A method for english video split learning based on large model

Abstract

The present application relates to the technical field of video splitting, and particularly relates to a method for English video splitting learning based on a large model. The method acquires teaching text and voice transcription text of a teaching video through ASR and OCR technologies respectively, cuts the text into single sentences, extracts vocabulary and stores it in a difficult vocabulary library, integrates videos in accordance with examination syllabus requirements according to user expected learning level and difficulty quantification indexes, and cuts the video length according to user learning duration requirements; collects user playback records, repetition rates, complete play rates and interactive feedback data, comprehensively evaluates user learning level and video teaching difficulty through a big data model, and dynamically adjusts the video content. The scheme fills the limitations of traditional static splitting methods, effectively reduces knowledge omission and improves learning efficiency through dynamic optimization and personalized customization of teaching content, and provides precise matching teaching resources for learners of different levels, significantly improving learning efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of video splitting, in particular to a method for English video splitting learning based on a large model. BACKGROUND

[0002] With the wide application of mobile devices and the continuous improvement of network speed, short and concise content with high traffic propagation effect has gradually been favored by major platforms, user groups and capital markets. As a new form of Internet dissemination, short videos not only quickly become popular, but also help promote and spread related video resources.

[0003] The patent document with the Chinese patent publication number CN110493637A discloses a video splitting method and device, the technical point of which is to obtain video content features of a to-be-processed video and time information corresponding to the video content features; determine a plot transition point of the to-be-processed video according to the video content features and the time information corresponding to the video content features; and split the to-be-processed video according to the plot transition point. By obtaining the video content features of the to-be-processed video and the time information corresponding to the video content features, and determining the plot transition point of the to-be-processed video according to the video content features and the time information corresponding to the video content features, the to-be-processed video can be automatically split according to the plot content. At the same time, the invention focuses on static splitting of video content and detection of plot transition, does not introduce user behavior data and interaction feedback, and lacks adaptability to dynamic changes and personalized needs of teaching content, resulting in low flexibility and accuracy of the splitting result in actual application. SUMMARY

[0004] Therefore, the present application provides a method for English video splitting learning based on a large model to overcome the problem of low accuracy of video splitting due to the low level of user feedback integration and dynamic teaching adjustment in the prior art which focuses on static video content splitting.

[0005] To achieve the above-mentioned purpose, the present application provides a method for English video splitting learning based on a large model, comprising,

[0006] obtaining input video data and processing the data into standard input data, dividing the standard input data into standard integrated data according to the examination difficulty outline requirements, and obtaining a plurality of integrated videos;

[0007] According to the actual learning duration requirement of the user, the integrated video is cut into a plurality of integrated video segments, including fragmented videos and long videos;

[0008] For any of the integrated video segments, actual video watching information of the user is acquired, and learning progress is analyzed based on the actual video watching information, a video content adjustment direction is determined according to an analysis result, a video content adjustment mode is determined based on the video content adjustment direction, and an adjustment result is fed back to the user end.

[0009] Further, acquiring input video data and processing the data into standard input data includes,

[0010] Acquiring a video teaching text and a voice transcription text;

[0011] The text content is cut into a plurality of individual sentences, the parts of speech of words and phrases in any individual sentence are extracted and stored in a corresponding difficulty vocabulary database, corresponding examination syllabus requirements or application ability requirements are determined according to learning level expectation values, and a plurality of integrated videos are integrated according to difficulty quantification indexes.

[0012] Further, the plurality of integrated videos integrated according to the difficulty quantification indexes include,

[0013] A standard matching interval is set, and an actual matching degree of the corresponding examination syllabus requirements or application ability requirements is determined according to the actual learning level expectation value of the user;

[0014] The actual matching degree is determined according to the standard matching interval, and the type of the integrated video content is determined according to the determination result, including a first video type, a second video type, and a third video type.

[0015] Further, the integrated video is cut into,

[0016] The integrated video is cut according to the user feedback:

[0017] If the user feedback is a selection of fragmented time quick learning requirement, the integrated video is cut into fragmented video;

[0018] If the user feedback is a selection of continuous deep learning requirement, the integrated video is cut into long video.

[0019] Further, the actual video watching information of the user is acquired, and the learning progress is analyzed based on the actual video watching information,

[0020] The actual video watching information includes a repetition rate, a complete play rate, and a repetition complete play index, the video watching information is evaluated and compared with a set preset threshold;

[0021] If the repetition complete play index meets the evaluation standard, it is determined that the user learning progress matches the difficulty progress requirement, and a user learning condition feedback measure is executed;

[0022] If the repeated complete play index does not meet the evaluation standard, it is determined that the user's learning progress and the current difficulty progress requirement do not match, the user's current learning bottleneck category is determined, and the video content adjustment direction is determined.

[0023] Further, determining the user's current learning bottleneck category and determining the video content adjustment direction includes,

[0024] The learning bottleneck category includes a first result, a second result, a third result, and a fourth result;

[0025] When the learning bottleneck category is determined to be the first result, a subtitle error adjustment program is executed;

[0026] When the learning bottleneck category is determined to be the second result, an example supplement program is executed;

[0027] When the learning bottleneck category is determined to be the third result, an anchor point play program is executed;

[0028] When the learning bottleneck category is determined to be the fourth result, a split interval adjustment program is executed;

[0029] Further, executing the subtitle error adjustment program includes,

[0030] The real-time delay duration of the subtitle is read and compared with the standard delay duration,

[0031] When the real-time delay duration is less than or equal to the standard delay duration, the subtitle is normal;

[0032] When the real-time delay duration is greater than the standard delay duration, the subtitle is abnormal;

[0033] If the determination result is that the subtitle is normal, the corresponding extension parameter is obtained by analyzing the learning content through a large model, and the video duration is extended according to the parameter;

[0034] If the determination result is that the subtitle is abnormal, the subtitle display parameter is adjusted and the subtitle text content is verified, the adjusted video viewing information is read and the information is evaluated;

[0035] If the evaluation result does not meet the evaluation standard and is excessively repeated, the corresponding extension parameter is obtained by analyzing the learning content through a large model, and the video duration is extended according to the parameter.

[0036] Further, executing the example supplement program includes,

[0037] According to the type of the current integrated video content, the teaching difficulty level is confirmed, and the language logic template under the difficulty is extracted;

[0038] A number of core words in the current integrated video content that match the teaching difficulty level are selected from the corresponding difficulty vocabulary library;

[0039] According to the language logic template, the semantics between the core words are associated, candidate example sentences are generated, and stored in the candidate example sentence information library, and whether to call is judged according to the actual needs of the user.

[0040] Further, the anchor point playing procedure comprises,

[0041] Confirming the starting position of the repeated interval, setting the anchor point at the starting position, and quickly playing back the repeated content from the anchor point;

[0042] Confirming the user playing behavior in the standard judgment duration, and dynamically adjusting the anchor point position.

[0043] Further, the split interval adjustment procedure comprises,

[0044] Extracting the playing information of the target fast-forward interval, evaluating the current learning level of the user according to the playing information and quantifying it into an actual learning level value;

[0045] Extracting the playing information of the integrated video, evaluating the current teaching level of the video according to the playing information and quantifying it into an actual teaching level value;

[0046] The playing information includes user historical learning records and interactive feedback;

[0047] Comparing the actual teaching level value with the actual learning level value,

[0048] When the actual teaching level value is greater than or equal to the actual learning level value, the target fast-forward interval is a not-yet-completely-mastered interval;

[0049] When the actual teaching level value is less than the actual learning level value, the target fast-forward interval is a completely mastered interval;

[0050] When the result is a completely mastered interval, the target fast-forward interval is removed.

[0051] Compared with the prior art, the beneficial effects of the present application are that by introducing a large model and multi-dimensional data fusion technology, intelligent splitting and personalized adjustment of video content are realized; compared with the traditional video splitting method which only relies on video content features and fixed time information, the present method first accurately extracts the key information in the video by using speech transcription and text analysis, and combines the actual viewing behavior of the user and the learning progress feedback to realize dynamic evaluation of the video content and introduce the complete play rate, repetition rate and other indicators, accurately identify the bottleneck area in the learning process of the user, and thus perform targeted optimization measures such as subtitle correction, example sentence supplementation, anchor point playing and split interval adjustment, eliminate redundant information while avoiding knowledge omission, realize dynamic customization and continuous optimization of teaching resources under different levels of teaching needs, and significantly improve learning efficiency and personalized teaching experience. BRIEF DESCRIPTION OF DRAWINGS

[0052] Figure 1 FIG. 1 shows a flowchart of a method for English video splitting learning based on a large model according to an embodiment of the present application;

[0053] Figure 2 FIG. 4 shows a logic decision diagram for integrating a plurality of integrated videos according to difficulty quantification indexes according to an embodiment of the present application;

[0054] Figure 3 FIG. 5 shows a logic decision diagram for determining a current learning bottleneck category of a user and judging a video content adjustment direction according to an embodiment of the present application;

[0055] Figure 4 FIG. 6 shows a logic decision diagram for executing a subtitle error adjustment program according to an embodiment of the present application. DETAILED DESCRIPTION

[0056] In order to make the objects and advantages of the present application clearer, the present application will be further described below with reference to embodiments. It should be understood that the specific embodiments described herein merely serve to explain the present application and should not be used to limit the present application.

[0057] The preferred embodiments of the present application will be described below with reference to the accompanying drawings. It should be understood by those skilled in the art that the embodiments merely serve to explain the technical principles of the present application and are not used to limit the protection scope of the present application.

[0058] It should be noted that, in the description of the present application, the terms indicating the direction or positional relationship such as "upper", "lower", "left", "right", "inner", "outer" and the like are based on the direction or positional relationship shown in the drawings, which is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application.

[0059] In addition, it should also be noted that, in the description of the present application, unless otherwise explicitly specified and limited, the terms "mounting", "connecting", "connection" should be understood in a broad sense, for example, can be fixed connection, can also be detachable connection, or integrally connected; can be mechanical connection, can also be electrical connection; can be directly connected, can also be indirectly connected through an intermediate medium, can be the communication inside two elements. Those skilled in the art can understand the specific meaning of the above terms in the present application according to the specific circumstances.

[0060] Please refer to Figure 1 FIG. 1 shows a flowchart of a method for English video splitting learning based on a large model according to an embodiment of the present application, and the present application provides a method for English video splitting learning based on a large model, which comprises,

[0061] Acquire input video data and process the data into standard input data, divide the standard input data according to the requirements of the examination difficulty outline to obtain a plurality of integrated videos;

[0062] According to the actual learning time requirement of the user, the integrated video is divided to obtain a plurality of integrated video segments, including fragmented video and long video;

[0063] For any integrated video segment, the actual video viewing information of the user is acquired, and the learning progress is analyzed based on the actual video viewing information, the video content adjustment direction is determined according to the analysis result, the video content adjustment mode is determined based on the video content adjustment direction, and the adjustment result is fed back to the user end;

[0064] By introducing a large model and multi-dimensional data fusion technology, intelligent splitting and personalized adjustment of video content are realized. Compared with the traditional video splitting method which only relies on video content features and fixed time information, this method first accurately extracts the key information in the video by using speech transcription and text analysis, and combines the actual viewing behavior and learning progress feedback of the user to realize dynamic evaluation of the video content and introduce the completion rate, repeat rate and other indicators., accurately identify the bottleneck area in the learning process of the user, and then perform subtitle correction, example supplement, anchor point playback and split interval adjustment optimization measures to eliminate redundant information while avoiding knowledge omission. Under different levels of teaching needs, dynamic customization and continuous optimization of teaching resources are realized, and the learning efficiency and personalized teaching experience are significantly improved.

[0065] Specifically, acquiring input video data and processing the data into standard input data includes,

[0066] Acquiring video teaching text and speech transcription text;

[0067] Split the text content into a plurality of individual sentences, extract the parts of speech of the words and phrases in any individual sentence, and store them in the corresponding difficulty vocabulary database, determine the corresponding examination outline requirements or application ability requirements according to the learning level expectation value, and integrate a plurality of integrated videos according to the difficulty quantification index;

[0068] In this embodiment, the steps of acquiring video teaching text and speech transcription text are,

[0069] By integrating an automatic speech recognition (ASR) system and an optical character recognition (OCR) technology, the teaching text and the speech transcription text in the teaching video are extracted;

[0070] The integrated automatic speech recognition system recognizes the speech in the video in real time to generate a preliminary text;

[0071] The optical character recognition technology recognizes the subtitle or board information embedded in the video;

[0072] The step of cutting the text content into individual sentences, extracting the parts of speech of the words and phrases in the individual sentences, and storing them in the corresponding difficulty vocabulary database in the embodiment is,

[0073] Cutting the text content according to the linguistic rules and punctuation boundaries, for the sentence "The students listen carefully and think actively.", it is identified as two logically continuous but independent semantic units according to the punctuation;

[0074] Extracting words and phrases and annotating the parts of speech, in the sentence "The quick brown fox jumps over the lazy dog.", "quick" is annotated as an adjective, "fox" as a noun, and "jumps" as a verb. The system stores the words and their part of speech information in the corresponding difficulty vocabulary database. The vocabulary meeting the requirements of the College English Test Band 4 is stored in the Band 4 difficulty database, and the vocabulary meeting the requirements of the College English Test Band 6 is classified into the Band 6 difficulty database;

[0075] After standardizing the two kinds of data, a unified standard input data is formed, providing a numerical basis for subsequent processing.

[0076] Referring to Figure 2 As shown in the figure, the logical decision diagram of the integrated video obtained by the embodiment of the application according to the difficulty quantification index is shown;

[0077] Specifically, the integrated video obtained according to the difficulty quantification index includes,

[0078] Setting a standard matching interval, determining the actual matching degree of the requirements of the examination syllabus or application ability according to the actual learning level expectation value of the user;

[0079] According to the standard matching interval, the actual matching degree is determined, and according to the determination result, the type of the integrated video content is determined, including the first video type, the second video type and the third video type;

[0080] In the embodiment, according to the learning level expectation value set by the user, the difficulty of the extracted vocabulary and sentences is quantitatively integrated by referring to the requirements of the corresponding examination syllabus. The difficulty quantification index calculation formula is:

[0081] D I =0.4*V ratio +0.35*S complexity +0.25*G difficulty

[0082] Wherein,

[0083] D I : actual matching degree;

[0084] Vratio : proportion of words beyond the basic vocabulary table;

[0085] S complexity : proportion of complex sentence structures;

[0086] G difficulty : frequency of complex grammatical structures;

[0087] 0.4, 0.35 and 0.25 are the weights of V ratio , S complexity and G difficulty ;

[0088] If the user sets the learning level expectation value as the fourth level examination, the matching interval is divided,

[0089] wherein [0.4, 0.6] is the standard matching interval for the fourth level examination;

[0090] When D I is in the interval [0.6, 0.8], the integrated video is the third video type, corresponding to the fourth level out-of-class difficulty content;

[0091] When D I is in the interval [0.4, 0.6], the integrated video is the second video type, corresponding to the fourth level standard difficulty content;

[0092] When D I is below 0.4, the integrated video is the first video type, corresponding to the fourth level basic difficulty content;

[0093] This quantitative method not only provides data support for the generation of integrated videos, but also ensures that the generated video content accurately matches the user's expectations in terms of difficulty.

[0094] Specifically, the segmentation of the integrated video includes,

[0095] Segmenting the integrated video according to user feedback:

[0096] If the user feedback is the selection of fragmented time quick learning needs, the integrated video is segmented into fragmented videos;

[0097] If the user feedback is the selection of continuous deep learning needs, the integrated video is segmented into long videos;

[0098] Wherein, the fragmented video duration is within 3 to 7 minutes, meeting the fragmented time quick learning needs, and utilizing fragmented time for quick information transmission and review;

[0099] The long video duration is within 40 to 60 minutes, meeting the continuous deep learning needs, suitable for continuous deep explanation and complete presentation of complex concepts;

[0100] Different cuts of the integrated video reduce cognitive load in a fragmented time learning environment, facilitate repeated viewing, and strengthen memory effect; in a long time learning environment, it provides learners with systematic and coherent knowledge structure, helps to understand and digest more complex content, and meets the needs of continuous learning.

[0101] Specifically, actual video viewing information of the user is acquired, and learning progress is analyzed based on the actual video viewing information, including,

[0102] The actual video viewing information includes repetition rate, completion rate, and repetition completion index, and the video viewing information is evaluated and compared with a set preset threshold value;

[0103] If the repetition completion index meets the evaluation standard, it is determined that the user's learning progress matches the difficulty progress requirement, and a user learning feedback measure is executed;

[0104] If the repetition completion index does not meet the evaluation standard, it is determined that the user's learning progress does not match the current difficulty progress requirement, the user's current learning bottleneck category is determined, and the video content adjustment direction is determined;

[0105] The specific judgment standard and calculation formula of the repetition rate and the completion rate data in this embodiment are,

[0106] Repetition rate = (total repetition content duration / total video duration) x 100%;

[0107] Completion rate = (number of users who watched the video completely / total number of users who watched the video) x 100%;

[0108] In order to more finely reflect the influence of repetition and completion on learning effect, a comprehensive index, repetition completion index, can be introduced, and the calculation formula of the index is,

[0109] Repetition completion index = a x repetition rate + b x (1 - completion rate)

[0110] Wherein, a and b are weight coefficients, which can be dynamically adjusted according to learning difficulty;

[0111] Wherein, the first video type a = 0.6, b = 0.8;

[0112] The second video type a = 0.5, b = 0.5;

[0113] The second video type a = 0.4, b = 0.2;

[0114] The set preset threshold value = 0.4,

[0115] When the repetition completion index is greater than or equal to the preset threshold value, the video has too high repetition problem, which needs to be fed back and adjusted;

[0116] When the repeated complete play index is less than a preset threshold, the video content does not have a repetition problem;

[0117] The division method quantifies the redundancy of the video content and combines the actual viewing data of the user, thereby providing a scientific basis for intelligent feedback and optimization adjustment.

[0118] Referring to Figure 3 As shown in FIG. 4, which is a logical determination diagram for determining the current learning bottleneck category of the user and judging the adjustment direction of the video content according to an embodiment of the present application;

[0119] Specifically, determining the current learning bottleneck category of the user and judging the adjustment direction of the video content includes,

[0120] The learning bottleneck category includes a first result, a second result, a third result, and a fourth result.

[0121] When the learning bottleneck category is determined to be the first result, a subtitle error adjustment program is executed.

[0122] When the learning bottleneck category is determined to be the second result, an example supplement program is executed.

[0123] When the learning bottleneck category is determined to be the third result, an anchor point play program is executed.

[0124] When the learning bottleneck category is determined to be the fourth result, a split interval adjustment program is executed.

[0125] When the learning bottleneck category is determined to be the first result, the key points of the fragmented video are excessively repeated.

[0126] When the learning bottleneck category is determined to be the second result, the key points of the long video are excessively repeated.

[0127] When the learning bottleneck category is determined to be the third result, the non-key points of the fragmented video are excessively fast-forwarded.

[0128] When the learning bottleneck category is determined to be the fourth result, the non-key points of the long video are excessively fast-forwarded.

[0129] Referring to FIG. 5, which is a logical determination diagram for executing the subtitle error adjustment program according to an embodiment of the present application;

[0130] Specifically, executing the subtitle error adjustment program includes,

[0131] reading the real-time delay duration of the subtitle and comparing it with a standard delay duration,

[0132] When the real-time delay duration is less than or equal to the standard delay duration, the subtitle is normal.

[0133] When the real-time delay duration is greater than the standard delay duration, the subtitle is abnormal.

[0134] If the judgment result is that the subtitle is normal, the corresponding extension parameter is obtained by analyzing the learning content through the large model, and the video duration is extended according to the parameter;

[0135] If the judgment result is that the subtitle is abnormal, the subtitle display parameter is adjusted, the subtitle text content is checked, the adjusted video viewing information is read, and the information is evaluated;

[0136] If the evaluation result does not meet the evaluation standard and is excessively repeated, the corresponding extension parameter is obtained by analyzing the learning content through the large model, and the video duration is extended according to the parameter;

[0137] In this embodiment, the standard delay duration is set to 100 milliseconds, the open source tool LanguageTool is selected, and the subtitle text content is checked in combination with the Python script;

[0138] When the actual delay between the subtitle and the audio does not exceed 100 milliseconds, the subtitle synchronization is normal;

[0139] When the actual delay between the subtitle and the audio exceeds 100 milliseconds, the subtitle has a delay anomaly;

[0140] After detecting the subtitle anomaly, the subtitle display parameter is adjusted, the display parameters include the subtitle display start time, position and font, and the subtitle text is automatically checked to ensure that the text content is consistent with the voice content;

[0141] The standard repetition rate is set to 50%, and the parameter calculation formula is,

[0142] T = (actual repetition rate - standard repetition rate) x 0.4

[0143] In this embodiment, the actual repetition rate is 65%, T = 15 x 0.4 = 6 seconds;

[0144] After the adjustment is completed, the system will collect real-time viewing data again, if the evaluation result still does not meet the standard and the repetition rate is high, the above calculation process will be repeated, and the video duration will be gradually extended, until the evaluation standard is met or the upper limit of the duration is reached;

[0145] While ensuring the accuracy of the subtitle synchronization, the video duration is dynamically adjusted according to the user viewing behavior, and the learning experience of the fragmented key learning content is improved.

[0146] Specifically, the execution example supplement program includes,

[0147] According to the type of the current integrated video content, the teaching difficulty level is confirmed, and the language logic template under the difficulty is extracted;

[0148] A number of core words in the current integrated video content that match the teaching difficulty level are selected from the corresponding difficulty vocabulary library;

[0149] According to the language logic template, the semantics between the core words are associated, candidate example sentences are generated, and stored in the candidate example sentence information library, and whether to call is judged according to the actual needs of the user;

[0150] In the context of learning the attributive clause, the attributive clause pattern is selected as the language logic template;

[0151] The difficult vocabulary library is the college English test four vocabulary library, and the system extracts the key words meeting the four level difficulty requirements by matching the vocabulary appearing in the video content with the vocabulary in the vocabulary library;

[0152] The example sentence generation is based on the basic structure of the template attributive clause to form a reasonable sentence,

[0153] Among them, the candidate example sentence includes video example sentence and text example sentence;

[0154] The video example sentence is extracted from the video teaching text and image;

[0155] The text example sentence is extracted from the speech transcription text;

[0156] After the system checks the grammar and semantics of the two types of candidate example sentences, the newly added candidate example sentences that pass the screening are marked in the repeated interval of the video, so as to facilitate the learners to review and refer to the targeted.

[0157] Specifically, the anchor point playback program includes,

[0158] Confirming the starting position of the repeated interval, setting an anchor point at the starting position to quickly play back the repeated content from the anchor point;

[0159] Confirming the user's playback behavior in the standard judgment time, to dynamically adjust the anchor point position;

[0160] Specifically, the split interval adjustment program includes,

[0161] Extracting the playback information of the target fast-forward interval, evaluating the current learning level of the user according to the playback information and quantifying it into an actual learning level value;

[0162] Extracting the playback information of the integrated video, evaluating the current teaching level of the video according to the playback information and quantifying it into an actual teaching level value;

[0163] Among them, the playback information includes user historical learning records and interactive feedback;

[0164] Comparing the actual teaching level value with the actual learning level value,

[0165] When the actual teaching level value is greater than or equal to the actual learning level value, the target fast-forward interval is the interval that has not been completely mastered;

[0166] When the actual teaching level value is less than the actual learning level value, the target fast-forward interval is the complete mastery interval;

[0167] When the result is the complete mastery interval, the target fast-forward interval is removed.

[0168] The target fast-forward interval is a video interval corresponding to high-frequency fast-forward actual operation.

[0169] In this embodiment, the standard fast-forward operation frequency value is four times per minute, and the user's actual fast-forward operation frequency value is obtained and compared with the standard fast-forward operation frequency value,

[0170] When the user's actual fast-forward operation frequency value is greater than or equal to the standard fast-forward operation frequency value, the actual operation is high-frequency fast-forward.

[0171] When the user's actual fast-forward operation frequency value is less than the standard fast-forward operation frequency value, the actual operation is non-high-frequency fast-forward.

[0172] Read the play information, extract the user's historical learning records, including the repetition rate, the complete play rate, and the historical play records; collect user interaction feedback data, including pause, mark, and active jump operation, to form an interaction feedback score;

[0173] In this embodiment, the step is processed and evaluated by a large model, and the specific steps are,

[0174] Read and normalize each data, and construct a feature vector containing dimensions such as repetition rate, complete play rate, interaction feedback, and viewing duration; and adopt noise reduction and missing value filling techniques to ensure data quality;

[0175] Input the preprocessed feature vector into a neural network model pre-trained on English teaching input videos, and analyze the historical play records, repetition rate, complete play rate, and after-school exercise completion through the model, to output an actual learning level value L user in the range of 0 to 100, which reflects the user's mastery of the overall content of the current integrated video through the target fast-forward interval;

[0176] Extract text and video frame features from the integrated video content, construct a video teaching difficulty feature vector including vocabulary difficulty, average sentence length, long sentence proportion, and grammar structure complexity, input the big data model, and output a video actual teaching level value L video in the range of 0 to 100, which reflects the difficulty of the current integrated video content teaching;

[0177] Compare the actual teaching level value L video with the actual learning level value L user

[0178] ​When the actual teaching level value L video is greater than or equal to the actual learning level value L user , the target fast-forward interval is the not-yet-completely-mastered interval;

[0179] When the actual teaching level value L video is less than the actual learning level value L user , the target fast-forward interval is the completely-mastered interval;

[0180] Performing a content pruning operation on the completely-mastered target fast-forward interval removes the interval from the integrated video, leaving the not-yet-completely-mastered interval as the focus of subsequent review and learning by the user;

[0181] The comprehensive judgment method based on the big data model visualizes the learning level of the user and the teaching difficulty of the video by refining data collection, feature engineering, model reasoning, and interval comparison steps, and realizes dynamic adjustment and personalized optimization of the target fast-forward interval, so that the teaching method accurately meets the personalized needs of the user.

[0182] The technical solutions of the present application have been described in conjunction with the preferred embodiments shown in the accompanying drawings, but those skilled in the art will readily understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to related technical features without departing from the principles of the present application, and the technical solutions after such changes or replacements will fall within the protection scope of the present application.

[0183] The above description is only the preferred embodiments of the present application and is not intended to limit the present application; for those skilled in the art, the present application can have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included within the protection scope of the present application.

Claims

1. A method for large model-based English video split learning, characterized in that, The application relates to a video teaching method and device. The application comprises, standard input data is obtained by processing input video data, and the standard input data is divided according to examination difficulty outline requirements to obtain a plurality of integrated videos; the integrated videos are cut according to actual learning time requirements of a user to obtain a plurality of integrated video segments, including fragmented videos and long videos; for any integrated video segment, actual video watching information of the user is obtained, learning progress is analyzed based on the actual video watching information, a video content adjustment direction is determined according to the analysis result, a video content adjustment mode is determined based on the video content adjustment direction, and the adjustment result is fed back to the user end; the actual video watching information of the user is obtained, and learning progress is analyzed based on the actual video watching information, including, the video watching information is evaluated and compared with a preset threshold value according to repetition rate, completion rate and repetition completion index; if the repetition completion index meets the evaluation standard, it is determined that the learning progress of the user matches the difficulty progress requirement, and a user learning condition feedback measure is executed; if the repetition completion index does not meet the evaluation standard, it is determined that the learning progress of the user does not match the current difficulty progress requirement, a user current learning bottleneck type is determined, and a video content adjustment direction is determined; the user current learning bottleneck type is determined, and the video content adjustment direction is determined, including, the learning bottleneck type includes a first result, a second result, a third result and a fourth result; when the learning bottleneck type is determined as the first result, a subtitle error adjustment program is executed; when the learning bottleneck type is determined as the second result, an example supplement program is executed; when the learning bottleneck type is determined as the third result, an anchor point playing program is executed; 2. The method of claim 1, wherein, when the learning bottleneck type is determined as the fourth result, a cutting interval adjustment program is executed. the input video data is obtained, and the data is processed into standard input data, including, video teaching texts and voice transcription texts are obtained; 3. The method of claim 2, wherein, the text content is cut into a plurality of individual sentences, the parts of speech of words and phrases in any individual sentence are extracted and stored in a corresponding difficulty vocabulary database, corresponding examination outline requirements or application ability requirements are determined according to learning level expectation values, and a plurality of integrated videos are obtained according to difficulty quantization indexes. the plurality of integrated videos are obtained according to the difficulty quantization indexes, including, a standard matching interval is set, actual matching degrees of the corresponding examination outline requirements or application ability requirements are determined according to actual learning level expectation values of the user; 4. The method of claim 1, wherein, the actual matching degrees are determined according to the standard matching interval, the types of the integrated video contents are determined according to the determination result, including a first video type, a second video type and a third video type. the integrated videos are cut, including, the integrated videos are cut according to user feedback: if the user feedback is a selection of fragmented time fast learning requirement, the integrated videos are cut into fragmented videos; 5. The method of claim 1, wherein, if the user feedback is a selection of continuous deep learning requirement, the integrated videos are cut into long videos. the subtitle error adjustment program is executed, including, a real-time delay time length of a subtitle is read, and is compared with a standard delay time length, when the real-time delay time length is less than or equal to the standard delay time length, the subtitle is normal; when the real-time delay time length is greater than the standard delay time length, the subtitle is abnormal; If the judgment result is that the subtitles are normal, the corresponding extension parameter is obtained by analyzing the learning content through the large model, and the video duration is extended according to the parameter; If the judgment result is that the subtitles are abnormal, adjust the subtitle display parameter and check the subtitle text content, read the adjusted video viewing information and evaluate the information; If the evaluation result does not meet the evaluation standard and is excessively repetitive, the corresponding extension parameter is obtained by analyzing the learning content through the large model, and the video duration is extended according to the parameter.

6. The method of large model-based English video split learning according to claim 1, wherein, The execution example supplement program includes, According to the type of the current integrated video content, confirm the teaching difficulty level, and extract the language logic template under the difficulty level; Filter out a number of core words in the current integrated video content that match the teaching difficulty level in the corresponding difficulty vocabulary library; According to the language logic template, associate the semantics between each of the core words, generate candidate example sentences, and store them in the candidate example sentence information library. Whether to call according to the actual needs of the user.

7. The method of claim 1, wherein, The anchor point playback program includes, Confirm the starting position of the repeated interval, set the anchor point at the starting position, and quickly play back the repeated content from the anchor point; Confirm the user's playback behavior within the standard judgment duration to dynamically adjust the anchor point position.

8. The method for large model-based English video split learning according to claim 1, wherein, The cut interval adjustment program includes, Extract the playback information of the target fast-forward interval, evaluate the user's current learning level according to the playback information and quantify it into an actual learning level value; Extract the playback information of the integrated video, evaluate the current teaching level of the video according to the playback information and quantify it into an actual teaching level value; Wherein, the playback information includes user historical learning records and interactive feedback; Compare the actual teaching level value with the actual learning level value, When the actual teaching level value is greater than or equal to the actual learning level value, the target fast-forward interval is the interval that has not been completely mastered; When the actual teaching level value is less than the actual learning level value, the target fast-forward interval is the completely mastered interval; Wherein, when the result is the completely mastered interval, remove the target fast-forward interval.

Images