Teaching process analysis method and system based on big data

Abstract

The invention relates to the field of teaching analysis, and discloses a teaching process analysis method and system based on big data, and the method comprises the steps: collecting multi-modal data through a sensor group disposed at a teaching site, and processing and extracting time sequence parameters directly related to a learning cognitive state: obtaining a heart rate variability sequence based on visual data; obtaining pen point pressure and writing speed based on the intelligent writing tool; performing data preprocessing, including timestamp alignment, denoising and standardization processing, on the physiological state parameters and the writing kinetic parameters to form a regular time sequence data set; inputting the aligned heart rate variability sequence and the writing kinetic parameters into a preset fusion model for calculation, and outputting a cognitive input index of the student; and carrying out correlation comparison on the time sequence of the cognitive input degree index and the timeline of the teaching behavior of the teacher, and analyzing the average cognitive input degree change of the student corresponding to different teaching contents and activities, so as to evaluate the teaching effect.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of teaching analysis, in particular to a teaching process analysis method and system based on big data. BACKGROUND

[0002] In the field of teaching process analysis, traditional evaluation methods mostly rely on teachers' subjective observation, students' self-report or simple classroom interaction records. These methods are often limited by strong subjectivity, incomplete data collection and poor timeliness, making it difficult to objectively and accurately reflect students' real cognitive state in the learning process. With the development of educational informatization, some technical means such as video monitoring and electronic whiteboard have been introduced into the teaching environment to record teaching behavior, but these methods still lack in-depth quantitative analysis of students' physiological state and cognitive process.

[0003] In recent years, big data technology has gradually increased its application in teaching analysis, such as learning analysis through students' learning behavior logs and online learning platform data. However, these methods mostly focus on behavior-level statistics and fail to effectively combine students' physiological indicators (such as heart rate variability) and writing dynamics parameters (such as pen pressure and writing speed) and other multi-modal data, making it impossible to fully represent students' cognitive load, thought fluency and other deep cognitive states. In addition, the existing technology lacks an effective fusion mechanism for multi-source heterogeneous data, making it difficult to achieve dynamic and continuous evaluation of students' cognitive input, and even more difficult to finely correlate cognitive input with specific teaching behaviors of teachers, resulting in limited accuracy and practicality of teaching effectiveness evaluation. SUMMARY

[0004] The purpose of the present application is to provide a teaching process analysis method and system based on big data, which solves the above technical problems.

[0005] The purpose of the present application can be achieved by the following technical solutions: A teaching process analysis method based on big data, comprising: S1, collecting multi-modal data through a sensor group deployed in the teaching site and processing and extracting time series parameters directly related to learning cognitive state: Based on visual data, heart rate variability sequence is calculated from student facial video stream using remote photoplethysmography technology as a physiological state parameter representing cognitive load; Based on intelligent writing tools, real-time collection and calculation of pen tip pressure and writing speed time series signals are performed as writing dynamics parameters representing thought fluency; S2, data preprocessing of the physiological state parameters and writing dynamics parameters, including timestamp alignment, denoising and standardization processing to form a regular time series data set; S3. Input the aligned heart rate variability sequence and the writing dynamics parameters into a preset fusion model for calculation, and output the student's cognitive engagement index. S4. Correlate and compare the time series of the cognitive engagement index with the timeline of teachers' teaching behaviors, analyze the changes in students' average cognitive engagement corresponding to different teaching content and activities, and thus evaluate the teaching effect.

[0006] As a further technical solution, the preset fusion model is implemented through the following calculation formula: ; in, This is an indicator of cognitive engagement. The writing engagement level is obtained by integrating the pen tip pressure signal and writing speed signal within a preset time window, and then normalizing the result to obtain the writing engagement level. This represents the average level of writing engagement, derived from historical data analysis. The cognitive focus level is obtained by calculating the high-frequency power of heart rate variability within a preset time window and then normalizing it. Based on historical data analysis Mean; , The weighting coefficients are determined based on historical data.

[0007] As a further technical solution, the specific calculation method for writing engagement KR is as follows: ; in, For the pen tip at all times Real-time pressure, ~ For time window, This represents the upper limit of the pressure applied to the pen tip. For the writing process at all times Real-time speed, This represents the upper limit of writing speed. , The weighting coefficients are determined based on historical data.

[0008] As a further technical solution The calculation method for GH is as follows: The time intervals between consecutive heartbeats are extracted from photoplethysmography (PPG) signals or electrocardiogram (ECG) signals to obtain a time-sequential RR interval sequence. , ,..., ;in The total number of heartbeat intervals contained in the RR interval sequence; Perform a fast Fourier transform on the RR interval sequence to obtain the power spectral density. ; Integrating the power spectral density in the standard high-frequency band, from 0.15 Hz to 0.4 Hz, the expression is: ; in, For the corresponding frequency The power below.

[0009] As a further technical solution, step S1 also includes: Based on the visual data, the eye-tracking algorithm is used to calculate the student's visual attention focus trajectory on the teaching display area, and the duration of gaze in specific knowledge point areas is statistically analyzed. In step S3, the gaze duration and cognitive engagement index are processed according to preset processing rules to obtain an updated cognitive engagement index.

[0010] As a further technical solution, the specific method for processing the gaze duration and cognitive engagement index according to preset rules is as follows: The cognitive attention threshold was determined based on historical teaching data and cognitive science experimental conclusions, and was used to determine whether students were in a state of cognitive overload. The cognitive attention level (GH) is compared with the cognitive attention threshold. If GH < the cognitive attention threshold, the student is considered to be in a state of cognitive overload. The updated cognitive engagement index... Calculate using the following formula: ; otherwise, ; in, This refers to the maximum possible duration of student gaze in the teaching display area within a preset time window, determined based on the duration of the teaching scenario and historical gaze data statistics. The duration of student attention in a specific knowledge area. , This is a reference coefficient.

[0011] As a further technical solution, the process of correlating and comparing the time series of cognitive engagement indicators with the timeline of teachers' teaching behaviors in step S4 includes: S41. Based on the audio and video analysis of the teacher's lecture recordings, or the pre-set lesson plan timeline, determine the start and end times of different teaching content and teaching activities to form a teaching behavior timeline. S42. Divide the teaching behavior timeline into several analysis periods with a preset time window as the interval; S43. Calculate the average cognitive engagement index of students in each class during each analysis period to obtain the cognitive engagement sequence corresponding to the teaching behavior sequence. S44. Calculate the Pearson correlation coefficient or dynamic time-normalized distance between the cognitive engagement sequence and the preset teaching effect reference sequence to quantify the overall correlation strength between teaching behavior and students' cognitive engagement.

[0012] As a further technical solution, the process of analyzing the changes in students' average cognitive engagement corresponding to different teaching content and activities in step S4, thereby evaluating the teaching effectiveness, includes: S45. Identify the peak and trough periods in the cognitive engagement sequence; S46. Mark the teaching behaviors corresponding to the peak periods as efficient teaching behaviors, and mark the teaching behaviors corresponding to the valley periods as teaching behaviors to be optimized. Specifically, step S45 includes the following steps: S451. Calculate the mean and standard deviation of the class's average cognitive engagement sequence throughout the entire teaching process; S452, Set dynamic threshold: Peak period threshold = mean + k1 * standard deviation; Valley threshold = Mean - k² * Standard deviation; Wherein, k1 and k2 are preset sensitivity coefficients, both ranging from 0.5 to 2.0; S453. Mark the period when the cognitive engagement index is consistently higher than the peak period threshold within a continuous time window as the peak period. S454. Mark the period when the cognitive engagement index is consistently below the threshold of the trough period within a continuous time window as the trough period.

[0013] A teaching process analysis system based on big data, the system being used to implement the aforementioned teaching process analysis method based on big data.

[0014] The beneficial effects of this invention are: (1) This invention breaks through the limitations of traditional teaching assessment that relies on subjective reports or single behavioral indicators. By simultaneously collecting and integrating students' physiological state parameters (such as heart rate variability) and writing dynamics parameters, a multi-dimensional cognitive engagement assessment system is constructed. Based on the cross-validation of physiological signals and behavioral actions, it can more realistically and accurately reflect students' internal cognitive load and thinking fluency, advancing the assessment from external behavioral observation to the analysis level of internal cognitive state, and providing unprecedented objective data support for understanding students' learning experience.

[0015] (2) This invention uses timestamp alignment technology to make a detailed correlation and comparison between the time series of students’ cognitive engagement indicators and the timeline of teachers’ teaching behavior. This dynamic correlation mechanism makes the analysis no longer limited to the overall evaluation of the classroom effect, but can locate specific teaching links and content. This can clearly reveal which specific teaching activities or knowledge points can effectively improve students’ average cognitive engagement, and which links may lead to students’ cognitive overload or distraction. This advances the evaluation of teaching effectiveness from whether it is effective to why it is effective and how to optimize it. It provides teachers with direct and quantitative basis for precise teaching reflection and strategy adjustment.

[0016] (3) Based on the efficient teaching behaviors and teaching behaviors to be optimized identified by the above correlation analysis, this invention lays a solid foundation for building a data-driven teaching optimization closed loop. It can automatically identify the peak and valley periods in the teaching process and correspond these key periods with specific teaching behaviors, so that teaching improvement is no longer based solely on experience, but on the cognitive feedback of the student group, that is, the accumulated teaching behavior and cognitive input response data, providing a data foundation for the subsequent improvement of efficient teaching models. Attached Figure Description

[0017] The invention will now be further described with reference to the accompanying drawings.

[0018] Figure 1 This is a diagram illustrating the method steps of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1 As shown, this invention is a teaching process analysis method based on big data, comprising: S1. Collect multimodal data through sensor arrays deployed at the teaching site, and process and extract time-series parameters directly related to the learning and cognitive state: Based on visual data, heart rate variability sequences were calculated from student facial video streams using remote photoplethysmography (TPM) technology, serving as physiological state parameters characterizing cognitive load. Based on intelligent writing tools, the temporal variation signals of pen tip pressure and writing speed are collected and calculated in real time, serving as writing dynamics parameters that characterize the fluency of thought. S2. Perform data preprocessing on the physiological state parameters and writing dynamics parameters, including timestamp alignment, denoising and standardization, to form a regular time-series dataset; S3. Input the aligned heart rate variability sequence and the writing dynamics parameters into a preset fusion model for calculation, and output the student's cognitive engagement index. S4. Correlate and compare the time series of the cognitive engagement index with the timeline of teachers' teaching behaviors, analyze the changes in students' average cognitive engagement corresponding to different teaching content and activities, and thus evaluate the teaching effect.

[0021] The preset fusion model is implemented through the following calculation formula: ; in, This is an indicator of cognitive engagement. The writing engagement level is obtained by integrating the pen tip pressure signal and writing speed signal within a preset time window, and then normalizing the result to obtain the writing engagement level. This represents the average level of writing engagement, derived from historical data analysis. The cognitive focus level is obtained by calculating the high-frequency power of heart rate variability within a preset time window and then normalizing it. Based on historical data analysis Mean; , The weighting coefficients are determined based on historical data. , The value range of is usually [0, 1], and satisfies + =1, to ensure Normalization. The initial value can be set to... =0.5, =0.5, then adjust based on historical data.

[0022] The specific calculation method for writing engagement KR is as follows: ; in, For the pen tip at all times Real-time pressure, ~ For time window, This represents the upper limit of the pressure applied to the pen tip. For the writing process at all times Real-time speed, This represents the upper limit of writing speed. , The weighting coefficients are determined based on historical data. Specifically, they can be based on historical writing data, and the relative importance of pressure and speed to writing engagement can be determined through principal component analysis (PCA) or factor analysis. The values ​​range from [0, 1]. + =1.

[0023] The calculation method for GH is as follows: The time intervals between consecutive heartbeats are extracted from photoplethysmography (PPG) signals or electrocardiogram (ECG) signals to obtain a time-sequential RR interval sequence. , ,..., ;in The total number of heartbeat intervals contained in the RR interval sequence; Perform a fast Fourier transform on the RR interval sequence to obtain the power spectral density. ; Integrating the power spectral density in the standard high-frequency band, from 0.15 Hz to 0.4 Hz, the expression is: ; in, For the corresponding frequency The power below.

[0024] In this embodiment, two different but intrinsically related data sources are collected: intrinsic physiological state parameters, namely heart rate variability sequences, and extrinsic behavioral dynamic parameters, namely pen tip pressure and writing speed. The intrinsic physiological state parameters are extracted non-contactly from facial videos using remote photoplethysmography (TP). Since these parameters are directly related to the human autonomic nervous system and reflect the cognitive load level induced by cognitive tasks such as thinking, memory, and comprehension, they are an introverted, non-actively controlled indicator. The extrinsic behavioral dynamic parameters are collected through intelligent writing tools. These parameters directly reflect the fluency of thought and the intensity of motor output when students perform core learning tasks, making them an extroverted, active indicator. Preprocessing these two types of data—timestamp alignment, noise reduction, and standardization—is a prerequisite for all subsequent analysis. This not only solves the data consistency problem but, more importantly, lays the foundation for establishing the dynamic correlation between the two types of parameters over time. Then, using a preset fusion model, the specific expression is: The cognitive engagement index was calculated. On the one hand, historical reference values, namely historical averages, are introduced. Instead of directly integrating the absolute values ​​of current writing engagement and cognitive focus, these values ​​are compared with the student's own historical average. , It is a current value / historical reference value, which realizes individualized normalization and effectively eliminates individual differences. A student with naturally strong writing force and a student with weak writing force, a student with a high resting heart rate and a student with a low resting heart rate are placed on their respective basis to measure their relative changes, making the assessment results more fair and accurate. Meanwhile, the (1 + ratio) structure in the formula ensures that the calculation result is always positive, and when the current performance is equal to the historical average, the value in parentheses is 2, forming a stable benchmark; when the performance is better than the average, the output value increases, which helps to capture positive changes. Finally, instead of a simple weighted average, the Euclidean norm form using the square root of the sum of squares is adopted. This nonlinear model has the characteristic of highlighting weaknesses and amplifying strengths; if students perform well in both dimensions, The value will increase significantly; if a certain dimension is severely low, for example, the heart rate shows high load but writing stops, it may indicate confusion, and it will effectively inhibit [the activity]. The increase in values ​​allows the fusion index to more sensitively and reasonably reflect complex cognitive states; weighting coefficients and This gives the model flexibility, allowing the contribution of physiological and behavioral data to be adjusted according to the characteristics of different educational stages and disciplines.

[0025] For the formula By using definite integrals to measure the time window ~ Within, the cumulative effect of writing behavior; because learning is a process, instantaneous pressure or speed peaks are not very meaningful, while continuous and stable output is the true manifestation of fluent thinking. A high score means that the student has maintained an active writing state throughout the entire period. and The original signal is normalized to the [0,1] interval, eliminating dimensions and weights. and The model can distinguish the different levels of importance of pen pressure and speed for "fluency of thought." For example, pen pressure may be more important in calligraphy practice, while speed may be more important in shorthand.

[0026] For the formula Based on the established scientific consensus of heart rate variability (GH) spectrum analysis, the high-frequency power of GH is related to respiratory sinus arrhythmia and is mainly regulated by the parasympathetic nervous system, namely the resting and digestive systems. A higher GH value indicates that the parasympathetic nervous system is active and the body is in a relaxed, low-stress state. This means that the individual has better cognitive resource reserves (such as working memory and attention), higher cognitive flexibility and resilience, and a stronger ability to cope with stress and recover quickly after cognitive tasks. Therefore, by calculating the power of the 0.15-0.4Hz band, i.e., the standard high-frequency band, and using it as a positive correlation indicator of cognitive focus, a higher GH value means higher cognitive focus.

[0027] The above technical solution addresses the problem that traditional teaching evaluations, when acquiring student behavioral parameters, often remain at a macro, superficial, and outcome-oriented level, such as attendance, head-raising rate, number of hand-raisings, and the correctness of classroom answers. These parameters have a weak correlation with actual cognitive processes and are easily influenced by subjective judgment and performative learning. Therefore, this technical solution abandons these superficial indicators and instead explores and defines a new category of behavioral parameters that is substantially related to teaching evaluation and can directly map the internal cognitive process—writing dynamics parameters. The pen tip pressure and writing speed selected in this solution are not arbitrary behavioral signals, but rather high-value parameters that have been carefully considered and are closely coupled with cognitive states. Pen pressure is not only a manifestation of physical force, but also an external expression of cognitive load, emotional state, and concentration. When students encounter difficult problems, engage in deep thinking, or are highly focused, their neuromuscular tension increases, often unconsciously leading to increased pen pressure. Conversely, when attention is scattered or thinking is stagnant, pen pressure may become lighter and unstable. This makes pen pressure a sensitive indicator for detecting the intensity of cognitive effort and the degree of mental block.

[0028] Writing speed directly reflects the fluency of thought; writing quickly on familiar content indicates smooth knowledge retrieval; while a sudden slowdown or pause in writing speed may mean that one is understanding new knowledge, organizing thoughts, or encountering obstacles; the temporal changes in writing speed are like a tachymeter for thought, intuitively demonstrating the smoothness of the cognitive process.

[0029] The focus of this scheme is not on detecting whether students are writing, but on precisely quantifying how they write. Therefore, it represents a paradigm shift from qualitative judgment to quantitative analysis. For example, in a traditional approach, teachers only observe student A writing with their head down, a binary yes / no state. This solution, however, focuses on capturing instances where, during the two minutes of explaining the core formula, student A's pen pressure consistently exceeds their average by 15%, while their writing speed remains stable, thus calculating their writing engagement. The value is high, thus it is a continuous, quantifiable, and individual-based deep behavioral description.

[0030] It is precisely because of the inherent connection between the aforementioned behavioral parameters and cognitive processes that they have a substantial relationship with teaching evaluation. On the one hand, they achieve uninterrupted and objective quantification of the thinking process. Since writing is the most core, natural, and frequent intellectual activity in classroom teaching, the intelligent pen tool transforms this inherent and invisible thinking activity into recordable and analyzable dynamic time-series data without interfering with the normal teaching process. This provides teaching evaluation with objective insights into students' learning process rather than just learning outcomes. On the other hand, it forms a behavioral-physiological dual-channel verification with physiological parameters, greatly improving the reliability of the assessment. This is because using pen pressure or speed alone may be ambiguous. For example, high pen pressure may be due to habitual force rather than focus. However, this solution integrates it with the intrinsic physiological indicator of heart rate variability.

[0031] For example: If increased pen pressure and smooth writing are detected in students (high) ), while heart rate variability high-frequency power decreases (high) This indicates high cognitive focus, which strongly suggests that the student is deeply engaged and learning efficiently. Conversely, if pen pressure is disordered and writing stagnates (low cognitive focus), it indicates a lack of focus. Meanwhile, if heart rate variability shows high cognitive load, it may indicate that the student is experiencing difficulty in understanding and is confused and struggling. The above cross-validation makes the judgment of cognitive engagement no longer a guess, but a reliable conclusion based on the fusion of multi-source data.

[0032] At the same time, it also provides precise positioning for teaching intervention; teachers not only know that the lesson was not effective, but can also pinpoint which specific knowledge point caused the students' general confusion by tracing back the time period when writing engagement and cognitive focus decreased simultaneously; the ability to accurately link specific teaching behaviors with students' deep cognitive responses is one of the core technical effects of this solution, making precise teaching and personalized intervention possible.

[0033] Step S1 also includes: Based on the visual data, the eye-tracking algorithm is used to calculate the student's visual attention focus trajectory on the teaching display area, and the duration of gaze in specific knowledge point areas is statistically analyzed. In step S3, the gaze duration and cognitive engagement index are processed according to preset processing rules to obtain an updated cognitive engagement index.

[0034] In this embodiment, the student's visual attention focus trajectory and the duration of gaze at specific knowledge point areas are incorporated into the analysis system, enabling the assessment of cognitive engagement to move from macroscopic physiological and behavioral representations to microscopic knowledge point attention levels. In actual teaching scenarios, students' physiological or writing parameters may show a superficial state of focus, but their visual attention may not be focused on the core teaching content. If cognitive engagement is calculated based solely on the original parameters, misjudgment of engagement is likely to occur. However, the gaze data obtained by this solution through eye tracking can directly reflect students' willingness to actively pay attention to different knowledge points. For example, prolonged focus on key concept areas often indicates that students are actively engaging in deep thinking; conversely, brief pauses or a loss of attention on key content may suggest comprehension difficulties or a lack of interest. Subsequent updates combining gaze duration with cognitive engagement indicators will further refine the results to better reflect students' actual learning and cognitive states, avoiding the limitations of a single data dimension.

[0035] The specific method for processing the gaze duration and cognitive engagement index according to preset rules is as follows: The cognitive attention threshold was determined based on historical teaching data and cognitive science experimental conclusions, and was used to determine whether students were in a state of cognitive overload. The cognitive attention level (GH) is compared with the cognitive attention threshold. If GH < the cognitive attention threshold, the student is considered to be in a state of cognitive overload. The updated cognitive engagement index... Calculate using the following formula: ; otherwise, ; in, This refers to the maximum possible duration of student gaze in the teaching display area within a preset time window, determined based on the duration of the teaching scenario and historical gaze data statistics. The duration of student attention in a specific knowledge area. , This is a reference coefficient. The attenuation coefficient, typically in the range of [0.1, 0.5], represents the attenuation factor during cognitive overload. The degree of reduction; based on historical data on cognitive overload. The correction ratio is determined; The enhancement factor, typically in the range [0, 1], is used to adjust the fixation duration under non-overload conditions. By analyzing historical data on gaze duration and The correlation analysis was used to determine the values. Specific values ​​can be set through linear regression or expert experience, for example... =0.3, =0.5.

[0036] Cognitive focus as referred to in this invention It is an indicator characterizing students' cognitive resource availability and neurophysiological flexibility. A higher baseline GH value indicates that students are in a low-stress state conducive to focused learning. This invention identifies anomalies indicating cognitive resource exhaustion caused by teaching tasks by setting dynamic thresholds based on individual historical data. Decrease, i.e., cognitive overload; cognitive attention threshold is based on the student's individual historical baseline. The set dynamic value; In this embodiment, the actual significance of gaze duration differs between different cognitive states of students, namely cognitive overload and normal cognition. When students are in a state of cognitive overload, even if they gaze at a specific knowledge point for a long time, it may be a passive gaze due to difficulty in understanding, rather than active engagement. If gaze duration is still positively correlated with engagement in the conventional way, the actual level of cognitive engagement will be seriously overestimated. Therefore, this rule first judges the cognitive state through a cognitive focus threshold, and then adjusts the indicator calculation method accordingly: when cognitive overload occurs, the coefficient correction avoids the indicator being artificially high, and truly reflects the cognitive difficulties of students in digesting content; under normal cognitive state, the positive impact of actively paying attention to knowledge points is amplified in an exponential form, reasonably reflecting the active thinking behavior behind gaze duration.

[0037] The process of correlating and comparing the time series of cognitive engagement indicators with the timeline of teachers' teaching behaviors in step S4 includes: S41. Based on the audio and video analysis of the teacher's lecture recordings, or the pre-set lesson plan timeline, determine the start and end times of different teaching content and teaching activities to form a teaching behavior timeline. S42. Divide the teaching behavior timeline into several analysis periods with a preset time window as the interval; S43. Calculate the average cognitive engagement index of students in each class during each analysis period to obtain the cognitive engagement sequence corresponding to the teaching behavior sequence. S44. Calculate the Pearson correlation coefficient or dynamic time-normalized distance between the cognitive engagement sequence and the preset teaching effect reference sequence to quantify the overall correlation strength between teaching behavior and students' cognitive engagement.

[0038] This embodiment defines in detail the specific process of comparing the time series of cognitive engagement with the timeline of teachers' teaching behavior, which solves the technical problem in traditional teaching analysis that it is difficult to accurately match teaching behavior with student feedback and that the correlation results lack quantitative basis. It promotes the upgrading of teaching effectiveness evaluation from subjective experience judgment to data-driven objective quantitative analysis.

[0039] The process of analyzing changes in students' average cognitive engagement corresponding to different teaching content and activities in step S4, thereby evaluating teaching effectiveness, includes: S45. Identify the peak and trough periods in the cognitive engagement sequence; S46. Mark the teaching behaviors corresponding to the peak periods as efficient teaching behaviors, and mark the teaching behaviors corresponding to the valley periods as teaching behaviors to be optimized. Specifically, step S45 includes the following steps: S451. Calculate the mean and standard deviation of the class's average cognitive engagement sequence throughout the entire teaching process; S452, Set dynamic threshold: Peak period threshold = mean + k1 * standard deviation; Valley threshold = Mean - k² * Standard deviation; Wherein, k1 and k2 are preset sensitivity coefficients, both ranging from 0.5 to 2.0; S453. Mark the period when the cognitive engagement index is consistently higher than the peak period threshold within a continuous time window as the peak period. S454. Mark the period when the cognitive engagement index is consistently below the threshold of the trough period within a continuous time window as the trough period.

[0040] In this embodiment, a complete evaluation logic is constructed around analyzing changes in cognitive engagement and identifying efficient and under-optimized teaching behaviors. This solves the technical pain points of traditional teaching evaluation, such as the difficulty in locating specific effective / ineffective teaching links and the lack of clear optimization directions. It extends the evaluation of teaching effectiveness from the overall result description to precise improvement guidance for specific links, forming a closed loop of evaluation and optimization.

[0041] A teaching process analysis system based on big data, the system being used to implement the aforementioned teaching process analysis method based on big data.

[0042] It should be noted that the calculation formulas and all parameters involved in the calculations in this invention have been dimensionless beforehand. The process of dimensionless processing is well known in the industry and will not be described here.

[0043] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A teaching process analysis method based on big data, characterized in that, include: S1. Collect multimodal data through sensor arrays deployed at the teaching site, and process and extract time-series parameters directly related to the learning and cognitive state: Based on visual data, heart rate variability sequences were calculated from student facial video streams using remote photoplethysmography (TPM) technology, serving as physiological state parameters characterizing cognitive load. Based on intelligent writing tools, the temporal variation signals of pen tip pressure and writing speed are collected and calculated in real time, serving as writing dynamics parameters that characterize the fluency of thought. S2. Perform data preprocessing on the physiological state parameters and writing dynamics parameters, including timestamp alignment, denoising and standardization, to form a regular time-series dataset; S3. Input the aligned heart rate variability sequence and the writing dynamics parameters into a preset fusion model for calculation, and output the student's cognitive engagement index. S4. Correlate and compare the time series of the cognitive engagement index with the timeline of teachers' teaching behaviors, analyze the changes in students' average cognitive engagement corresponding to different teaching content and activities, and thus evaluate the teaching effect.

2. The teaching process analysis method based on big data according to claim 1, characterized in that, The preset fusion model is implemented through the following calculation formula: ； in, As an indicator of cognitive engagement; The writing engagement level is obtained by integrating the pen tip pressure signal and writing speed signal within a preset time window, and then normalizing the result to obtain the writing engagement level. This represents the average level of writing engagement, derived from historical data analysis. The cognitive focus level is obtained by calculating the high-frequency power of heart rate variability within a preset time window and then normalizing it. Based on historical data analysis Mean; , The weighting coefficients are determined based on historical data.

3. The teaching process analysis method based on big data according to claim 2, characterized in that, The specific calculation method for writing engagement KR is as follows: ； in, For the pen tip at all times Real-time pressure, ~ For time window, This represents the upper limit of the pen tip pressure. For the writing process Real-time speed, This represents the upper limit of writing speed. , The weighting coefficients are determined based on historical data.

4. The teaching process analysis method based on big data according to claim 2, characterized in that, The calculation method for GH is as follows: The time intervals between consecutive heartbeats are extracted from photoplethysmography (PPG) signals or electrocardiogram (ECG) signals to obtain a time-sequential RR interval sequence. , ,..., ;in The total number of heartbeat intervals contained in the RR interval sequence; Perform a fast Fourier transform on the RR interval sequence to obtain the power spectral density. ; Integrating the power spectral density in the standard high-frequency band, from 0.15 Hz to 0.4 Hz, the expression is: ; in, For the corresponding frequency The power below.

5. The teaching process analysis method based on big data according to claim 1 or 2, characterized in that, Step S1 also includes: Based on the visual data, the visual attention focus trajectory of students on the teaching display area is calculated using an eye-tracking algorithm, and the duration of gaze dwell in specific knowledge point areas is statistically analyzed. In step S3, the gaze duration and cognitive engagement index are processed according to preset processing rules to obtain an updated cognitive engagement index.

6. The teaching process analysis method based on big data according to claim 1, characterized in that, The specific method for processing the gaze duration and cognitive engagement index according to preset rules is as follows: The cognitive attention threshold was determined based on historical teaching data and cognitive science experimental conclusions, and was used to determine whether students were in a state of cognitive overload. The cognitive attention level (GH) is compared with the cognitive attention threshold. If GH < the cognitive attention threshold, the student is considered to be in a state of cognitive overload. The updated cognitive engagement index... Calculate using the following formula: ; otherwise, ; in, This refers to the maximum possible duration of student gaze in the teaching display area within a preset time window, determined based on the duration of the teaching scenario and historical gaze data statistics. The duration of student attention in a specific knowledge area. , This is a reference coefficient.

7. The teaching process analysis method based on big data according to claim 1, characterized in that, The process of correlating and comparing the time series of cognitive engagement indicators with the timeline of teachers' teaching behaviors in step S4 includes: S41. Based on the audio and video analysis of the teacher's lecture recordings, or the pre-set lesson plan timeline, determine the start and end times of different teaching content and teaching activities to form a teaching behavior timeline. S42. Divide the teaching behavior timeline into several analysis periods with a preset time window as the interval; S43. Calculate the average cognitive engagement index of students in each class during each analysis period to obtain the cognitive engagement sequence corresponding to the teaching behavior sequence. S44. Calculate the Pearson correlation coefficient or dynamic time-normalized distance between the cognitive engagement sequence and the preset teaching effect reference sequence to quantify the overall correlation strength between teaching behavior and students' cognitive engagement.

8. The teaching process analysis method based on big data according to claim 7, characterized in that, The process of analyzing changes in students' average cognitive engagement corresponding to different teaching content and activities in step S4, thereby evaluating teaching effectiveness, includes: S45. Identify the peak and trough periods in the cognitive engagement sequence; S46. Mark the teaching behaviors corresponding to the peak periods as efficient teaching behaviors, and mark the teaching behaviors corresponding to the valley periods as teaching behaviors to be optimized. Specifically, step S45 includes the following steps: S451. Calculate the mean and standard deviation of the class's average cognitive engagement sequence throughout the entire teaching process; S452, Set dynamic threshold: Peak period threshold = mean + k1 * standard deviation; Valley threshold = Mean - k² * Standard deviation; Wherein, k1 and k2 are preset sensitivity coefficients, both ranging from 0.5 to 2.0; S453. Mark the period when the cognitive engagement index is consistently higher than the peak period threshold within a continuous time window as the peak period. S454. Mark the period when the cognitive engagement index is consistently below the threshold of the trough period within a continuous time window as the trough period.

9. A teaching process analysis system based on big data, characterized in that, The system is used to implement the big data-based teaching process analysis method according to any one of claims 1-8.

Images