Interactive teaching method and system based on teaching video

Abstract

The invention discloses an interactive teaching method and system based on a teaching video, and relates to the technical field of information, and the method comprises the following steps: 1, collecting the surface behavior data of a student watching the teaching video for at least three times recently, and the corresponding knowledge evaluation data; 2, analyzing the collected behavior data and knowledge evaluation data through a machine learning algorithm to obtain a concentration state distribution diagram of students in each section of video class, and judging the knowledge absorption ability of the students in combination with the corresponding knowledge evaluation data; step 3, performing targeted optimization processing on teaching video contents in the video library according to the knowledge absorption capability of the students, and inserting an interaction task set adaptive to the students into corresponding time nodes in the teaching videos; and step 4, pushing the optimized teaching video to a student terminal, receiving task completion data of the student in real time, and continuing to optimize the interaction task according to the task completion data to complete video teaching.

Description

Technical Field

[0001] This invention relates to the field of information technology, specifically to an interactive teaching method and system based on instructional videos. Background Technology

[0002] With the rapid development of information technology, instructional videos have become a core carrier of knowledge transmission, providing valuable learning resources for a large number of students. To facilitate students' acquisition of knowledge from these videos and improve their learning outcomes, providing an effective teaching method is essential. However, existing teaching methods suffer from simplistic and untargeted interactive formats, failing to adapt to the different knowledge bases and learning paces of users. Some students, with strong comprehension or prior knowledge, may show inattentiveness while watching instructional videos but still achieve high scores on in-class quizzes. Applying the same interactive mechanisms and tasks to all students of varying levels will lead to boredom among these students, negatively impacting teaching effectiveness and diminishing their learning experience. Summary of the Invention

[0003] This invention provides an interactive teaching method based on instructional videos, designed to address the needs of students who exhibit inattentiveness while watching instructional videos but still achieve good academic results.

[0004] This invention provides the following technical solution: an interactive teaching method based on instructional videos, comprising the following steps:

[0005] Step 1: Collect surface behavior data and corresponding knowledge assessment data from students during the most recent viewing of instructional videos;

[0006] Step 2: Analyze the collected behavioral data and knowledge assessment data using machine algorithms to obtain a distribution map of students' attention status in each video lesson. Combine this with the corresponding knowledge assessment data to determine students' knowledge absorption ability.

[0007] Step 3: Optimize the teaching video content in the video library according to the students' knowledge absorption ability, and insert interactive task sets adapted to the students into the corresponding time nodes in the teaching videos;

[0008] Step 4: Push the optimized teaching video to the student terminal, and teach the students in a dialogue or interactive manner. Based on the completion data of the interactive tasks, repeat steps 1-3 to continue to optimize the teaching video content.

[0009] As a preferred embodiment of the present invention, in step 1, the surface behavior data includes the percentage of fast-forwarding, the percentage of focused gaze, and the percentage of interactive participation; the knowledge assessment data includes the assessment accuracy rate.

[0010] As a preferred embodiment of the present invention, step 2 includes the following steps:

[0011] The collected surface behavior data is defined into computable quantitative indicators. The quantified quantitative indicators are integrated into a focus state feature vector, and the focus state feature vector is standardized using the Min-Max normalization algorithm.

[0012] First, samples with clearly defined high focus and clearly defined low focus are selected based on a preset combination of quantitative indicator thresholds. Then, the remaining samples are input into the random forest model. The random forest algorithm is used to obtain the probability of each sample being highly focused. If the output probability value is greater than the set threshold X, it is considered highly focused; otherwise, it is considered low focused.

[0013] The video segments identified as having low focus are correlated with the knowledge assessment data of the corresponding knowledge points, and the focus status assessment results are corrected based on the assessment results.

[0014] Based on the revised focus state assessment results, the system outputs the student's focus state at each time point in the teaching video, the corresponding knowledge point difficulty coefficient, and mastery level, thereby determining the strength of the student's knowledge absorption ability.

[0015] As a preferred embodiment of the present invention, associating video periods identified as low-focus with knowledge assessment data of corresponding knowledge points includes the following steps: locating video periods identified as low-focus in the preliminary focus state assessment results, extracting the teaching knowledge points corresponding to those periods, and querying the difficulty coefficient of the knowledge points and the students' accuracy rate in assessing those knowledge points.

[0016] As a preferred embodiment of the present invention, the step of correcting the focus state determination result based on the evaluation result includes: if the evaluation accuracy rate is ≥ threshold M, maintaining the low focus determination and marking it as effective low focus; if the evaluation accuracy rate is < threshold N, maintaining the low focus determination and marking it as ineffective low focus; if the evaluation accuracy rate is N~M, further determining whether the fast-forward percentage data only refers to fast-forwarding of basic knowledge, and if so, correcting the low focus to medium focus.

[0017] As a preferred embodiment of the present invention, the concentration state distribution is presented in the form of a time axis, including the concentration state of each time period, the corresponding knowledge point difficulty coefficient and mastery status. If the student shows a lack of focus in learning but the assessment result is that the student has mastered the knowledge, then the student is a student with strong knowledge absorption ability. If the student shows focus in learning but the assessment result is that the student has not mastered the knowledge, then the student is a student with weak knowledge absorption ability.

[0018] As a preferred embodiment of the present invention, step 3, the step of optimizing the teaching video content in the video library, is as follows:

[0019] Perform content analysis on the instructional videos, identify the difficulty level of all knowledge points in the videos, and insert interactive markers into the videos:

[0020] If the student has a strong ability to absorb knowledge, interactive tasks for knowledge points with high difficulty levels should be placed at the end of the knowledge point, while interactive tasks for knowledge points with low difficulty levels should be placed within the knowledge point. If the student has a weak ability to absorb knowledge, interactive tasks for knowledge points with high difficulty levels should be placed within the knowledge point, while interactive tasks for knowledge points with low difficulty levels should be placed at the end of the knowledge point.

[0021] As a preferred embodiment of the present invention, the interactive task set includes in-depth expansion interactive tasks for knowledge points corresponding to low attention periods, and precise reinforcement interactive tasks for knowledge points to be consolidated related to low attention periods; the in-depth expansion interactive tasks include knowledge point transfer and application questions and cross-chapter association analysis questions, and the precise reinforcement interactive tasks include questions for identifying easily confused points and practical questions for breaking down steps.

[0022] As a preferred embodiment of the present invention, during the teaching process, the system collects students' surface behavior data in real time. If students fast-forward through videos, look away from the screen, or skip interactive tasks, the system determines whether to push online interactive tasks based on the video content.

[0023] An interactive teaching system based on instructional videos, applicable to the aforementioned interactive teaching methods, comprising:

[0024] The data acquisition module is used to collect students' surface behavior data during the process of watching teaching videos, as well as corresponding knowledge assessment data; the data acquisition module connects to the video playback terminal to capture playback operation logs, links with the assessment system to synchronize answer results and incorrect question information, and supports multi-terminal data fusion;

[0025] The learning feature analysis module is used to analyze the collected behavioral data and knowledge assessment data through machine learning algorithms to obtain the distribution map of students' attention status in each video lesson. Combined with the corresponding knowledge assessment data, it judges the students' knowledge absorption ability. The learning feature analysis module has a built-in attention status recognition model and a knowledge point mastery matrix construction unit, and supports real-time data analysis and periodic updates.

[0026] The teaching video optimization module is used to optimize the teaching video content in the video library according to the student's knowledge absorption ability, and insert interactive task sets adapted to the student into the corresponding time nodes in the teaching video; the interactive task generation module includes a task template library, which includes in-depth expansion templates and precise reinforcement templates.

[0027] The teaching and feedback module is used to push optimized teaching videos to student terminals, conduct teaching for students in a dialogue or interactive manner, and further optimize the teaching video content based on the completion data of interactive tasks.

[0028] The present invention has the following beneficial effects:

[0029] 1. By analyzing the collected surface behavior data and knowledge assessment data through machine learning algorithms, we can obtain data on students' attention status distribution and knowledge mastery level, and display them in the form of a timeline. Whether it is a machine or a person, we can intuitively see the attention status at each time period in each video lesson, providing data support for the insertion of subsequent interactive tasks. By combining the difficulty coefficient of the knowledge points in each time period and the knowledge assessment data, we can identify the types of students with effective low attention and ineffective low attention, and determine the strength of the student's knowledge absorption ability. Based on the student's situation, we can develop appropriate teaching video content and interactive tasks for students with effective and ineffective low attention, so as to avoid students becoming bored and improve the teaching effect.

[0030] 2. In identifying students with low focus, this method accurately distinguishes between effective and ineffective low focus, avoiding misjudgment of students' learning status. Traditional focus assessment methods rely on single behavioral indicators, easily misinterpreting students' actions such as fast-forwarding through basic knowledge segments, simply not paying attention to the screen, or skipping interactive tasks involving already mastered basic knowledge as distraction. This method, through multi-dimensional behavioral quantification and correction based on knowledge assessment results, accurately identifies students with high knowledge mastery who exhibit low focus. This distinction avoids over-intervention for high-performing learners while ensuring targeted intervention for genuinely distracted students.

[0031] 3. By using three indicators—fast-forward rate, eye focus rate, and interactive participation rate—in subsequent data analysis, if students fast-forward through basic knowledge segments, fail to focus on the video content during the storytelling phase, or actively skip interactive tasks involving basic knowledge segments, these data will be actively removed. This significantly reduces the computational load on the model and allows students to maintain their original learning state without any proactive intervention in their learning, thus improving learning efficiency. Attached Figure Description

[0032] Figure 1 This is a block diagram of the overall logic of the teaching method in this invention;

[0033] Figure 2 This is a logic block diagram of step 2 in this invention;

[0034] Figure 3 A logic block diagram for adapting interactive tasks in this invention;

[0035] Figure 4 This is a structural block diagram of the teaching system in this invention. Detailed Implementation

[0036] The technical solutions of the embodiments of this specification will be explained and described below with reference to the accompanying drawings. However, the following embodiments are only preferred embodiments of this specification and not all of them. Other embodiments obtained by those skilled in the art based on the embodiments in the implementation methods without creative effort are all within the protection scope of this specification.

[0037] The terms "first," "second," "third," etc., in the description, claims, and accompanying drawings are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such processes, methods, products, or apparatus.

[0038] In the following description, terms such as “inner,” “outer,” “upper,” “lower,” “left,” and “right” are used only to facilitate the description of the embodiments and to simplify the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this specification.

[0039] All data involved in this application are information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0040] Before introducing the technical solutions described in this manual, a brief introduction to the application background of the technical solutions will be given first.

[0041] Online education breaks down geographical and time barriers, providing access to a wide range of high-quality course resources. For students in areas with relatively limited educational resources, it undoubtedly opens a door to a broader ocean of knowledge. The flexible scheduling of online learning allows students to choose their class times according to their own schedules, which is especially important for busy individuals. Compared to the high costs of traditional education, online courses are generally more affordable, enabling more people to access higher education or professional skills training.

[0042] Therefore, this manual provides an interactive teaching method based on instructional videos, such as... Figure 1-4As shown, machine learning algorithms are used to analyze collected surface behavior data and knowledge assessment data to obtain data on students' attention state distribution and knowledge mastery level, which are then displayed in a timeline format. This allows both machines and individuals to intuitively observe the attention state at different times within each video lesson, providing data support for the insertion of subsequent interactive tasks. By combining the difficulty coefficient of knowledge points at each time point with knowledge assessment data, the strength of each student's knowledge absorption ability can be determined. Based on the student's situation, appropriate teaching video content and interactive tasks are designed to prevent student boredom and improve teaching effectiveness. Specifically, the steps include:

[0043] An interactive teaching method based on instructional videos includes the following steps:

[0044] Step 1: Collect surface behavior data and corresponding knowledge assessment data from students during the most recent viewing of instructional videos;

[0045] The surface behavior data mentioned above includes the percentage of fast-forwarding, the percentage of focused gaze, and the percentage of interactive participation; the knowledge assessment data includes the assessment accuracy rate.

[0046] To facilitate the quantification of the above indicators, the fast-forward percentage is defined as: (total fast-forward time / total video clip time) × 100%, and the gaze focus percentage is defined as: (time the gaze lingers on the video content area / total viewing time) × 100%. The interactive participation percentage is defined as: (number of completed interactions / total number of push interactions) × 100%.

[0047] The fast-forward percentage has two dimensions of meaning. First, the smaller the fast-forward percentage, the more focused the students are. Second, the larger the fast-forward percentage, the less focused the students are. When students fast-forward through the segments explaining basic knowledge, it directly reflects their active selection of video content. This is the core operational indicator for distinguishing between effective and ineffective low focus.

[0048] The percentage of time students' eyes are focused can directly reflect the level of concentration from the perspective of where visual attention falls. It is a direct mapping of the physiological state of concentration. If the eyes frequently leave the video content area, it is likely that the student is in a state of distraction. Moreover, it can be collected in real time through terminal devices, so the data acquisition cost is low and the timeliness is high. However, it should be noted that some video content may not require students to keep their eyes on the computer screen, such as storytelling or background knowledge. This often happens in history teaching, so each video content needs to be analyzed specifically.

[0049] The percentage of interactive participation is a direct reflection of the level of engagement in learning tasks. It can verify the learning commitment behind the viewing behavior and reflect the students' attitudes toward the interactive task, such as liking or disliking it.

[0050] The following details the data collection steps for fast-forward percentage, eye focus percentage, and interactive participation percentage:

[0051] First, the data collection steps for fast-forward percentage.

[0052] Define the scope and period of data collection: The data collection targets are the students' "most recent at least 3" viewings of instructional videos, and each viewing must be associated with a complete video segment (including video ID, chapter knowledge point annotations, and total video duration) to ensure data continuity and relevance, meeting the needs of subsequent analysis of students' knowledge absorption capabilities.

[0053] Accessing video playback terminals to obtain operation logs: Through the data acquisition module of the teaching system, establish a connection with the video playback terminals used by students (such as computers, tablets, mobile phones, etc.) to capture and store playback operation logs in real time; the logs should include the start time and end time of video playback, the trigger time of each fast-forward operation, the video progress at the start of fast-forward, and the video progress at the end of fast-forward.

[0054] Extract core data related to fast forward: Filter out the records corresponding to "fast forward operation" from the playback operation log, calculate the duration of each fast forward operation, and sum them up to obtain the "total fast forward duration" during the video viewing process. Remove the duration of fast forward basic knowledge segments, and at the same time, extract the "total duration of the corresponding video segment" for that viewing.

[0055] Calculate the fast-forward percentage: Calculate the fast-forward percentage of each video segment during each viewing using the formula "Fast-forward percentage = (Total fast-forward duration / Total duration of corresponding video segment) × 100%".

[0056] The calculated fast-forward percentage data is associated with and labeled with the specific time points, knowledge points, and number of views of the corresponding videos, and stored in the system database. It also supports data fusion from multiple terminals to ensure data integrity.

[0057] Second, the data collection steps for the percentage of eye focus.

[0058] First, check whether the video playback terminal used by students has eye tracking function or supports screen area landing point recognition, and obtain authorization from students through the teaching system to allow the terminal to capture eye landing point data during video viewing.

[0059] Screen area division and real-time capture: In the video playback interface, the video content area (i.e., the core area where the teaching video is displayed) and the non-video content area (such as window borders, comment areas, device desktops, etc.) are pre-divided; throughout the student's video viewing process, the specific point where the student's gaze lands is captured in real time through the terminal camera or screen sensing technology, and the duration of the gaze staying in the video content area is recorded.

[0060] Calculate the total viewing time and effective viewing time: Extract the total viewing time of this viewing session from the video playback terminal logs; at the same time, summarize the "time spent watching the video content area" in the above steps, and remove the time spent when the viewer was not paying attention to the video content and was in the storytelling stage or reciting with their eyes closed.

[0061] Calculate the percentage of time your eyes are focused on the video: Use the formula "Percentage of time your eyes are focused on the video content area / Total viewing time" to calculate the percentage of time your eyes are focused on the video in each viewing session.

[0062] Data verification and correlation: Verify the captured gaze data and exclude abnormal data; store the final gaze focus percentage data in association with the corresponding video time point, knowledge point, and number of views, and synchronize it to the system database to support subsequent matching and analysis with knowledge assessment data.

[0063] Third, the steps for collecting data on the percentage of interactive participation.

[0064] Define the scope and statistical period of interactive tasks: Clearly define interactive tasks as adaptive tasks inserted into teaching videos. The collection period should be consistent with the proportion of fast-forwarding and the proportion of eye focus. The push time, video knowledge point, and task type of each interactive task should be marked.

[0065] Interact with the interactive task system to obtain push and participation logs: Through the data collection module of the teaching system, connect with the interactive task generation and push system to obtain two core logs: one is the interactive task push log, which records the total number of interactive tasks pushed in each viewing session, the unique ID of each task, and the push time; the other is the student interaction participation log, which records the student's response behavior to each pushed task, such as whether they clicked on the task or submitted the task answer.

[0066] Define "effective participation" and count the number of participants: Define the standard for "interactive participation" based on the log: students complete the operation of "clicking to open the task and submitting part or complete answers" for the interactive task pushed; based on this standard, count the number of interactive tasks actually participated in by students in this viewing session from the "student interactive participation log", and remove the number of interactive tasks that students actively skipped the basic knowledge segments.

[0067] Calculate the percentage of interactive participation: Calculate the percentage of interactive participation in each view using the formula "Interactive participation percentage = (Number of actual interactive tasks / Total number of pushed interactive tasks) × 100%".

[0068] Data association and supplementary annotation: The interactive participation rate data is associated and stored with the corresponding video time nodes, knowledge points, interactive task types, and number of views; at the same time, the reasons for not participating in the task are supplemented and annotated, providing a basis for optimizing the frequency of inserting interactive tasks in the future, and supporting data synchronization across multiple terminals.

[0069] Step 2: Analyze the collected behavioral data and knowledge assessment data using machine learning algorithms to obtain data on students' attention status distribution and knowledge mastery level.

[0070] Specifically, the following steps are included:

[0071] Step 2.1: Define calculable quantitative indicators for the collected surface behavior data, integrate the quantified quantitative indicators into a focus state feature vector, and use the Min-Max normalization algorithm to standardize the focus state feature vector.

[0072] The core formula of the Min-Max normalization algorithm is: (Mapped to [0,1])

[0073] x is the original data point, and min(X) and max(X) are the minimum and maximum values ​​of the dataset X, respectively; X norm It is the normalized value, and its range is strictly limited to [0, 1].

[0074] The structure of the focus state feature vector is [fast forward percentage, gaze focus percentage, interactive participation percentage]; the standardization process uses the Min-Max normalization algorithm to map the values ​​of each quantitative indicator to the 0-1 range, eliminating the dimensional differences between different indicators.

[0075] Step 2.2: First, based on the preset combination of quantitative indicator thresholds, select samples with clear high focus and clear low focus. Then, input the remaining samples into the random forest model and use the random forest algorithm to obtain the probability of each sample being highly focused. If the output probability value is greater than the set threshold X, it is considered highly focused; otherwise, it is considered low focused.

[0076] Specifically, the threshold combination for high focus is defined as fast-forwarding percentage ≤ 5%, visual focus percentage ≥ 90%, and interactive participation percentage ≥ 80%, while the threshold combination for low focus is defined as any one of the following: fast-forwarding percentage > 30%, visual focus percentage ≤ 10%, and interactive participation percentage < 20%.

[0077] The formula for the output of the random forest with a high focus probability is: ;

[0078] Where K is the number of decision trees, used to balance accuracy and efficiency; K can be 100, 200, ...

[0079] p k (1|x) is the high-focus prediction probability of the k-th tree for sample x, and p k = (sample x number of samples with the label "high focus" falling into the leaf node) / (total number of samples in that leaf node).

[0080] In the process of classifying the model optimization, a random forest model is adopted, using samples labeled with behavioral data and knowledge assessment results as training data. The unfiltered samples are classified and the probability of them belonging to a high-focus state is output. The threshold X is set to 60% to 65%, preferably 60%, to judge the high-focus state. When the high-focus probability P(1|x) output by the model is greater than or equal to the set threshold X, it is judged as high focus. When the high-focus probability output by the model is less than the set threshold X, it is judged as low focus.

[0081] The sample construction method for labeling behavioral data and knowledge assessment results is as follows: if a student's behavioral data shows low focus and the corresponding knowledge point assessment accuracy is ≥85%, it is labeled as low focus - mastered; if a student's behavioral data shows low focus and the corresponding knowledge point assessment accuracy is <60%, it is labeled as low focus - not mastered; if a student's behavioral data shows high focus and the corresponding knowledge point assessment accuracy is ≥85%, it is labeled as high focus - mastered; if a student's behavioral data shows high focus and the corresponding knowledge point assessment accuracy is <60%, it is labeled as high focus - not mastered.

[0082] Step 2.3: Associate the video segments identified as having low focus with the difficulty coefficients and knowledge assessment data of the corresponding knowledge points, and correct the focus state determination results based on the assessment results.

[0083] The process of linking video segments initially identified as having low focus with knowledge assessment data for the corresponding knowledge points involves the following steps: locating video segments identified as having low focus in the initial focus assessment results, extracting the corresponding teaching knowledge points for those segments, and querying the difficulty coefficient of the knowledge points and the students' accuracy rate in assessing those knowledge points.

[0084] The threshold M ranges from 85% to 90%, with 85% being preferred. The threshold N ranges from 60% to 65%, with 60% being preferred.

[0085] The steps for correcting the focus state judgment result based on the assessment results include: if the assessment accuracy rate is ≥85%, maintain the low focus judgment and mark it as effective low focus; if the assessment accuracy rate is <60%, maintain the low focus judgment and mark it as ineffective low focus; if the assessment accuracy rate is 60%-84%, further determine whether the fast-forward percentage data only refers to fast-forwarding of basic knowledge, and if so, correct the low focus to medium focus.

[0086] Step 2.4: Based on the corrected focus state judgment results, output the student's focus state, the corresponding knowledge point difficulty coefficient and mastery status at each time period of the teaching video, and determine the strength of the student's knowledge absorption ability.

[0087] The distribution of focus status is presented in the form of a timeline, including the focus status of each time period, the corresponding difficulty coefficient of the knowledge points, and the mastery status. If a student is not focused on learning basic knowledge segments but the assessment result is that they have mastered the knowledge, then the student is considered to have strong knowledge absorption ability. If a student is focused on learning basic knowledge segments but the assessment result is that they have not mastered the knowledge, then the student is considered to have weak knowledge absorption ability.

[0088] The data processing method involved in step 2 has the following effects:

[0089] First, in identifying students with low focus, this method accurately distinguishes between effective and ineffective low focus, avoiding misjudgment of students' learning status. Traditional focus assessment methods rely on single behavioral indicators, easily misinterpreting students' actions such as fast-forwarding through basic knowledge segments, simply looking away from the screen, or skipping interactive tasks involving already mastered basic knowledge as distraction. This method, through multi-dimensional behavioral quantification and correction based on knowledge assessment results, accurately identifies students with high knowledge mastery who exhibit low focus. This distinction avoids over-intervention for high-performing learners while ensuring targeted intervention for genuinely distracted students.

[0090] Secondly, by using three indicators—fast-forward rate, eye focus rate, and interactive participation rate—in subsequent data analysis, if students fast-forward through basic knowledge segments, fail to focus on the video content during the storytelling phase, or actively skip interactive tasks involving basic knowledge segments, these data will be actively removed. This significantly reduces the computational load on the model and allows students to maintain their original learning state without any proactive intervention in their learning, thus improving learning efficiency.

[0091] Third, it reduces the cost of subjective judgment for teachers and enables intelligent teaching management. In traditional teaching, teachers need to manually judge students' attention status through classroom observation and homework analysis, which is inefficient and highly subjective. This method automatically collects and quantifies multi-dimensional behavioral data, automatically classifies and corrects it using machine learning models, and outputs a structured distribution of attention status, directly providing data support for the design of teaching videos and significantly reducing teachers' manual costs.

[0092] Step 3: Based on the student's knowledge absorption ability, optimize the teaching video content in the video library accordingly, and insert interactive task sets adapted to the student into the corresponding time points in the teaching videos.

[0093] Furthermore, in step 3, the steps for optimizing the teaching video content in the video library are as follows:

[0094] Perform content analysis on the instructional videos, identify the difficulty level of all knowledge points in the videos, and insert interactive markers into the videos:

[0095] If the student has a strong ability to absorb knowledge, interactive tasks for knowledge points with high difficulty levels should be placed at the end of the knowledge point, while interactive tasks for knowledge points with low difficulty levels should be placed within the knowledge point. If the student has a weak ability to absorb knowledge, interactive tasks for knowledge points with high difficulty levels should be placed within the knowledge point, while interactive tasks for knowledge points with low difficulty levels should be placed at the end of the knowledge point.

[0096] For students with strong knowledge absorption abilities, they don't need excessive focus on basic knowledge segments. Adding interactive tasks within these segments won't negatively impact their learning. Furthermore, inserting in-depth extension tasks or targeted reinforcement tasks for related knowledge points within these segments prevents students from becoming bored and facilitates further understanding of the knowledge point, as well as reviewing and reinforcing related concepts. Conversely, placing more challenging interactive tasks at the end of video segments keeps students highly focused, preventing interruptions or frustration, and thus avoiding boredom.

[0097] For students with weak knowledge absorption abilities, they actually need more focus on basic knowledge segments. Embedding interactive tasks at the end of video segments can help them maintain a high level of concentration while learning that knowledge point, avoiding interruptions or frustration caused by the interactive tasks, thus preventing students from becoming bored with them. Adding interactive tasks with examples within segments of difficult knowledge provides students with some time to absorb the information, which can actually help them digest and understand that knowledge point.

[0098] In addition, the length of the video content can be optimized for these two types of students. For students with strong knowledge absorption ability, the content of the basic knowledge explanation can be shortened appropriately, while for students with weak knowledge absorption ability, the content of the difficult knowledge explanation can be extended appropriately, so that these two types of students can carry out targeted learning and reinforcement.

[0099] Furthermore, during the teaching process, the system collects students' surface behavior data in real time. If students fast-forward through videos, look away from the screen, or skip interactive tasks, the system will determine whether to push online interactive tasks based on the video content.

[0100] If the videos in that time slot contain storytelling or memorization exercises that pique curiosity, such as reciting with eyes closed, then there's no need to push online interactive tasks. If the videos in that time slot contain examples that require concentration, then the appropriate task needs to be determined based on the student's ability to absorb knowledge. If students actively fast-forward through or skip basic knowledge segments in interactive tasks, then there's no need to push online interactive tasks. This simplifies the model's computational load and allows for timely adjustments to learning strategies for students with and without effective low concentration.

[0101] Furthermore, the interactive task set includes in-depth expansion interactive tasks targeting knowledge points corresponding to low-focus periods, and precise reinforcement interactive tasks targeting knowledge points to be consolidated related to low-focus periods; the in-depth expansion interactive tasks include knowledge point transfer and application questions and cross-chapter association analysis questions, and the precise reinforcement interactive tasks include questions to distinguish easily confused points and step-by-step breakdown and practical questions.

[0102] For student A mentioned above, based on their attention span distribution during the "Solution of First-Order Linear Differential Equations" instructional video, the content of the instructional videos in the video library was adjusted. For example, for the instructional videos on "Triple Integrals," including Knowledge Point 1: Definition and Geometric Meaning of Triple Integrals (0-4 minutes, Difficulty Coefficient 0.8), Knowledge Point 2: Calculation Method of Triple Integrals in Cartesian Coordinate System (4-10 minutes, Difficulty Coefficient 1), and Knowledge Point 3: Application of Triple Integrals in Cylindrical Coordinate System (10-15 minutes, Difficulty Coefficient 0.6), the explanation time for Knowledge Points 1 and 2 can be compressed. At the same time, the basic probability question-and-answer interactive tasks originally embedded in Knowledge Points 1 and 2 can be removed. Then, in-depth expansion interactive tasks or precise reinforcement interactive tasks corresponding to the knowledge points to be consolidated can be inserted into Knowledge Points 1 and 2. The interactive tasks originally embedded in the middle of the video segments in Knowledge Point 3 can be removed, and the interactive tasks can be inserted at the end of Knowledge Point 3. This allows students to maintain a high level of focus when learning difficult knowledge points and improves learning effectiveness.

[0103] Step 4: Push the optimized teaching video to the student terminal, and teach the students in a dialogue or interactive manner. Based on the completion data of the interactive tasks, repeat steps 1-3 to continue to optimize the teaching video content.

[0104] The above teaching method is suitable for students with strong knowledge absorption abilities. By compressing the explanation time of basic knowledge videos and inserting in-depth expansion and precise reinforcement of solidified knowledge while watching basic knowledge videos, it not only increases students' learning interest and improves learning efficiency, but also avoids wasting time due to uniform interactive tasks and fills some hidden gaps in knowledge during learning.

[0105] An interactive teaching system based on instructional videos, applicable to the aforementioned interactive teaching methods, comprising:

[0106] The data acquisition module is used to collect students' surface behavior data during the process of watching teaching videos, as well as corresponding knowledge assessment data; the data acquisition module connects to the video playback terminal to capture playback operation logs, links with the assessment system to synchronize answer results and incorrect question information, and supports multi-terminal data fusion;

[0107] The learning feature analysis module is used to analyze the collected behavioral data and knowledge assessment data through machine learning algorithms to obtain the distribution map of students' attention status in each video lesson. Combined with the corresponding knowledge assessment data, it judges the students' knowledge absorption ability. The learning feature analysis module has a built-in attention status recognition model and a knowledge point mastery matrix construction unit, and supports real-time data analysis and periodic updates.

[0108] The teaching video optimization module is used to optimize the teaching video content in the video library according to the student's knowledge absorption ability, and insert interactive task sets adapted to the student into the corresponding time nodes in the teaching video; the interactive task generation module includes a task template library, which includes in-depth expansion templates and precise reinforcement templates.

[0109] The teaching and feedback module is used to push optimized teaching videos to student terminals, conduct teaching for students in a dialogue or interactive manner, and further optimize the teaching video content based on the completion data of interactive tasks.

[0110] The embodiments described above are merely preferred embodiments of this specification and are not intended to limit the scope of this specification. Any modifications and improvements made by those skilled in the art to the technical solutions of this specification without departing from the spirit of this specification should fall within the protection scope defined by the claims of this specification.

Claims

1. An interactive teaching method based on instructional videos, characterized in that, Includes the following steps: Step 1: Collect surface behavior data and corresponding knowledge assessment data from students during the most recent viewing of instructional videos; Step 2: Analyze the collected behavioral data and knowledge assessment data using machine algorithms to obtain a distribution map of students' attention status in each video lesson. Combine this with the corresponding knowledge assessment data to determine students' knowledge absorption ability. Step 3: Optimize the teaching video content in the video library according to the students' knowledge absorption ability, and insert interactive task sets adapted to the students into the corresponding time nodes in the teaching videos; Step 4: Push the optimized teaching video to the student terminal, and teach the students in a dialogue or interactive manner. Based on the completion data of the interactive tasks, repeat steps 1-3 to continue to optimize the teaching video content.

2. The interactive teaching method based on instructional videos according to claim 1, characterized in that, In step 1, the surface behavior data includes the percentage of fast-forwarding, the percentage of focused gaze, and the percentage of interactive participation; the knowledge assessment data includes the assessment accuracy rate.

3. The interactive teaching method based on instructional videos according to claim 2, characterized in that, Step 2 includes the following steps: The collected surface behavior data is defined into computable quantitative indicators. The quantified quantitative indicators are integrated into a focus state feature vector, and the focus state feature vector is standardized using the Min-Max normalization algorithm. First, samples with clearly defined high focus and clearly defined low focus are selected based on a preset combination of quantitative indicator thresholds. Then, the remaining samples are input into the random forest model. The random forest algorithm is used to obtain the probability of each sample being highly focused. If the output probability value is greater than the set threshold X, it is considered highly focused; otherwise, it is considered low focused. The video segments identified as having low focus are correlated with the knowledge assessment data of the corresponding knowledge points, and the focus status assessment results are corrected based on the assessment results. Based on the revised focus state assessment results, the system outputs the student's focus state at each time point in the teaching video, the corresponding knowledge point difficulty coefficient, and mastery level, thereby determining the strength of the student's knowledge absorption ability.

4. The interactive teaching method based on instructional videos according to claim 3, characterized in that, The process of associating video segments identified as having low focus with knowledge assessment data for corresponding knowledge points includes the following steps: locating video segments identified as having low focus in the initial focus status assessment results, extracting the teaching knowledge points corresponding to those segments, and querying the difficulty coefficient of the knowledge points and the students' accuracy rate in assessing those knowledge points.

5. The interactive teaching method based on instructional videos according to claim 3, characterized in that, The steps for correcting the focus state judgment result based on the evaluation results include: if the evaluation accuracy rate is ≥ threshold M, maintain the low focus judgment and mark it as effective low focus; if the evaluation accuracy rate is < threshold N, maintain the low focus judgment and mark it as ineffective low focus; if the evaluation accuracy rate is N~M, further determine whether the fast-forward percentage data only refers to fast-forwarding of basic knowledge, and if so, correct the low focus to medium focus.

6. The interactive teaching method based on instructional videos according to claim 3, characterized in that, The distribution of focus status is presented in the form of a time axis, including the focus status of each time period, the corresponding difficulty coefficient of the knowledge points, and the mastery status. If a student shows a lack of focus in learning but the assessment result is that the student has mastered the knowledge, then the student is a student with strong knowledge absorption ability. If a student shows focus in learning but the assessment result is that the student has not mastered the knowledge, then the student is a student with weak knowledge absorption ability.

7. The interactive teaching method based on instructional videos according to claim 1, characterized in that, Step 3 involves optimizing the teaching video content in the video library as follows: Perform content analysis on the instructional videos, identify the difficulty level of all knowledge points in the videos, and insert interactive markers into the videos: If the student has a strong ability to absorb knowledge, interactive tasks for knowledge points with high difficulty levels should be placed at the end of the knowledge point, while interactive tasks for knowledge points with low difficulty levels should be placed within the knowledge point. If the student has a weak ability to absorb knowledge, interactive tasks for knowledge points with a high difficulty level should be set within the knowledge point's range, while interactive tasks for knowledge points with a low difficulty level should be set at the end of the knowledge point.

8. The interactive teaching method based on instructional videos according to claim 1, characterized in that, The interactive task set includes in-depth expansion interactive tasks targeting knowledge points corresponding to low focus periods, and precise reinforcement interactive tasks targeting knowledge points related to low focus periods that need to be consolidated. The in-depth expansion interactive tasks include knowledge point transfer and application questions and cross-chapter association analysis questions. The precise reinforcement interactive tasks include questions on easily confused points and step-by-step breakdown and practical questions.

9. The interactive teaching method based on instructional videos according to claim 1, characterized in that, During the teaching process, the system collects students' surface behavior data in real time. If students fast-forward through videos, look away from the screen, or skip interactive tasks, the system will determine whether to push online interactive tasks based on the video content.

10. An interactive teaching system based on instructional videos, characterized in that, include: The data acquisition module is used to collect surface behavioral data of students while watching instructional videos, as well as corresponding knowledge assessment data; The data acquisition module connects to the video playback terminal to capture playback operation logs, synchronizes answer results and incorrect question information with the assessment system, and supports multi-terminal data fusion; The learning feature analysis module is used to analyze the collected behavioral data and knowledge assessment data through machine learning algorithms to obtain the distribution map of students' attention status in each video lesson. Combined with the corresponding knowledge assessment data, it judges the students' knowledge absorption ability. The learning feature analysis module has a built-in attention status recognition model and a knowledge point mastery matrix construction unit, and supports real-time data analysis and periodic updates. The teaching video optimization module is used to optimize the teaching video content in the video library according to the student's knowledge absorption ability, and insert interactive task sets adapted to the student into the corresponding time nodes in the teaching video; the interactive task generation module includes a task template library, which includes in-depth expansion templates and precise reinforcement templates. The teaching and feedback module is used to push optimized teaching videos to student terminals, conduct teaching for students in a dialogue or interactive manner, and further optimize the teaching video content based on the completion data of interactive tasks.

Images