homework scores and subject progress goal achievement data, and summarized to generate a learning feedback data set; using the feedback data as input, the trained feedback evaluation model is used to perform feedback evaluation to obtain student learning feedback values.

[0022] Furthermore, the learning feedback value is predicted and the corresponding predicted value is obtained, which is arranged along time and the corresponding trend graph is constructed; a similarity analysis is performed between the obtained trend graph and the expected trend graph. If the obtained trend similarity is lower than the similarity threshold, the students' learning process data is collected to guide students to set learning goals using the SMART principle.

[0023] Furthermore, the corresponding learning features are obtained after feature extraction of the student learning profile, and the pre-trained path planning algorithm generates the corresponding learning intervention path for the student based on the learning features;

[0024] After collecting student learning resources, they are summarized to generate a learning resource library; based on the correspondence between learning intervention paths and learning resources, a learning resource push strategy is generated by a content-based recommendation algorithm.

[0025] Furthermore, the learning status data such as the students’ learning progress and learning goal achievement rate are collected and recorded during the pre-set learning observation period, and the reminder value is generated from the learning status data. ;

[0026] If the reminder value When the learning threshold is lower than the preset reminder threshold, a learning reminder is sent to the outside; several reminder methods are collected and set in advance, and a reminder method library is generated; students' feedback data and preference data on learning reminders are collected, and the pre-trained recommendation algorithm selects a reminder method in the reminder method library.

[0027] Furthermore, the reminder value is generated by the learning status data The way is as follows:

[0028] ;

[0029] Where: and is the weight coefficient, , ,and ;

[0030] is the learning progress at time t, is the target progress value, is the learning performance at time t, For target scores; For the observation period Any time point on.

[0031] Furthermore, if the number of learning reminders issued in the current reminder cycle exceeds expectations, the ratio of the reminder value exceeding the reminder threshold is used as the reminder ratio. , the reminder density is generated by the time interval between learning reminders and the reminder ratio , if the reminder density If the density exceeds the preset threshold, an external performance analysis instruction is issued; among which, a reminder density is generated. The way is as follows:

[0032] ;

[0033] Where: For the i Second and j The reminder ratio difference between learning reminders, is the mean of the differences, is a very small number greater than 0, is the number threshold, is the number of reminders; is the length of the reminder cycle, is the time attenuation coefficient, is the interval of learning reminder at time t; weight coefficient, , , ,and .

[0034] Furthermore, student performance feedback data is collected, and after receiving the performance analysis instruction, the trained performance recognition model is used to analyze the student performance feedback data, obtain the student's current stage of learning weakness data, and generate a learning feedback report from the learning feedback data, learning status data and learning weakness data.

[0035] Furthermore, after aligning several feedback nodes before and after the intervention in the student learning process, the learning feedback value on the feedback node Generate target completion If the goal completion Does not exceed the preset completion threshold and issues optimization instructions to the outside world;

[0036] Furthermore, the learning feedback value on the feedback node Generate target completion The way is as follows:

[0037] ;

[0038] in, and Before and after intervention i The learning feedback value on the feedback node, and is the corresponding mean, i is the feedback node number, , n is the number of feedback nodes, For the i The median completion value of the feedback nodes, is the corresponding mean.

[0039] Furthermore, after receiving the optimization instruction, the students’ intervention feedback data are collected and the principal component analysis is performed on the intervention feedback data to obtain the influencing factors that affect the goal completion;

[0040] The pre-trained genetic algorithm is used to optimize the learning resource push strategy according to the influencing factors to obtain the optimized learning resource push strategy.

[0041] Furthermore, learning sharing content is created through new media channels and topic tags are added to collect sharing feedback data, including support rate, interaction rate and sharing rate, and then a sharing feedback data set is generated after aggregation; the sharing coefficient is generated from the sharing feedback data in the sharing feedback data set ;

[0042] If the sharing coefficient When the preset sharing threshold is exceeded, immediate incentive feedback measures will be implemented;

[0043] If the obtained sharing coefficient does not exceed the preset sharing threshold, the pre-trained LDA algorithm is used to learn the feedback report for topic extraction to obtain the corresponding feedback topic; learning intervention path optimization is used as the target word, and a learning intervention path optimization knowledge graph is pre-constructed.

[0044] Furthermore, the output interval of the optimization suggestion Constraints are imposed, and after the output interval that meets the constraints, the optimization suggestions are output. The constraints are as follows:

[0045] ;

[0046] Where: The number of times learning sharing content is created, For the i The sharing coefficient of times, is the corresponding mean, For time node and The difference in target completion between is the expected value of the difference in target completion, is the corresponding difference threshold; is an indicator function, which takes the value 1 when the condition is met, otherwise it takes the value 0. The weight coefficient is , ,and .

[0047] The digital media literacy intervention system for adolescents based on big data includes a feedback data analysis unit, which generates a learning feedback value for students after collecting their learning feedback data, and performs prediction and trend analysis on the learning feedback value. If the change trend is inconsistent with expectations, a student learning profile is generated based on the student's learning process data and learning goals;

[0048] The progress reminder unit generates learning intervention paths from student learning profiles, recommends corresponding learning resources to students based on learning resource push strategies, and collects learning status data to generate reminder values. , in the reminder value Send out learning reminders to the outside world when it is lower than expected, and switch the reminder mode when it is not confirmed within the scheduled time;

[0049] Feedback report generation unit, which generates reminder density from the status data of received learning reminders , in the reminder density When the score is higher than expected, the performance feedback data will be analyzed and a learning feedback report will be generated;

[0050] Strategy optimization unit, which generates target completion by learning feedback values ​​before and after intervention in the student learning process If the goal completion The completion degree that does not exceed the preset completion threshold is obtained by principal component analysis. Optimize the learning resource push strategy based on the impact factor of the learning resource.

[0051] Sharing feedback unit guides students to create learning sharing content through new media channels and collect sharing feedback data, and generate sharing coefficients based on sharing feedback data , according to the sharing coefficient Execute immediate incentive feedback measures, or output optimization suggestions for intervention path optimization by learning the intervention path optimization knowledge graph.

[0052] (III) Beneficial effects

[0053] The present invention provides a digital media literacy intervention system and method for teenagers based on big data, which has the following beneficial effects:

[0054] 1. By collecting learning feedback data to generate learning feedback values, students' current learning outcomes and learning status can be evaluated, and timely adjustments can be made when the learning effect fails to meet expectations; comprehensive stage evaluation of learning outcomes can achieve holistic evaluation and avoid the instability of single evaluation.

[0055] 2. Construct corresponding learning paths and learning resource push strategies for students based on their learning profiles, and provide customized interventions on students’ learning process and message goals; match and set intervention paths based on students’ learning profiles, and provide targeted and customized assistance to students’ learning literacy when they study on online learning platforms.

[0056] 3. Continuously collect students’ learning feedback and learning literacy and generate reminder values , evaluate the students’ current learning status and learning results, and remind students when the current learning results fail to meet expectations to prevent students from continuing to be in a low learning state.

[0057] 4. Guided by students’ reminder feedback data and preference data, students can be given the opportunity to choose reminder methods adaptively, which can improve push efficiency and increase students’ attention. When the number of learning reminders received is relatively high at the current stage, learning reminders can be used as a starting point to build reminder density. , to achieve a secondary evaluation of the current learning effect, so that students or intervention personnel can make further responses accordingly.

[0058] 5. Through data analysis and identification, students' learning weaknesses can be quickly determined and dealt with in a targeted manner. By regularly generating and outputting learning feedback reports and providing feedback on the current learning status and learning results, students' learning progress can be tracked.

[0059] 6. Based on the degree of goal completion It can form an effective evaluation of the current intervention effect, verify whether the current learning intervention has achieved the expected results, and re-customize learning resources for students so that students can obtain more matching learning resources on their current learning path.

[0060] 7. Generate sharing coefficients through collected sharing feedback data By introducing external evaluation and new media channels, a comprehensive evaluation of students' current learning outcomes can be conducted. The reliability of the evaluation is higher. By taking incentive feedback measures in this way, it can also play a positive role in promoting and motivating students and increase their enthusiasm for learning.

[0061] 8. After determining the current learning intervention path, optimize and adjust it. When the student’s current learning status cannot achieve the expected results, it can be adjusted in time. By constraining and adjusting the output interval of the learning intervention path, avoid the impact of high-frequency adjustments on student learning effects. As a display of learning content and learning effects, promote the growth of students’ new media literacy. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It is a flow chart of the digital media literacy intervention method for teenagers based on big data of the present invention;

[0063] Figure 2 It is a schematic diagram of the digital media literacy intervention method system for teenagers based on big data of the present invention. DETAILED DESCRIPTION

[0064] 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.

[0065] See also Figure 1 The present invention provides a method for intervening in digital media literacy of adolescents based on big data, including:

[0066] Step 1: After collecting students’ learning feedback data, generate students’ learning feedback values, and make predictions and trend analysis on the learning feedback values. If the changing trend is inconsistent with expectations, generate students’ learning profiles based on students’ learning process data and learning goals;

[0067] The step 1 includes the following contents:

[0068] Step 101: Collect student learning feedback data, including study time distribution, test scores, homework scores, and subject progress goal achievement, and summarize them to generate a learning feedback data set;

[0069] The machine learning algorithm is trained with the labeled sample data to generate a trained feedback evaluation model; the feedback data is used as input, and the trained feedback evaluation model is used to perform feedback evaluation to obtain the learning feedback value of the student at the current time node;

[0070] When in use, after a student has completed a period of learning, the learning feedback data is collected to generate a learning feedback value, and the student's current learning achievement and learning status are evaluated based on the learning feedback value. If the learning effect fails to meet expectations, it can be adjusted in time;

[0071] Step 102: Use time series analysis to predict the learning feedback value and obtain the corresponding prediction value. After summarizing the obtained prediction value with the historical learning feedback value, arrange them along time and construct a corresponding trend graph;

[0072] Perform similarity analysis between the obtained trend graph and the expected trend graph, and evaluate the student's current learning trend based on the trend similarity obtained by the analysis. If the obtained trend similarity is lower than the preset similarity threshold, it means that the student's current learning trend and learning status are in a poor state and need to be intervened in time. At this time, send an archive construction instruction to the outside;

[0073] When in use, as further feedback, the learning feedback values ​​at several consecutive time nodes are summarized and subjected to similarity analysis or distance calculation, and a comprehensive stage evaluation is conducted on the learning outcomes at the current stage, which can achieve a holistic evaluation and avoid the instability of a single evaluation.

[0074] Step 103: After receiving the file building instruction, collect the student's learning process data, including the student's learning activity participation and learning outcomes, as well as learning interests, learning grades, learning styles, and learning and performance goals;

[0075] And guide students to set learning goals using the SMART principle, generate student learning files based on students' learning process data and learning goals; add timestamps to student learning files and update them in real time;

[0076] The SMART principle is a method of goal setting and management that aims to ensure that the goals set are specific, measurable, achievable, relevant, and time-bound. This principle is widely used in personal development, project management, corporate strategic planning and other fields to help people set and achieve goals more effectively. The following is a detailed introduction to the meaning of each letter of the SMART principle:

[0077] Specific: Goals must be clear and specific and cannot be vague or too broad. Specificity requires a clear definition of what you want to achieve, including details such as the scope, quantity, and quality of the goal. For example, a specific goal might be "Increase sales by 15% in the next three months."

[0078] Measurable: Goals must be measurable in some way so that you can track progress and determine whether they have been successfully achieved. Measurability requires setting clear evaluation criteria or indicators, which can be quantity (such as sales, output), quality (such as customer satisfaction), time (such as the deadline for completing the task), etc.

[0079] Achievable: Goals must be realistic and achievable, not unrealistic fantasies. Achievability requires assessing resources and capabilities to ensure that goals are set based on actual conditions and can be achieved through hard work.

[0080] Relevant: Goals must be aligned with your overall goals, personal or organizational mission, vision, and strategic plan. Relevance requires making sure the goal is meaningful and moves you toward your long-term goals, rather than distracting from them.

[0081] Time-bound: The goal must have a clear deadline or time frame so that you can set a reasonable schedule and complete it on time. Time-bound requires setting a specific timeline for the goal, which can be days, weeks, months, or years, depending on the size and complexity of the goal.

[0082] By following the SMART principle, you can set clearer and more challenging goals and increase the likelihood of achieving them. At the same time, the SMART principle also helps to improve individual and team efficiency, responsibility and sense of achievement.

[0083] When using, combine the contents in steps 101 to 103:

[0084] As a further content, after collecting the learning process data and determining the corresponding learning goals, we can determine the corresponding learning profiles for the students, build the corresponding learning paths and learning resource push strategies for the students based on the learning profiles, and achieve customized intervention in the students' learning process and message goals.

[0085] In the existing intervention methods for digital media literacy of adolescents, big data and new media are usually introduced when students conduct independent learning through online learning platforms. New media and big data are used to recommend corresponding learning resources to students, and personalized customization of learning resources is achieved, so as to improve learning efficiency and improve learning feedback. However, when students are learning independently, in the existing intervention methods, the intervention path for students' learning is usually determined by the students' current learning feedback data. When there is a major change in the students' learning status, the pre-constructed intervention path and the change of intervention methods will have a large lag. When there is a deviation in the personalized customization strategy of learning resources and the pushed content, it cannot be adjusted quickly, which easily leads to the existing intervention methods failing to achieve the expected improvement effect.

[0086] Step 2: Generate learning intervention paths from student learning profiles, recommend corresponding learning resources to students based on learning resource push strategies, and collect learning status data to generate reminder values , in the reminder value Send out learning reminders to the outside world when it is lower than expected, and switch the reminder mode when it is not confirmed within the scheduled time;

[0087] The step 2 includes the following contents:

[0088] Step 201: extract features from the student's learning profile to obtain corresponding learning features, and use a pre-trained path planning algorithm to generate a corresponding learning intervention path for the student based on the learning features;

[0089] Collect student learning resources, such as reference books, study notes, multimedia learning resources, online courses, and online question banks, and generate a learning resource library after aggregation; generate a learning resource push strategy based on the content-based recommendation algorithm according to the correspondence between the learning intervention path and the learning resources, execute the learning resource push strategy, recommend corresponding learning resources to students based on their learning progress, and provide customized content recommendations to students;

[0090] When in use, the intervention path is matched and set based on the student's learning profile, and when the student is studying on the online learning platform, the student's learning literacy can be assisted in a targeted and customized manner.

[0091] Step 202: After the learning resource recommendation is completed, the learning status data such as the student's learning progress and learning goal achievement rate are collected and recorded within a preset learning observation period, and the learning status data set is generated after being aggregated;

[0092] A reminder value is generated from the learning status data to indicate whether a learning reminder is needed for the student. , as follows:

[0093] ;

[0094] Where: and is the weight coefficient, , ,and , the weight coefficient can be obtained by referring to the hierarchical analysis method; is the learning progress at time t, is the target progress value, is the learning performance at time t, For target scores; For the observation period Any time point on

[0095] The analytic hierarchy process (AHP) is a structured and systematic decision support technology. It decomposes complex multi-criteria decision problems into a series of interrelated and hierarchical sub-problems or factors, uses expert judgment to compare and quantitatively evaluate the importance of factors at each level, and finally obtains the relative advantages and disadvantages of each decision plan through synthetic calculation, providing decision makers with a clear and quantitative basis for decision making. This method combines qualitative and quantitative analysis, effectively improving the scientificity and accuracy of decision-making.

[0096] Pre-set reminder thresholds based on historical data and management expectations of students’ learning status;

[0097] If the reminder value obtained When it is lower than the preset reminder threshold, it means that the student’s current learning status or learning progress is lower than expected and needs to be adjusted and optimized. At this time, a learning reminder is issued to the outside world;

[0098] When used, as a further content, on the basis of completing the initial customized assistance for students, the learning feedback effect and learning literacy of students are continuously collected, and the reminder value is generated from the collected data , according to the reminder value The current learning status and learning results of students can be evaluated. When the current learning results fail to meet expectations, reminders can be sent to students to prevent them from continuing to be in a low learning state.

[0099] Step 203: pre-collect and set a number of reminder methods, such as instant messaging, email, and video conferencing, and generate a reminder method library;

[0100] After collecting students' feedback and preference data on learning reminders, when it is necessary to send learning reminders to students, the pre-trained recommendation algorithm selects a reminder method from the reminder method library; if the student fails to confirm the learning reminder within the scheduled time, the current reminder method is switched;

[0101] When using, combine the contents in steps 201 and 203:

[0102] As a further content, when it is necessary to remind students, using students' reminder feedback data and preference data as a guide to adaptively choose reminder methods for students can improve push efficiency. Of course, you can also choose the reminder method that students dislike the most to increase students' attention.

[0103] Step 3: Generate reminder density from the state data of receiving learning reminders , in the reminder density When the score is higher than expected, the performance feedback data will be analyzed and a learning feedback report will be generated;

[0104] The step three includes the following contents:

[0105] Step 301: If the number of learning reminders issued in the current reminder cycle exceeds the expected number, the ratio of the reminder value exceeding the reminder threshold is used as the reminder ratio. , the reminder density is generated by the time interval between learning reminders and the reminder ratio , as follows:

[0106] ;

[0107] Where: For the i Second and jThe reminder ratio difference between learning reminders, is the mean of the differences, is a very small number greater than 0, used to avoid the denominator being 0. is the number threshold, is the number of reminders; is the length of the reminder cycle, is the time attenuation coefficient, ranging from 0 to 1, and e is 2.718. is the interval of learning reminder at time t; weight coefficient, , , ,and ;

[0108] Pre-set the density threshold based on historical data and management expectations of students’ learning status and results;

[0109] If the reminder density If the density is higher than the preset threshold, it means that the student's current learning status is poor. This may be due to low interest or low matching degree in the current learning resources. The current learning resources can be enriched. At this time, a performance analysis instruction is issued to the outside.

[0110] When used, as a further content, if the number of learning reminders received in the current stage is large, it means that in the current stage, students have not achieved better results when studying. Use learning reminders as a starting point to build reminder density , realize the secondary evaluation of the current learning effect, so that students or intervention personnel can make further response processing accordingly;

[0111] Step 302: train the machine algorithm with the labeled sample data to obtain a trained performance recognition model; collect student performance feedback data, such as learning test scores and the degree of mastery of knowledge points; after receiving the performance analysis instruction, use the trained performance recognition model to analyze and identify the student's performance feedback data, obtain the student's current stage of learning weakness data, and generate a learning feedback report based on the learning feedback data, learning status data and learning weakness data;

[0112] When using, combine the contents in steps 301 and 302:

[0113] As a further content, after collecting several feedback data from students, through data analysis and identification, we can quickly determine the students' learning weaknesses and deficiencies, and deal with them in a targeted manner, for example, providing targeted learning guidance or learning resource support, etc.; by regularly generating and outputting learning feedback reports, we can provide feedback on the current learning status and learning results, and track the students' learning progress.

[0114] Step 4: Generate target completion from learning feedback values ​​before and after intervention in students’ learning process If the goal completion The completion degree that does not exceed the preset completion threshold is obtained by principal component analysis. Optimize the learning resource push strategy based on the impact factor of the learning resource.

[0115] The step 4 includes the following contents:

[0116] Step 401: After executing the feedback incentive for the students, collect the students' learning feedback data and re-acquire the learning feedback values ​​on several feedback nodes, align several feedback nodes before and after the intervention in the students' learning process, and use the learning feedback values ​​on the feedback nodes to calculate the learning feedback values. Generate target completion The way is as follows:

[0117] ;

[0118] in, and Before and after intervention i The learning feedback value on the feedback node, and is the corresponding mean, i is the feedback node number, , n is the number of feedback nodes, For the i The median completion value of the feedback nodes, is the corresponding mean;

[0119] Pre-set completion thresholds based on historical data and expectations of the effectiveness of student learning interventions;

[0120] If the goal is completed If the completion rate does not exceed the preset threshold, it means that the effect of the student learning intervention in the current stage has not met expectations, and the current intervention method may need to be optimized and adjusted. At this time, an optimization instruction is issued to the outside.

[0121] When in use, the goal completion degree is constructed by the learning feedback value collected in real time , based on the goal completion It can form an effective evaluation of the current intervention effect and verify whether the current learning intervention has achieved the expected effect. If the expected effect has been achieved, the current learning intervention method can be continued to improve students' learning literacy and new media literacy;

[0122] Step 402: after receiving the optimization instruction, collect the student's learning status and learning feedback report, learning resource recommendation data, new media content recommendation data, and learning intervention path data, summarize them as intervention feedback data, perform principal component analysis on the intervention feedback data, and obtain influencing factors that affect the goal completion;

[0123] The pre-trained genetic algorithm is used to optimize the learning resource push strategy according to the influencing factors, and the optimized learning resource push strategy is obtained. After the optimized learning resource push strategy is executed, the corresponding learning resources can be recommended according to the students' learning progress;

[0124] When using, combine the contents in steps 401 and 402:

[0125] On the basis of optimizing the learning intervention path, if the current intervention plan fails to achieve better improvement results, as a way of compensation, starting from the push of learning resources, we will re-customize learning resources for students so that students can obtain more matching learning resources on the current learning path.

[0126] Step 5: Guide students to create learning sharing content through new media channels and collect sharing feedback data, and generate sharing coefficients based on sharing feedback data , according to the sharing coefficient Execute immediate incentive feedback measures, or output optimization suggestions for intervention path optimization by learning intervention path optimization knowledge graph;

[0127] The step five includes the following contents:

[0128] Step 501: guide students to create learning sharing content and add topic tags through new media channels. After a preset sharing stage, collect sharing feedback data through new media channels, including support rate, interaction rate, and sharing rate, and generate a sharing feedback data set after aggregation;

[0129] Step 502: Generate a sharing coefficient from the sharing feedback data in the sharing feedback data set , among which, the support rate , interaction rate and sharing rate Perform linear normalization as follows:

[0130] ;

[0131] Where: and is the weight coefficient, , ,and , the value is consistent with the previous value;

[0132] Pre-set sharing thresholds based on historical data and management expectations of the effects of sharing learning content;

[0133] If the sharing coefficient is obtained If the preset sharing threshold is exceeded, it means that students have obtained good feedback after sharing their learning content and experience through new media. At this time, immediate incentive feedback measures can be implemented, such as recommending teachers' online lectures to students, extending break time, or pre-requested rewards.

[0134] When in use, the sharing coefficient is generated through the collected sharing feedback data By introducing external evaluation and new media channels, a comprehensive evaluation of students' current learning outcomes can be conducted, which will make the evaluation more reliable. At the same time, by taking incentive feedback measures, it can also play a positive role in promoting and motivating students and increase their enthusiasm for learning.

[0135] Step 503: If the obtained sharing coefficient does not exceed the preset sharing threshold, the pre-trained LDA algorithm is used to learn the feedback report to perform topic extraction and obtain the corresponding feedback topic;

[0136] Taking learning intervention path optimization as the target word, after deep retrieval and entity relationship building, a learning intervention path optimization knowledge graph is pre-built; based on the correspondence between the feedback topic and the learning intervention path optimization suggestions, the learning intervention path optimization knowledge graph outputs the corresponding optimization suggestions;

[0137] When used, as a way to extract learning feedback reports, after determining the current learning intervention path, it can be optimized and adjusted. New learning intervention methods can be determined, and when the student's current learning status cannot achieve the expected effect, it can be adjusted in time;

[0138] Step 504: Output interval of optimization suggestions Constraints are imposed, and after the output interval that meets the constraints, the optimization suggestions are output. The constraints are as follows:

[0139] ;

[0140] Where: The number of times learning sharing content is created, For the i The sharing coefficient of times, is the corresponding mean, For time node and The difference in target completion between is the expected value of the difference in target completion, is the corresponding difference threshold; is an indicator function, which takes the value 1 when the condition is met, otherwise it takes the value 0. The weight coefficient is , ,and , the value is consistent with the previous value;

[0141] When using, combine the contents in steps 501 to 504:

[0142] When using it, considering that the student's learning path should be in a relatively stable state, the learning intervention path cannot be adjusted frequently. At this time, by constraining and adjusting the output interval of the learning intervention path, the impact of high-frequency adjustment on the student's learning effect is avoided; and as a further content, as a display of learning content and learning effect, the growth of students' new media literacy is promoted;

[0143] Constructing a knowledge graph with "learning intervention path optimization" as the target word requires a series of detailed and systematic steps. The following is a specific construction process:

[0144] Preparation for knowledge graph construction: Clarify the goal: Determine the core theme of the knowledge graph as "learning intervention path optimization". Data source determination: Identify and collect relevant structured, semi-structured and unstructured data, which may come from academic papers, professional books, online courses, educational forums, etc. Tool and technology preparation: Prepare tools and technologies for knowledge extraction, knowledge fusion and knowledge computing, such as deep learning models, graph databases, etc.

[0145] Knowledge extraction: Entity extraction: Extract entities related to learning intervention path optimization from data sources, such as "learners", "learning resources", "intervention strategies", etc. Use named entity recognition (NER) technology, combined with deep learning models (such as LSTM, CRF, etc.) for automatic extraction. Relationship extraction: Identify and extract the relationship between entities, such as "learners-use-learning resources", "learning resources-influence-learning effects", etc. Use relationship extraction technology, combined with deep learning models (such as TransR, RESCAL, etc.) for automatic extraction. Attribute extraction: Extract attribute information of entities, such as learners' age, gender, learning style, etc.; type, difficulty, duration, etc. of learning resources. Use data mining technology to mine the relationship pattern between entity attributes and attribute values ​​from text.

[0146] Knowledge fusion: Entity alignment: Align entities in multiple data sources to ensure the consistency of the same entity in different data sources. Use entity linking technology to disambiguate entities and resolve coreferences. Relationship integration: Integrate relationships in different data sources to eliminate contradictions and ambiguities. Use ontology libraries and model layers to manage and integrate relationships. Attribute integration: Integrate the attributes of entities to ensure the accuracy and completeness of the attributes. Use data cleaning and normalization technology for processing.

[0147] Knowledge processing: Ontology construction: Build an ontology library based on the domain knowledge of learning intervention path optimization. Use IDEF5, Skeletal Methodology and other ontology construction methods for construction. Knowledge reasoning: Use logic-based reasoning, graph-based reasoning and deep learning-based reasoning technology to perform knowledge reasoning and relationship completion. Discover new learning intervention paths and optimization strategies through reasoning of knowledge graphs. Quality assessment: Perform quality assessment on the knowledge in the knowledge graph to ensure the accuracy and reliability of the knowledge. Use indicators such as confidence and accuracy for assessment.

[0148] Knowledge storage and query: Choose a storage method: According to the scale and application scenarios of the knowledge graph, choose a suitable storage method, such as a graph database (such as Neo4J, GraphSQL, etc.). Design a query language: Design a query language suitable for the knowledge graph, such as SPARQL, etc. Ensure that the query language can efficiently access and query the data in the knowledge graph.

[0149] Iteration and update of knowledge graph: Data update: Update the knowledge graph in a timely manner with the emergence of new learning intervention path optimization data and knowledge. Use the update mechanism of the data layer to add or update entities, relationships, and attributes. Model optimization: Optimize and adjust the deep learning model according to the actual application effect. Use new data and features to train the model to improve the accuracy and efficiency of the model. Feedback and iteration: Collect user feedback and actual application effects to iterate and optimize the knowledge graph. Continuously improve the construction process and technical methods of the knowledge graph.

[0150] See also Figure 2 The present invention provides a digital media literacy intervention system for adolescents based on big data, including:

[0151] The feedback data analysis unit collects students' learning feedback data to generate students' learning feedback values, and makes predictions and trend analyses on the learning feedback values. If the changing trend is inconsistent with expectations, the student learning profile is generated based on the students' learning process data and learning goals.

[0152] The progress reminder unit generates learning intervention paths from student learning profiles, recommends corresponding learning resources to students based on learning resource push strategies, and collects learning status data to generate reminder values. , in the reminder value Send out learning reminders to the outside world when it is lower than expected, and switch the reminder mode when it is not confirmed within the scheduled time;

[0153] Feedback report generation unit, which generates reminder density from the status data of received learning reminders , in the reminder density When the score is higher than expected, the performance feedback data will be analyzed and a learning feedback report will be generated;

[0154] Strategy optimization unit, which generates target completion by learning feedback values ​​before and after intervention in the student learning process If the goal completion The completion degree that does not exceed the preset completion threshold is obtained by principal component analysis. Optimize the learning resource push strategy based on the impact factor of the learning resource.

[0155] Sharing feedback unit guides students to create learning sharing content through new media channels and collect sharing feedback data, and generate sharing coefficients based on sharing feedback data , according to the sharing coefficient Execute immediate incentive feedback measures, or output optimization suggestions for intervention path optimization by learning the intervention path optimization knowledge graph.

[0156] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0157] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0158] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only some logical function divisions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0159] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0160] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A digital media literacy intervention method for adolescents based on big data, characterized by: include, After collecting students’ learning feedback data, generate students’ learning feedback values, and make predictions and trend analysis on the learning feedback values. If the changing trend is inconsistent with expectations, generate student learning files based on students’ learning process data and learning goals; The learning intervention path is generated from the student learning profile, and the corresponding learning resources are recommended to the students according to the learning resource push strategy. The learning status data is collected to generate the reminder value LPv(t). When the reminder value LPv(t) is lower than expected, a learning reminder is issued to the outside. When it is not confirmed within the predetermined time, the reminder mode is switched. Among them, the method of generating the reminder value LPv(t) from the learning status data is as follows: Where: α and β are weight coefficients, 0≤α≤1, 0≤β≤1, and α+β=1; P(t) is the learning progress at time t, T is the target progress value, G(t) is the learning performance at time t, and A is the target performance; τ is any time node in the observation period [0, t]; If the number of learning reminders issued in the current reminder cycle exceeds expectations, the ratio of the reminder value exceeding the reminder threshold is used as the reminder ratio Xp, and the reminder density Txj is generated by the time interval between the received learning reminders and the reminder ratio; If the reminder density Txj is higher than the preset density threshold, a performance analysis instruction is issued to the outside. The reminder density Txj is generated as follows: Where: Xp ij Xp is the difference between the reminder ratio of the i-th and j-th learning reminders, a is the mean of the difference, ξ is a very small number greater than 0, U is the number threshold, N is the number of reminders; T is the length of the reminder cycle, λ is the time decay coefficient, Gp(t) is the interval of learning reminders at time t; weight coefficient, 0≤ζ≤1, 0≤X≤1, 0≤μ≤1 and ζ+X+μ=1; Generate reminder density Txj from the state data of receiving learning reminders. When the reminder density Txj is higher than expected, perform data analysis on the performance feedback data and generate a learning feedback report. The target completion degree Yo is generated by the learning feedback values ​​before and after the intervention in the student learning process. If the target completion degree does not exceed the preset completion degree threshold of Yo, the influencing factors affecting the target completion degree Yo are obtained by principal component analysis to optimize the learning resource push strategy; Guide students to create learning sharing content through new media channels and collect sharing feedback data, and generate sharing coefficient Fxp from sharing feedback data, among which the support rate Cv, interaction rate Dp and sharing rate Fv are linearly normalized as follows: Where: α and β are weight coefficients, 0≤α≤1, 0≤β≤1, and α+β=1; Execute instant incentive feedback measures based on the sharing coefficient Fxp, or output optimization suggestions for intervention path optimization from the learning intervention path optimization knowledge graph; After aligning several feedback nodes before and after the intervention in the student learning process, the learning feedback value Ko on the feedback node generates the target completion Yo. If the target completion Yo does not exceed the preset completion threshold, an optimization instruction is issued to the outside. The method of generating the target completion Yo from the learning feedback value Ko on the feedback node is as follows: Among them, AKo i and BKo i are the learning feedback values ​​on the i-th feedback node before and after the intervention, AKo a With BKo a is the corresponding mean, i is the feedback node number, i = 1, 2, ..., n, n is the number of feedback nodes, Zo i is the intermediate value of the completion degree of the i-th feedback node, Zo a is the corresponding mean; After receiving the optimization instruction, collect students' intervention feedback data, perform principal component analysis on the intervention feedback data, and obtain the influencing factors that affect the goal completion; use the pre-trained genetic algorithm to optimize the learning resource push strategy based on the influencing factors to obtain the optimized learning resource push strategy; Create learning sharing content through new media channels and add topic tags, collect sharing feedback data, including support rate, interaction rate and sharing rate, and generate a sharing feedback data set after aggregation; generate a sharing coefficient Fxp from the sharing feedback data in the sharing feedback data set, and if the sharing coefficient Fxp exceeds the preset sharing threshold, execute immediate incentive feedback measures; If the obtained sharing coefficient Fxp does not exceed the preset sharing threshold, the pre-trained LDA algorithm is used to extract topics from the learning feedback report to obtain the corresponding feedback topics; learning intervention path optimization is used as the target word, and a learning intervention path optimization knowledge graph is pre-constructed; Constrain the output interval JoG of the optimization suggestion. After the output interval meets the constraint conditions, the optimization suggestion is output. The constraint method is as follows; Where: m is the number of times learning and sharing content is created, Fxp i is the i-th sharing coefficient, Fxp a is the corresponding mean, ΔYo(t i , t i+1 ) is the time node t i With t i+1 The difference in target completion between b is the expected value of the difference in target completion, Δ∈ is the corresponding difference threshold; ||() is the indicator function, which takes the value 1 when the condition is met, otherwise it takes the value 0, and the weight coefficient is 0≤β≤1, 0≤α≤1, and α+β=1.

2. The method for intervention in digital media literacy for adolescents based on big data as claimed in claim 1, characterized in that: Collect students' learning feedback data, including study time distribution, test scores, homework scores, and subject progress goal achievement data, and summarize them to generate a learning feedback data set; Taking the feedback data as input, the trained feedback evaluation model is used to perform feedback evaluation and obtain the student learning feedback value.

3. The method for intervention in digital media literacy for adolescents based on big data as claimed in claim 2, characterized in that: Predict the learning feedback value and obtain the corresponding predicted value, arrange it along time and build the corresponding trend chart; A similarity analysis is performed between the obtained trend graph and the expected trend graph. If the obtained trend similarity is lower than the similarity threshold, the students' learning process data is collected and the students are guided to set learning goals using the SMART principle.

4. The method for intervention in digital media literacy for adolescents based on big data as claimed in claim 3, characterized in that: After extracting features from the student's learning profile, the corresponding learning features are obtained, and the pre-trained path planning algorithm generates the corresponding learning intervention path for the student based on the learning features; After collecting student learning resources, they are summarized to generate a learning resource library; based on the correspondence between learning intervention paths and learning resources, a learning resource push strategy is generated by a content-based recommendation algorithm.

5. The method for intervention in digital media literacy for adolescents based on big data as claimed in claim 4, characterized in that: Collect and record students' learning status data such as learning progress and learning goal achievement rate during a pre-set learning observation period, and generate a reminder value LPv(t) from the learning status data; if the reminder value LPv(t) is lower than the preset reminder threshold, send a learning reminder to the outside; collect and set several reminder methods in advance, summarize and generate a reminder method library, collect students' feedback data and preference data on learning reminders, and select a reminder method from the reminder method library using a pre-trained recommendation algorithm.

6. The method for intervention in digital media literacy for adolescents based on big data as claimed in claim 5, characterized in that: Collect students' performance feedback data, and after receiving the performance analysis instruction, use the trained performance recognition model to analyze the students' performance feedback data to obtain the students' current learning weaknesses data; A learning feedback report is generated from the learning feedback data, learning status data and learning weakness data.

7. A digital media literacy intervention system for adolescents based on big data, applying the method according to any one of claims 1 to 6, characterized in that: include, The feedback data analysis unit collects students' learning feedback data to generate students' learning feedback values, and makes predictions and trend analyses on the learning feedback values. If the changing trend is inconsistent with expectations, the student learning profile is generated based on the students' learning process data and learning goals. The progress reminder unit generates a learning intervention path from the student's learning profile, recommends corresponding learning resources to students based on the learning resource push strategy, collects learning status data to generate a reminder value LPv(t), sends a learning reminder to the outside when the reminder value LPv(t) is lower than expected, and switches the reminder mode when it is not confirmed within the scheduled time; A feedback report generating unit generates a reminder density Txj from the state data of receiving learning reminders, and when the reminder density Txj is higher than expected, performs data analysis on the performance feedback data and generates a learning feedback report; The strategy optimization unit generates the target completion Yo from the learning feedback values ​​before and after the intervention in the student learning process. If the target completion Yo does not exceed the preset completion threshold, the principal component analysis is used to obtain the influencing factors affecting the target completion Yo, and the learning resource push strategy is optimized; The sharing feedback unit guides students to create learning sharing content through new media channels and collect sharing feedback data. The sharing coefficient Fxp is generated from the sharing feedback data. Immediate incentive feedback measures are implemented based on the sharing coefficient Fxp, or the learning intervention path optimization knowledge graph outputs optimization suggestions for intervention path optimization.

Images