Learning path planning method based on habit analysis

Abstract

The invention provides a learning path planning method based on habit analysis, and belongs to the field of learning path planning. The problem of complex learning path planning is solved; the method specifically comprises the following steps: S1, obtaining personalized information of a learner; obtaining a learning target of a learner and a learning preset time length to form planning data; s2, judging the continuous learning duration of the learner according to the personalized information of the learner; performing preliminary planning on a learning path of the learner in combination with a learning target of the learner to obtain a first planned route; s3, analyzing the learning state of the learner according to the first planned route, arranging a review plan for the learner, and optimizing the first planned route according to the review plan; s4, acquiring a preset learning time length of the learner, and adjusting the optimized route to complete route planning; according to the method, the learning path of the learner is planned, data reference is provided for the learner, and the learning efficiency of the learner is improved.

Description

Technical Field

[0001] The present invention discloses a learning path planning method based on habit analysis, and relates to the field of learning path planning. Background Art

[0002] The existing learning path planning methods have the following shortcomings: Lack of personalization: Existing methods often plan learning paths based on fixed rules (e.g., "learn basic courses first, then advanced courses") or preset templates. These methods fail to fully consider individual learners' differences (e.g., knowledge base, learning style, cognitive ability, etc.), resulting in low learning efficiency and a poor learning experience. Lack of explainability and transparency: Existing learning path planning is primarily generated through black-box models such as deep learning. This makes it difficult to explain the recommendation logic, and learners cannot understand why specific courses or sequences are recommended. During the learning process, learners struggle to intuitively perceive their learning progress and goal achievement. Insufficient dynamic adaptability: Path planning lacks a dynamic adjustment mechanism and cannot flexibly adjust the content difficulty or progress based on learners' real-time performance (such as the accuracy of answering questions and the degree of distraction). When learners' abilities do not match the difficulty, it is easy to waste time or cause frustration. Summary of the Invention

[0003] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide a learning path planning method based on habit analysis, aiming to solve the complex problems of learning path planning.

[0004] To achieve the above objectives, the present invention is implemented through the following technical solutions: a learning path planning method based on habit analysis, the planning method includes: Step S1: Acquire learner's personalized information and planning data; Step S2: Determine the learner's continuous learning duration based on the learner's personalized information; perform preliminary planning of the learner's learning path based on the continuous learning duration and combined with the planning data to obtain a first planned route; Step S3: Based on the first planned route, the learner's learning completion is tested. Based on the learner's learning completion and combined with the learner's personalized information, the learner's learning status is analyzed, a review plan is arranged for the learner, and the first planned route is optimized based on the review plan; Step S4: Dynamically adjust the optimized route according to the planning data to complete the route planning.

[0005] Furthermore, the specific steps of step S1 are as follows: Step S11: acquiring real-time learning video data of the learner through a video capture device, and preprocessing the video data to generate high-quality video data; Step S12: Inputting the high-quality video data into a multimodal large model, the multimodal large model performs end-to-end deep visual understanding of the video frames to generate visual features containing information such as the learner's expression, attention, and posture; the multimodal large model further integrates the visual features with other modal data such as voice collected simultaneously to generate a comprehensive feature representation; Step S13: Based on the comprehensive feature representation, the video sequence is analyzed using a self-attention mechanism to generate quantified student state parameters, which constitute personalized information of the learner.

[0006] Furthermore, the specific steps of step S13 are as follows: Get the association set tzj(i), extract the number of association values ​​t in the association set; record the association values ​​in the association set as tzz(1), tzz(2), ..., tzz(t); Step S1312: Obtain correlation values ​​tzz(1), tzz(2), ..., tzz(t); construct a correlation function hs based on the correlation values. The correlation function is as follows: ; Obtain the correlation value of the learner's facial image to obtain xtz(1) to xtz(t); calculate it in combination with the correlation function hs to obtain the learner's correlation function xhs; ; Calculate the difference between the learner's correlation function xhs and the correlation function hs to obtain the difference in the direction from x1 to xt; ; The absolute values ​​of the differences in the direction from x1 to xt are accumulated and calculated to obtain the detection value; the expression type corresponding to the minimum detection value is obtained, and the learner's emotional intensity is obtained according to the self-attention mechanism. The expression category and emotional intensity are combined to obtain the learner's expression information.

[0007] Furthermore, the specific steps of step S13 also include: Obtain the focused attention area, connect the learner's line of sight starting point with the center of the focused attention area as the standard line; record the length of the standard line as bzc; perform edge detection on the focused attention area; test the angle between the line of sight and the standard line when the line of sight is at the edge of the focused attention area, and obtain the edge angle by; obtain the attention angle interval [0, by] from the edge angle; Obtain the learner's gaze point, connect it with the center of the area where attention is focused, and record the distance between the gaze point and the center of the area where attention is focused as lyz; combine this with the length of the standard line bzc to calculate the learner's gaze angle α; When the learner’s sight angle α∈[0, by], the learning sight is judged to be focused; When the learner's gaze is focused, the learner's blinking frequency is detected to obtain the learner's blinking frequencies zpl(1) to zpl(z) in the daily state; the average of the blinking frequencies zpl(1) to zpl(z) is calculated to obtain the average frequency jpl; the learner's real-time blinking frequency spl is obtained; if spl>jpl, it is judged that the learner is not paying attention.

[0008] Furthermore, the multimodal large model adopts a Transformer-based architecture, which divides video frames into image blocks and directly inputs them into the encoder. It learns the intrinsic correlation between image blocks through the self-attention mechanism to understand the image content as a whole, eliminating the need to pre-extract facial key points or design manual features.

[0009] Furthermore, the student status parameters include but are not limited to: the student's expression type; concentration level; emotion intensity and learning behavior.

[0010] Furthermore, the specific steps of step S2 are as follows: Step S21: Obtaining personalized information about the learner during learning, and monitoring the learner's facial expression type, emotional intensity, and concentration level in real time; and determining the learner's continuous learning time based on the learner's facial expression type, emotional intensity, and concentration level during learning; Step S22: Obtain the learner's learning goal, and decompose the learner's learning goal based on the learner's continuous learning time to obtain learning units; calculate the correlation between learning units to obtain correlation values, and use the correlation values ​​to preliminarily plan the learner's learning path.

[0011] Furthermore, the specific steps of step S22 are as follows: Step S221: Obtain the learner's learning goal and the learner's continuous learning time; obtain the estimated learning time for the learning goal; and divide the estimated learning time into learning units according to the learner's continuous learning time. Step S222: Obtain the number of learning units xgs; obtain the time required for the learner to learn each learning unit dsj(1), dsj(2), ..., dsj(xgs); combine the learning units in pairs to obtain combined units; obtain the time zh2 required for the learner to learn the combined units; The correlation value glz2 between learning units is calculated based on the time required for each learning unit dsj (1) to dsj (xgs) and the time required for the learner to learn the combined unit zh; ; Where: glz2(u, v) represents the association value between learning units u and v; zh2(u, v) represents the time required to learn the combination of learning units u and v; dsj(u) refers to the time required to learn the u-th learning unit, and dsj(v) refers to the time required to learn the v-th learning unit. Step S223: Based on the combination unit, add a learning unit; obtain the time zh3 required for the learner to learn the combination unit; calculate the associated value according to the calculation formula of step S222 to obtain glz3; similarly obtain zh4 to zhxgs; calculate their corresponding associated values; accumulate the associated values, and take the combination route corresponding to the largest associated value as the first planned route.

[0012] Furthermore, the specific steps of step S3 are as follows: Step S31: The learner studies according to the first planned route; during the learner's learning process, the learner is given in-class exercises to obtain the learner's learning behavior; based on the learner's learning behavior, the learner's answer information is obtained; and based on the answer information, the learner's learning completion degree is calculated; Step S32: Analyze the learner's learning status based on the learner's personalized information and the learner's learning completion degree. Calculate the review weight based on the learner's learning status and arrange a review plan for the learner based on the review weight. Combine the review plan with the first planned route and optimize the first route.

[0013] Furthermore, the specific steps of step S31 are as follows: Step S311: Obtain the total score zfz of the in-class exercise; obtain the score kfz deducted due to mistakes in the in-class exercise; obtain the score wfz deducted due to unanswered questions in the in-class exercise; and calculate the learner's error value cwz; Step S312: Get the estimated time ygt for the in-class practice; get the number of questions tsl; get the learner's hand video; record the learner's hand pause time tds when completing each question t ; Get the learner's completion time wct of the in-class exercise; Calculate the learner's jam value kdz; ; Step S313: Calculate the learner's learning completion degree based on the learner's error value and the learner's jam value.

[0014] Furthermore, the specific steps of step S32 are as follows: Step S321: Obtain the learner's personalized information, i.e., the time series data of the student's state parameters generated by the multimodal large model; extract from the time series data the total duration qxt of the learner's negative emotional state and the total duration zyt of the learner's inattention during the learning process; obtain the learner's total learning time xxt; use the ratio of the learner's negative emotional time to the learner's learning time as the first weight, and use the ratio of the learner's inattention time to the learner's learning time as the second weight; Step S322: Obtain the learner's learning completion degree wcd; calculate by combining the first weight and the second weight, and multiply the learning completion degree by the first weight and the second weight to obtain the review weight fxq; Arrange time to review the learning units according to the review weights, sort the review weights in ascending order, and review in sequence according to the ranking of the review weights to obtain a review plan; use the review plan as a supplement to the first planned route to optimize the first planned route to obtain an optimized route.

[0015] Furthermore, the specific steps of step S4 are as follows: Step S41: Obtain the current learning time dqt, the preset learning time yst, and the ranking of the current learning unit in the optimized route; calculate the ratio of the ranking to the number of learning units to obtain the learning progress; and compare the ratio of the current learning time to the preset learning time with the learning progress. Step S42: The ratio of the current learning time to the preset learning time is recorded as the expected value; if the expected value is greater than the learning progress, the optimization route is adjusted: Calculate the difference between the expected value and the learning progress to obtain the learning deviation; calculate the learning deviation and the preset learning time to obtain the deviation time, and obtain the number of remaining learning units in the optimized route; obtain the adjusted time based on the deviation time and the number of remaining learning units, and compress the learning units in the optimized route based on the adjusted time to complete the route planning.

[0016] Compared with the prior art, the present invention has the following beneficial effects: Improve learners' personalized choices: Learners independently select their learning goals and study time. By judging learners' attention and analyzing their expressions, their study time can be improved. Based on the improved study time, learning goals are broken down into learning units, and learning paths are planned based on the learning units. By allowing learners to independently select their learning goals and study time, learners' learning needs are optimized, unnecessary content can be avoided from being repeated, and learning efficiency can be improved through the use of learning units. Planning learning paths based on time: Based on learning units, we analyze the learning time required for different learning unit combinations through big data. Based on the combination of learning units, we obtain time changes and optimize the learner's learning time. This path planning method helps learners achieve their learning goals within a certain time. Real-time adjustment of learning paths: The present invention adjusts the learning paths in real time according to the learner's learning status; optimizes the learner's learning experience, arranges more review time for difficult learning content, and reasonably arranges the time for different learning units in combination with preset time constraints. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments with reference to the following drawings: Figure 1 Schematic diagram of the method of the present invention; Figure 2 This is a schematic diagram of expression detection according to the present invention; Figure 3 This is a schematic diagram of attention recognition according to the present invention; Figure 4 This is a schematic diagram of the path planning of the present invention. DETAILED DESCRIPTION

[0018] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] Example 1 See also Figure 1 , the learning path planning methods based on habit analysis include: Step S1: Obtain the learner's personalized information; obtain the learner's learning goals and preset learning duration to form planning data; Step S11: acquiring real-time learning video data of the learner through a video capture device, and preprocessing the video data to generate high-quality video data; Step S12: Inputting the high-quality video data into a multimodal large model, the multimodal large model performs end-to-end deep visual understanding of the video frames to generate visual features containing information such as the learner's expression, attention, and posture; the multimodal large model further integrates the visual features with other modal data such as voice collected simultaneously to generate a comprehensive feature representation; It should be noted that the multimodal large model adopts a Transformer-based architecture, which divides video frames into image blocks and directly inputs them into the encoder. It learns the intrinsic correlation between image blocks through the self-attention mechanism to understand the image content as a whole, eliminating the need to pre-extract facial key points or design manual features.

[0020] Step S13: Based on the comprehensive feature representation, the video sequence is analyzed using a self-attention mechanism to generate quantified student state parameters, which constitute personalized information of the learner.

[0021] It should be noted that the student status parameters include but are not limited to: student expression categories, such as confusion, concentration or fatigue; concentration levels, such as high, medium and low; emotional intensity; and learning behaviors, such as wandering eyes, frequent blinking or changes in sitting posture.

[0022] See also Figure 2 Step S131: Detect the learner's facial expression; obtain an expression recognition dataset and classify the expression into categories, such as happiness, sadness, surprise, and confusion; learn the intrinsic associations between image blocks of the same type of expression through a self-attention mechanism to obtain association values; perform statistics on the association values ​​to obtain an association set tzj(i); construct an association function hs based on the association set; It should be noted that expression classification refers to the detection and classification of human facial expressions using the ultra-lightweight face detection model RFB-320; Obtain the learner's real-time learning video; obtain the learner's facial image, obtain the correlation value of the learner's facial image based on the self-attention mechanism, and perform expression detection on the learner in combination with the correlation function to obtain the detection result.

[0023] Step S1311: Obtain an association set tzj(i), extract the number of association values ​​t in the association set; record the association values ​​in the association set as tzz(1), tzz(2), ..., tzz(t); Step S1312: Obtain correlation values ​​tzz(1), tzz(2), ..., tzz(t); construct a correlation function hs based on the correlation values. The correlation function is as follows: ; It should be noted that x1, x2, ..., xt are not calculated values, but represent the correlation values ​​in different directions and are only used in the judgment process; Step S1313: Obtain the correlation values ​​of the learner's facial image, denoted as xtz(1), xtz(2), ..., xtz(t); perform calculations based on the correlation function hs to obtain the learner's correlation function xhs; ; Calculate the difference between the learner's correlation function xhs and the correlation function hs to obtain the difference in the directions of x1, x2, ..., xt; ; For example, xtz(1)-tzz(1) is the difference in the x1 direction; The absolute values ​​of the differences in the directions of x1, x2, ..., xt are accumulated and calculated to obtain the detection value; the expression category corresponding to the minimum detection value is obtained, and the learner's emotional intensity is obtained according to the self-attention mechanism. The expression category and emotional intensity are combined to obtain the learner's expression information.

[0024] Step S132: Obtain the learner's visual learning area and record it as the attention focus area; perform a line of sight test on the learner; determine whether the learner's attention is focused, and optimize the attention judgment based on the learner's blinking frequency to obtain the learner's attention information; It should be noted that gaze tracking (Eye Tracking) is a technology that uses technical means to capture and analyze the trajectory of human eye movements to determine the user's gaze direction or area of ​​attention. Gaze detection is performed based on infrared light source + optical imaging: a near-infrared light source (such as an LED) is used to illuminate the eyes, and a camera is used to capture the positional relationship between the corneal reflection (Purkinje's spot) and the center of the pupil. The gaze point is calculated using a geometric model (such as a 3D eyeball model).

[0025] See also Figure 3 Step S1321: Obtain the focused attention area, connect the starting point of the learner's line of sight with the center point of the focused attention area as a standard line; record the length of the standard line as bzc; perform edge detection on the focused attention area; test the angle between the line of sight and the standard line when the line of sight is at the edge of the focused attention area, and obtain the edge angle by; obtain the attention angle interval [0, by] from the edge angle; It should be noted that the attention focus area is the display area of ​​learning content during the learner's learning process.

[0026] Step S1322: Obtain the learner's gaze point through the gaze detection device, connect the gaze point with the center point of the attention focus area, and record the distance between the gaze point and the center point of the attention focus area as lyz. Combined with the length of the standard line bzc, the learner's gaze angle α is calculated. The specific calculation process is as follows: ; When the learner's line of sight angle α∈[0,by], the learner's line of sight is judged to be focused.

[0027] Step S1323: When the learner's line of sight is focused, the learner's blinking frequency is detected to obtain the learner's blinking frequencies zpl(1), zpl(2), ... zpl(z) in z daily states; the average of the blinking frequencies zpl(1) to zpl(z) is calculated to obtain the average frequency jpl; the learner's real-time blinking frequency spl is obtained; if spl>jpl, the learner is judged to be inattentive; the learner's concentration is obtained based on the learner's attention; Step S2: Determine the learner's continuous learning time based on the learner's personalized information; obtain the learner's learning goals based on the planning data, and make a preliminary plan for the learner's learning path based on the continuous learning time to obtain a first planned route.

[0028] Step S21: Obtaining personalized information about the learner during learning, and monitoring the learner's facial expression type, emotional intensity, and concentration level in real time; and determining the learner's continuous learning time based on the learner's facial expression type, emotional intensity, and concentration level during learning; Step S22: Obtain the learner's learning goal, and decompose the learner's learning goal based on the learner's continuous learning time to obtain learning units; use big data to obtain the correlation between the learning units to obtain the correlation value, and use the correlation value to preliminarily plan the learner's learning path; Step S221: Obtain the learner's learning goal and the learner's continuous learning time; obtain the estimated learning time for the learning goal through big data; and divide the estimated learning time into learning units according to the learner's continuous learning time; It should be noted that obtaining the estimated learning time for a learning goal through big data means obtaining the learning time of the majority of people for the learning goal. For example, for mathematical functions, the learning time of no less than 100 learners for the mathematical function is obtained, and the scores and the average of the good and bad are added up. The total learning time is then summed up, and based on the number of learners, the average learning time of each learner is calculated as the estimated learning time.

[0029] See also Figure 4 ; Step S222: Obtain the number of learning units xgs; Obtain the time required for each learning unit of the learner dsj (1), dsj (2), ..., dsj (xgs) through big data; Combine the learning units in pairs to obtain combined units; Obtain the time required for the learner to learn the combined units zh2; ; It should be noted that zh2(1,1) and zh2(2,2) are only used to express data integrity and do not participate in the specific calculation process. zh2(1,2) indicates the time required to learn the second learning unit after learning the first learning unit. The correlation value glz2 between learning units is calculated based on the time required for each learning unit dsj(1), dsj(2), ..., dsj(xgs) and the time required for the learner to learn the combined unit zh; ; It should be noted that: glz2(u, v) represents the association value between learning units u and v, zh2(u, v) represents the time required to learn the combined unit consisting of learning units u and v; dsj(u) refers to the time required to learn the u-th learning unit, and dsj(v) refers to the time required to learn the v-th learning unit. The calculation is based on the ratio of the time required to learn two learning units to the time required to learn the two units separately. The ratio reflects the reduction in time required to learn the two learning units together. By taking the difference from 1, the proportional relationship between the association value and the reduction time is consistent. Step S223: Based on the combination unit, add a learning unit; obtain the time zh3 required for the learner to learn the combination unit; taking zh2 (1, 2) as an example, obtain zh3 (1, 2, 3), zh3 (1, 2, 4), ..., zh3 (1, 2, xgs); calculate the association value according to the calculation formula of step S222 to obtain glz3; similarly, obtain zh4, zh5, ..., zhxgs; calculate their corresponding association values; accumulate the association values, and use the combination route corresponding to the largest association value as the first planned route; Step S3: Based on the first planned route, the learner's learning completion is tested, and personalized information of the learner is obtained in real time. The learner's learning status is analyzed based on the learner's learning completion, and a review plan is arranged for the learner. Based on the review plan, the first planned route is optimized; Step S31: The learner studies according to the first planned route; during the learner's learning process, the learner is given in-class exercises to obtain the learner's learning behavior; based on the learner's learning behavior, the learner's answer information is obtained; and based on the answer information, the learner's learning completion degree is calculated; Step S311: Obtain the total score zfz of the in-class exercise; obtain the score kfz deducted due to mistakes in the in-class exercise; obtain the score wfz deducted due to unanswered questions in the in-class exercise; and calculate the learner's error value cwz; ; It should be noted that: for the score kfz deducted due to error, it indicates that the learner understands the question but still has problems; the score wfz deducted due to unanswered question indicates that the learner does not understand the question at all. The two should be distinguished by different weights. The present invention classifies the score deducted due to error by adding 1 / 2 weight. Step S312: Get the estimated time ygt for the in-class practice; get the number of questions tsl; get the learner's hand video; record the learner's hand pause time tds when completing each question t ; Get the learner's completion time wct of the in-class exercise; Calculate the learner's jam value kdz; ; It should be noted that the learner's hand pause time for each question refers to the time the learner pauses during the answering process, from the start of the question to the end. For example, for question 1, the learner begins answering at time 0 and finishes at time 10. During the time 0-10, the learner pauses at times 2-4 and 6-8. Therefore, the learner's hand pause time for each question is (4-2) + (8-6).

[0030] It should be noted that: the jam value is preset to 1; the sign is extracted by the ratio of the difference between the learner's in-class exercise completion time wct and the estimated in-class exercise time ygt and their absolute value. When it is a negative sign, it indicates that the learner has learned the learning unit well; the jam value is reduced; if the learner's completion time is long and the jam time is long, the jam value is increased.

[0031] Step S313: Calculate the learner's learning completion degree wcd based on the learner's error value and the learner's jam value; ; Step S32: Analyze the learner's learning status based on the learner's personalized information and the learner's learning completion degree. Calculate the review weight based on the learner's learning status and arrange a review plan for the learner based on the review weight. Combine the review plan with the first planned route and optimize the first route.

[0032] Step S321: Obtain the learner's personalized information, i.e., the time series data of the student's state parameters generated by the multimodal large model; extract from the time series data the total duration qxt of the learner's negative emotional state and the total duration zyt of the learner's inattention during the learning process; obtain the learner's total learning time xxt; use the ratio of the learner's negative emotional time to the learner's learning time as the first weight, and use the ratio of the learner's inattention time to the learner's learning time as the second weight; Step S322: Obtain the learner's learning completion degree wcd; calculate by combining the first weight and the second weight, and multiply the learning completion degree by the first weight and the second weight to obtain the review weight fxq; It should be noted that the learning completion degree, the first weight, the second weight and the review weight are in direct proportion. When any value of the learning completion degree, the first weight and the second weight increases, the review weight increases.

[0033] Arrange time for review of learning units according to review weights, sort the review weights in ascending order, and review in order according to the order of review weights to obtain a review plan; use the review plan as a supplement to the first planned route to optimize the first planned route to obtain an optimized route; Step S4: Obtain the learner's preset learning time, dynamically adjust the optimized route according to the preset learning time, and complete the route planning; Step S41: Obtain the current learning time dqt, the preset learning time yst, and the ranking of the current learning unit in the optimized route; calculate the ratio of the ranking to the number of learning units to obtain the learning progress; and compare the ratio of the current learning time to the preset learning time with the learning progress. Step S42: The ratio of the current learning time to the preset learning time is recorded as the expected value; if the expected value is greater than the learning progress, the optimization route is adjusted: Calculate the difference between the expected value and the learning progress to obtain the learning deviation; multiply the learning deviation by the preset learning time to obtain the deviation time, and obtain the number of remaining learning units in the optimized route; divide the deviation time by the number of remaining learning units to obtain the adjusted time, and compress the learning units in the optimized route according to the adjusted time to complete the route planning.

[0034] The above formulas are all dimensionless and calculated by taking their numerical values. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions. For example, if there are weight coefficients and proportional coefficients, the size of the settings is to quantify each parameter to obtain a specific value, which is convenient for subsequent comparison. Regarding the size of the weight coefficient and the proportional coefficient, as long as it does not affect the proportional relationship between the parameter and the quantized value, it is fine.

[0035] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the above-described embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-described embodiments within the technical scope disclosed by the present invention, or replace some of the technical features therein with equivalents. Such modifications, changes, or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A learning path planning method based on habit analysis, characterized in that: Planning methods include: Step S1: Acquire learner's personalized information and planning data; Step S2: Determine the learner's continuous learning duration based on the learner's personalized information; perform preliminary planning of the learner's learning path based on the continuous learning duration and combined with the planning data to obtain a first planned route; Step S3: Based on the first planned route, the learner's learning completion is tested. Based on the learner's learning completion and combined with the learner's personalized information, the learner's learning status is analyzed, a review plan is arranged for the learner, and the first planned route is optimized based on the review plan; Step S4: Dynamically adjust the optimized route according to the planning data to complete the route planning.

2. The learning path planning method based on habit analysis according to claim 1, characterized in that: The specific steps of step S1 are as follows: Step S11: acquiring real-time learning video data of the learner through a video capture device, and preprocessing the video data to generate high-quality video data; Step S12: Inputting the high-quality video data into a multimodal large model, the multimodal large model performs end-to-end deep visual understanding of the video frames to generate visual features containing information such as the learner's expression, attention, and posture; the multimodal large model further integrates the visual features with other modal data such as voice collected simultaneously to generate a comprehensive feature representation; Step S13: Based on the comprehensive feature representation, the video sequence is analyzed using a self-attention mechanism to generate quantified student state parameters, which constitute personalized information of the learner.

3. The learning path planning method based on habit analysis according to claim 2, characterized in that: The multimodal large model adopts a Transformer-based architecture, divides video frames into image blocks and directly inputs them into the encoder. It learns the intrinsic correlation between image blocks through the self-attention mechanism to understand the image content as a whole, eliminating the need to pre-extract facial key points or design manual features.

4. The learning path planning method based on habit analysis according to claim 2, characterized in that: The student status parameters include but are not limited to: the student's expression type; concentration level; emotion intensity and learning behavior.

5. The learning path planning method based on habit analysis according to claim 1, characterized in that: The specific steps of step S2 are as follows: Step S21: Obtaining personalized information about the learner during learning, and monitoring the learner's facial expression type, emotional intensity, and concentration level in real time; and determining the learner's continuous learning time based on the learner's facial expression type, emotional intensity, and concentration level during learning; Step S22: Obtain the learner's learning goal, and decompose the learner's learning goal based on the learner's continuous learning time to obtain learning units; The correlation between learning units is calculated to obtain the correlation value, and the learner's learning path is preliminarily planned based on the correlation value.

6. The learning path planning method based on habit analysis according to claim 5, characterized in that: The specific steps of step S22 are as follows: Step S221: Obtain the learner's learning goal and the learner's continuous learning time; obtain the estimated learning time for the learning goal; and divide the estimated learning time into learning units according to the learner's continuous learning time. Step S222: Obtain the number of learning units xgs; obtain the time required for the learner to learn each learning unit dsj(1), dsj(2), ..., dsj(xgs); combine the learning units in pairs to obtain combined units; obtain the time zh2 required for the learner to learn the combined units; The correlation value glz2 between learning units is calculated based on the time required for each learning unit dsj (1) to dsj (xgs) and the time required for the learner to learn the combined unit zh; ； Where: glz2(u, v) represents the association value between learning units u and v; zh2(u, v) represents the time required to learn the combination of learning units u and v; dsj(u) refers to the time required to learn the u-th learning unit, and dsj(v) refers to the time required to learn the v-th learning unit. Step S223: Based on the combination unit, add a learning unit; obtain the time zh3 required for the learner to learn the combination unit; calculate the associated value according to the calculation formula of step S222 to obtain glz3; similarly obtain zh4 to zhxgs; calculate their corresponding associated values; accumulate the associated values, and take the combination route corresponding to the largest associated value as the first planned route.

7. The learning path planning method based on habit analysis according to claim 1, characterized in that: The specific steps of step S3 are as follows: Step S31: The learner studies according to the first planned route; during the learner's learning process, the learner is given in-class exercises to obtain the learner's learning behavior; based on the learner's learning behavior, the learner's answer information is obtained; and based on the answer information, the learner's learning completion degree is calculated; Step S32: Analyze the learner's learning status based on the learner's personalized information and the learner's learning completion degree, calculate the review weight based on the learner's learning status, and arrange a review plan for the learner based on the review weight; Combine the review plan with the first planned route and optimize the first route.

8. The learning path planning method based on habit analysis according to claim 7, characterized in that: The specific steps of step S31 are as follows: Step S311: Obtain the total score zfz of the in-class exercise; obtain the score kfz deducted due to mistakes in the in-class exercise; obtain the score wfz deducted due to unanswered questions in the in-class exercise; and calculate the learner's error value cwz; Step S312: Get the estimated time ygt for the in-class practice; get the number of questions tsl; get the learner's hand video; record the learner's hand pause time tds when completing each question t ; Get the learner's completion time wct of the in-class exercise; Calculate the learner's jam value kdz; ； Step S313: Calculate the learner's learning completion degree based on the learner's error value and the learner's jam value.

9. The learning path planning method based on habit analysis according to claim 7, characterized in that: The specific steps of step S32 are as follows: Step S321: Obtain the learner's personalized information, i.e., the time series data of the student's state parameters generated by the multimodal large model; extract from the time series data the total duration qxt of the learner's negative emotional state and the total duration zyt of the learner's inattention during the learning process; obtain the learner's total learning time xxt; use the ratio of the learner's negative emotional time to the learner's learning time as the first weight, and use the ratio of the learner's inattention time to the learner's learning time as the second weight; Step S322: Obtain the learner's learning completion degree wcd; calculate by combining the first weight and the second weight, and multiply the learning completion degree by the first weight and the second weight to obtain the review weight fxq; Arrange time to review the learning units according to the review weights, sort the review weights in ascending order, and review them in order of the review weight ranking to obtain a review plan; The review plan is used as a supplement to the first planned route to optimize the first planned route and obtain an optimized route.

10. The learning path planning method based on habit analysis according to claim 1, characterized in that: The specific steps of step S4 are as follows: Step S41: Obtain the current learning time dqt, the preset learning time yst, and the ranking of the current learning unit in the optimized route; calculate the ratio of the ranking to the number of learning units to obtain the learning progress; and compare the ratio of the current learning time to the preset learning time with the learning progress. Step S42: The ratio of the current learning time to the preset learning time is recorded as the expected value; if the expected value is greater than the learning progress, the optimization route is adjusted; Calculate the difference between the expected value and the learning progress to obtain the learning deviation; calculate the learning deviation and the preset learning time to obtain the deviation time, and obtain the number of remaining learning units in the optimized route; obtain the adjusted time based on the deviation time and the number of remaining learning units, and compress the learning units in the optimized route based on the adjusted time to complete the route planning.

Images