A sign language learning progress tracking and personalized recommendation management system and method
By generating a learning progress tracking model and identifying learning preferences, a personalized learning path is constructed, which solves the problem that existing sign language learning resources cannot adapt to the individual differences of the hearing impaired, and achieves efficient and personalized sign language learning management.
Patent Information
- Application Number
- CN202411117786.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-15
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-08-15
AI Technical Summary
Existing sign language learning resources lack personalization and dynamic adjustment capabilities, failing to adapt to individual differences among hearing-impaired individuals with varying degrees of hearing loss, resulting in low learning efficiency.
By collecting basic information about hearing-impaired individuals, a learning progress tracking model is generated. Behavioral data is recorded in real time to adjust the model, identify learning preferences, build personalized learning path planning, and provide personalized learning path recommendations.
It improves the efficiency and effectiveness of sign language learning, meets the personalized needs of the hearing impaired, enhances learning motivation and participation, and optimizes the coherence and suitability of the learning path.
Smart Images

Figure CN118643073B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of a data processing system and method for management, and in particular to the technical field of education, and specifically to a management system and method for tracking the progress of sign language learning and personalized recommendation. Background Art
[0002] At present, sign language, as an important visual language, is a key tool for communication and expression for the hard of hearing who retain a certain degree of hearing. As society's demand for inclusive and barrier-free communication grows, sign language learning has received more extensive attention. However, traditional sign language learning resources, such as books and dictionaries, as well as video tutorials, although providing basic learning and imitation opportunities, often lack the ability to personalize and dynamically adjust learning paths, which is not efficient enough for learners at different learning stages and needs.
[0003] With the popularity of massive open online sign language learning courses, personalized recommendations have gradually broken through small-scale and are aimed at large-scale sign language learners. It focuses on providing personalized sign language recommendation learning through the collection and mining of massive sign language learning resources and process data, and continuously promotes sign language learners to complete learning tasks efficiently through progress tracking and optimizing the recommended content. Among them, the hard of hearing face different communication challenges from people with normal hearing. Different degrees of hearing loss lead to different learning situations. They need an educational tool that can adapt to their unique learning methods and rhythms. Each hard of hearing person with different degrees of hearing loss has different learning styles, learning progress, speed and interests. Personalized learning can improve learners' participation and efficiency and help them master sign language skills faster.
[0004] Therefore, it is necessary to improve the sign language learning recommendation management system in the prior art to solve the above problems. Summary of the invention
[0005] The present invention overcomes the deficiencies of the prior art and provides a sign language learning progress tracking and personalized recommendation management system and method, aiming to solve the problem of providing personalized sign language learning recommendations for hearing-impaired people with different degrees of hearing loss in the prior art.
[0006] To achieve the above object, the technical solution adopted by the present invention is: a method for tracking the progress of sign language learning and providing personalized recommendations, comprising:
[0007] S1. Collect basic information of the hard of hearing sign language learner and generate a first learning progress tracking model;
[0008] S2, real-time recording of the behavior data of the hard of hearing person in the sign language learning process, determining the degree of the hard of hearing person, modifying the first learning progress tracking model according to the degree of the hard of hearing, and obtaining the second learning progress tracking model to obtain the learning progress of the hard of hearing person;
[0009] S3, identify the learning preferences of the hard of hearing students based on the behavioral data in S2, build a dynamic evaluation model for learning preferences, and evaluate the preferences of the hard of hearing students for different sign language learning content, resources, and methods;
[0010] S4, according to the preference output in S3 and the learning progress of the corresponding hard of hearing person, a personalized learning path is planned for the hard of hearing person;
[0011] S5. Push the personalized learning path planning to the corresponding hard of hearing person.
[0012] In a preferred embodiment of the present invention, basic information of the hearing impaired is collected during registration, including: age, hearing impairment level, use of assistive devices, background in learning sign language, and learning goals; the hearing impairment levels include: severe hearing impairment, moderate hearing impairment, and mild hearing impairment.
[0013] In a preferred embodiment of the present invention, in step S1, the step of generating a first learning progress tracking model includes:
[0014] S11. Conduct correlation analysis using basic information of hard of hearing persons;
[0015] S12. Assign weights to basic information that affects learning progress;
[0016] S13. Output a first learning progress tracking model.
[0017] In a preferred embodiment of the present invention, the basic information is used as an influencing factor for constructing a first learning progress tracking model. and learning progress Calculate Pearson correlation coefficient : ;
[0018] in, It is The sample in The value of the influencing factors, For the The learning progress of samples, and is the average value, is the sample size of hard of hearing subjects;
[0019] Standardize the correlation coefficient : , m is the total number of influencing factors;
[0020] Calculate the entropy value of each influencing factor : ,in, , , It is The sample in The value of the influencing factors, is the coefficient of variation, It is Samples for each influencing factor The ratio of
[0021] Combined with the correlation adjustment, the coefficient of difference, ;
[0022] Use the adjusted coefficient of variation to calculate the weight of each indicator: ;
[0023] After calculating the weights, the corresponding first learning progress tracking model is obtained: , To predict learning progress, is the intercept term.
[0024] In a preferred embodiment of the present invention, in step S2, the behavior data includes the degree of hearing difficulty, and the degree of hearing difficulty is determined based on the acquired video volume, and the volume is used to numerically describe the degree of hearing difficulty.
[0025] In a preferred embodiment of the present invention, The video volume level set for the user. The default video volume level of the platform. is the total volume range, Set the volume difference between the user and the default volume;
[0026] Convert the volume difference to a level relative to the total scale: ;
[0027] When the user adjusts the volume during viewing, Introduce an adjustment factor for the number of times the user adjusts the volume within a fixed time , and finally the total hearing loss score is calculated as: ;
[0028] Adjusting the intercept term by hearing impairment ,in is a constant coefficient: ;
[0029] Get the second learning progress tracking model: .
[0030] In a preferred embodiment of the present invention, in step S3,
[0031] S31, converting the behavior data that can reflect the learning content preference, learning resource preference and learning method preference into numerical form;
[0032] S32, respectively calculating the preference of learning content preference, learning resource preference and learning method preference;
[0033] Learning content preference: For each learning content including sign language vocabulary, sentences or dialogues, calculate the user's viewing frequency and duration; ;
[0034] in, It means Content preference of each learning content, is the total amount of learning content, For users to watch The frequency of learning content, For users to watch The duration of the learning content; The product of the total frequency and duration of all content viewed by the user;
[0035] Learning resource preference: For each learning resource, including video tutorials, interactive exercises, and graphic teaching materials, record the user's access and usage frequency: ;in, It means The resource preference of the learning resource, is the total number of learning resources, User access and use frequency of resource generation; The total frequency of users accessing and using all resources;
[0036] Learning method preference: data on users’ practice completion and video viewing habits: ;in, Indicates Preference for learning method, is the total number of learning methods, Indicates that the user adopts The frequency of the method, Indicates the total frequency of all methods adopted by users;
[0037] S33, comprehensively calculating the learning preference and obtaining a dynamic evaluation model of the learning preference;
[0038] The content preference, resource preference and method preference are weighted and integrated to obtain the overall learning preference. : ; Calculated by entropy weight method The weights representing content preference, resource preference, and method preference are the same as those in S2.
[0039] In a preferred embodiment of the present invention, the current learning progress of the hard of hearing person is obtained from the second learning progress tracking model. , obtain the learning preference from the learning preference dynamic evaluation model , , ; represents the weights of content preference, resource preference, and method preference,
[0040] Build a personalized learning path model: ;
[0041] in: is the personalized learning path score; It is a content personalization scoring function based on learning progress and preference; It is a resource personalization scoring function based on learning progress and preference; It is a method-based scoring function for personalization based on learning progress and preference;
[0042] The personalization score function is defined as a weighted combination of learning progress and preference: ; ; ;in: is a constant between 0 and 1.
[0043] In a preferred embodiment of the present invention, according to Scoring, categorizing learning content, resources and methods The scores are arranged from high to low and are recommended for people with hearing loss.
[0044] The present invention provides a sign language learning progress tracking and personalized recommendation management system, comprising:
[0045] User information management module, which collects and stores basic information of users and manages them according to their categories: deaf, hard of hearing, and normal people;
[0046] The learning behavior tracking module monitors and records the user's behavior data on the sign language learning platform in real time, and builds a unique learning progress tracking model for each user category;
[0047] The learning preference evaluation module analyzes user behavior data, identifies and evaluates user preferences for different sign language learning resources and methods, and updates the learning preference dynamic evaluation model in real time;
[0048] The personalized recommendation engine module generates personalized learning resource and learning path recommendations based on the user's learning preference evaluation results and current learning progress;
[0049] The learning path optimization module optimizes and adjusts the personalized learning path for users based on the output of the personalized recommendation engine module to ensure the consistency and suitability of the learning content;
[0050] The user experience optimization module collects user feedback information, optimizes system iteration and recommendation algorithms; continuously improves user interface design, interactive functions and auxiliary tools, provides barrier-free access support, and enhances user learning experience.
[0051] The present invention solves the defects existing in the background technology and has the following beneficial effects:
[0052] (1) The present invention performs personalized learning path planning based on the learning preferences and learning progress of the hard of hearing person, selects the learning content, resources and methods with the highest scores, and constructs a personalized learning path, so that the learning path is more in line with the personalized needs and learning progress of the hard of hearing person, thereby improving the efficiency and effect of learning. The existing technology often only provides a general learning path or recommendation, while the present invention realizes personalized learning path planning, promotes the independent learning and self-improvement of the hard of hearing person, and improves the success rate and satisfaction of sign language learning.
[0053] (2) The present invention utilizes the combination of basic information of the hard of hearing and behavioral data. The present invention collects the basic information of the hard of hearing during registration and records the behavioral data of the hard of hearing in real time during the learning process. The weights of the influencing factors are calculated through correlation analysis and entropy weight method to generate a first learning progress tracking model. The model is dynamically adjusted based on the behavioral data, especially the degree of hearing loss, to obtain a second learning progress tracking model. This makes the learning progress tracking more personalized and accurate, and can better adapt to the learning needs and changes of the hard of hearing, thereby improving the effect and satisfaction of sign language learning and enabling the hard of hearing to learn more efficiently.
[0054] (3) The present invention constructs a dynamic learning preference evaluation model by analyzing the learning content preferences, learning resource preferences, and learning method preferences of the hard of hearing. The model updates and evaluates the learning preferences in real time, making personalized recommendations more in line with the actual needs and preferences of the hard of hearing, thereby improving the pertinence and effectiveness of learning, enhancing the learning motivation and participation of the hard of hearing, and improving the overall effect of sign language learning.
[0055] (4) When there is incoherence in the output learning path planning content, the present invention optimizes it through a path smoothing algorithm to ensure the coherence and suitability of the learning content, making the learning path smoother and more coherent, and improving the logic and systematicness of learning. The existing technology often only focuses on the recommendation of learning content, but ignores the coherence and logic of the learning path. The present invention realizes the optimization and adjustment of the learning path, enhances the learning experience and effect of the hard of hearing, and makes sign language learning smoother and more efficient.
[0056] (5) The present invention provides a sign language learning progress tracking and personalized recommendation management system, including multiple modules such as user information management, learning behavior tracking, learning preference assessment, personalized recommendation engine, learning path optimization and user experience optimization, so as to make sign language learning management more comprehensive and systematic, improve the efficiency and effectiveness of management, enhance the overall level and effect of sign language learning, and provide more high-quality and personalized learning services for the hard of hearing. BRIEF DESCRIPTION OF THE DRAWINGS
[0057] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art are briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work.
[0058] Figure 1 is a flow chart of a preferred embodiment of the present invention; DETAILED DESCRIPTION
[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0060] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below. Embodiment 1
[0061] A method for tracking sign language learning progress and providing personalized recommendations, comprising the steps of:
[0062] S1. Collect basic information of the hard of hearing sign language learner and generate a first learning progress tracking model;
[0063] S2, real-time recording of the behavior data of the hard of hearing person in the sign language learning process, determining the degree of the hard of hearing person, modifying the first learning progress tracking model according to the degree of the hard of hearing, and obtaining the second learning progress tracking model to obtain the learning progress of the hard of hearing person;
[0064] S3, identify the learning preferences of the hard of hearing students based on the behavioral data in S2, build a dynamic evaluation model for learning preferences, and evaluate the preferences of the hard of hearing students for different sign language learning content, resources, and methods;
[0065] S4, according to the preference output in S3 and the learning progress of the corresponding hard of hearing person, a personalized learning path is planned for the hard of hearing person;
[0066] S5. Push the personalized learning path planning to the corresponding hard of hearing person.
[0067] Hard of hearing refers to individuals who retain a certain degree of residual hearing. Hard of hearing people can hear some sounds, but cannot understand speech clearly or need louder sounds to hear clearly. The degree of hearing loss of hard of hearing people varies. Some are only mild and have little impact on daily communication, while others are more serious and need hearing aids or other auxiliary devices to improve hearing.
[0068] The communication challenges faced by hard of hearing people in sign language learning are different from those faced by people with normal hearing and people who are completely deaf. Due to differences in the degree of hearing loss, each hard of hearing person has a different learning style, progress, speed, and interests when learning sign language. Therefore, it is particularly important to design a system that can provide personalized learning recommendations and progress tracking for the hard of hearing.
[0069] In the present invention, S1 collects basic information when the user registers to understand the user's starting point and needs, ensuring that the learning system can provide a customized learning experience based on the user's individual differences; uses correlation analysis to determine the impact of factors on learning progress, assigns weights to each factor, and constructs the first learning progress tracking model, laying the foundation for subsequent learning progress tracking and personalized recommendations.
[0070] Specifically, basic information of the hard of hearing is collected during registration, including age, level of hard of hearing, use of assistive devices, background in learning sign language, and learning goals; the basic information is used as an influencing factor in constructing the first learning progress tracking model.
[0071] The levels of hearing loss include: "severe hearing loss", "moderate hearing loss" and "mild hearing loss"; severe hearing loss means that daily conversations can hardly be heard clearly, and one needs to speak loudly or use hearing aids to barely understand; moderate hearing loss means that one can hear some sounds in a quiet environment, but has difficulty understanding in noisy environments or long-distance conversations; mild hearing loss means that one can hear most sounds, but needs to concentrate or ask for repetition when there is loud background noise and the speaker's voice is soft.
[0072] The use of assistive devices includes: whether hearing assistive devices are used and the frequency of using hearing assistive devices; the background of learning sign language includes: whether there is experience in learning sign language and the learning time; the learning goals include: personal interests, professional needs, and social communication;
[0073] The first learning progress tracking model is a learning progress tracking framework initially established for each newly registered hard of hearing user in the sign language learning system. The first learning progress tracking model sets an initial learning path and progress evaluation criteria based on the user's basic information.
[0074] Specifically, the step of generating the first learning progress tracking model in S1 includes:
[0075] S11. Conduct correlation analysis using basic information of hard of hearing persons;
[0076] S12. Assign weights to basic information that affects learning progress;
[0077] S13. Output a first learning progress tracking model.
[0078] Correlation analysis is a statistical method used to evaluate whether there is a statistical association between two or more variables, as well as the strength and direction of this association. There are significant individual differences in the learning needs and abilities of hard of hearing people. In S11, through correlation analysis, the system identifies which factors have a significant impact on learning progress, thereby better understanding the uniqueness of each learner.
[0079] Specifically, the hearing impairment level was quantified into a numerical value: severe hearing impairment = 3, moderate hearing impairment = 2, mild hearing impairment = 1; the use of assistive devices was quantified into a binary variable: use = 1, no use = 0; the frequency of use was: low = 1, medium = 2, high = 3; the background of learning sign language was quantified: learning experience = 1, no learning experience = 0; the learning time was expressed in actual hours;
[0080] For each influencing factor and learning progress Calculate Pearson correlation coefficient : ;in, It is The sample in The value of the influencing factors, and is the average value, For the The learning progress of samples, is the sample size of hard of hearing subjects.
[0081] The weight is calculated based on the absolute value of the correlation coefficient to reflect the contribution of each factor to the learning progress, and the correlation coefficient is standardized. , to eliminate the influence of different dimensions: , m is the total number of influencing factors;
[0082] Combined with the above correlation coefficient, the entropy weight method is used to calculate the weights of the influencing factors. The entropy weight method is an objective weight allocation method that determines the weight based on the discreteness of the data rather than subjective judgment. The combination of the two ensures that the weight allocation not only considers the discreteness of the data, but also the actual correlation strength between the factors and the learning progress.
[0083] Calculate the entropy value of each influencing factor , the entropy value reflects the discrete degree of the influencing factors, and the calculation formula is as follows: , , , It is The sample in The value of the influencing factors, is the coefficient of variation, It is Samples for each influencing factor ratio.
[0084] Combined with the correlation adjustment, the coefficient of difference, ;
[0085] Use the adjusted coefficient of variation to calculate the weight of each indicator: ;
[0086] After calculating the weights, the corresponding first learning progress tracking model is obtained: , To predict learning progress, is the intercept term.
[0087] Since the first learning progress tracking model uses the hearing loss level that the hard of hearing person subjectively evaluates himself, the hearing loss level is quantified into a numerical value to facilitate the construction of the first learning progress tracking model. However, by monitoring the actual behavior data of the hard of hearing person, more accurate data can be provided to adjust the first learning progress tracking model. The purpose of step S1 is to establish an initial learning model, while the purpose of step S2 is to dynamically adjust the model to better adapt to the actual learning situation of the user.
[0088] In S2 of the present invention, the degree of hearing loss of the hard of hearing person can be obtained through behavioral data. The degree of hearing loss is inferred through the actual behavioral data generated by the user when using the platform. The degree of hearing loss is judged according to the volume of the video. The hard of hearing person who needs a higher volume has a more serious hearing loss. The volume is used to numerically describe the degree of hearing loss.
[0089] In S2, the platform's built-in log system is used to record and collect users' learning behavior data, including: video viewing time, number of pauses, segments watched repeatedly, exercise completion status, and interactive Q&A participation.
[0090] By analyzing the time, progress, and repeated viewing segments of the video, the system evaluates the time required for learners to understand the content and their learning efficiency; if users often repeatedly watch certain parts of the video or take longer to complete the learning task, this indicates that they need more time to understand the content; these behavioral data serve as influencing factors for the continuous optimization of the first learning progress tracking model, making the second learning progress model more in line with the current learning status of the hard of hearing.
[0091] In S2, the behavioral data are converted into numerical form, including: viewing time is converted into total viewing time or average viewing time; the number of pauses is converted into the actual number of times; the repeatedly watched segments are converted into the total time or number of repeated viewings; the exercise completion status is converted into the completion percentage; the interactive question and answer participation is converted into the number of participations or participation rate.
[0092] The behavioral data of the hard of hearing, including the time of watching the video, the number of pauses, the segments watched repeatedly, the completion of exercises, and the participation in interactive questions and answers, are added to S1 as new influencing factors. After calculating the new influencing factors, the corresponding correlation of the second learning progress tracking model is and weight .
[0093] set up The video volume level set for the user. The default video volume level of the platform. is the total volume range, Set the volume for the user and the difference from the default volume.
[0094] Convert the volume difference to a level relative to the total range: ;
[0095] If the user adjusts the volume while watching, Introduce an adjustment factor for the number of times the user adjusts the volume within a certain period of time The logarithmic function is used to reduce the impact of extreme adjustments on the score, while ensuring that even a single adjustment can be taken into account. Finally, the total hearing loss score is calculated as: ;
[0096] The hearing difficulty score takes into account the volume difference and adjustment frequency, which can more accurately reflect the user's hearing difficulty. ,in is a constant coefficient: ;
[0097] The second learning progress tracking model can be obtained: ; for The value of the influencing factors after adding behavioral data, Track the model's prediction learning progress for the second progress.
[0098] In step S2, the first learning progress tracking model is optimized and updated once a week. The learning progress and comprehension ability of each hard of hearing student change dynamically. Regularly updating the model can ensure that it always reflects the learner's current learning status and needs.
[0099] In S3 of the present invention, learning preference is crucial to personalized learning and can significantly affect learning outcomes. It analyzes users' learning preferences for different learning contents, resources, and methods, and uses a dynamic evaluation model to continuously track and update user preference changes to ensure the relevance of recommended content.
[0100] In step S3, learning preference refers to the preference or tendency of the hard of hearing person towards different learning content, resources and methods during the sign language learning process. Identifying learning preference can help the system provide personalized learning recommendations for each hard of hearing person more accurately, thereby improving learning outcomes and satisfaction.
[0101] Learning content preference: By analyzing the frequency and duration of how hard of hearing people view specific sign language words, sentences, or dialogues, the system can identify learning content that the user is more interested in or finds more useful.
[0102] Learning resource preference refers to the fact that hard of hearing people prefer to use certain types of learning resources, including video tutorials, interactive exercises, and graphic teaching materials. The system can identify this preference by recording the user's access and usage frequency of different types of resources.
[0103] Learning method preference: The method preference of hard of hearing people affects their learning experience. Specifically, some hard of hearing people prefer to consolidate their knowledge through repeated practice, while some hard of hearing people prefer to learn by watching demonstration videos. The system can identify this preference by analyzing the user's practice completion status, video viewing habits and other data.
[0104] S31, converting the behavior data that can reflect the learning content preference, learning resource preference and learning method preference into numerical form;
[0105] S32, respectively calculating the preference of learning content preference, learning resource preference and learning method preference;
[0106] Learning content preference: For each learning content including sign language vocabulary, sentences or dialogues, calculate the user's viewing frequency and duration; ;
[0107] in, It means Content preference of each learning content, is the total amount of learning content, For users to watch The frequency of learning content, For users to watch The duration of the learning content; It is the product of the total frequency and duration of all content viewed by the user.
[0108] Learning resource preference:
[0109] For each learning resource including: video tutorials, interactive exercises, graphic teaching materials, record the user's access and usage frequency: ;
[0110] in, It means The resource preference of the learning resource, is the total number of learning resources, User access and use frequency of resource generation; The total frequency of users accessing and using all resources.
[0111] Learning method preference:
[0112] Data on users’ exercise completion and video viewing habits: ;
[0113] in, Indicates Preference for learning method, is the total number of learning methods, Indicates that the user adopts The frequency of the method, Indicates the total frequency of all methods adopted by users.
[0114] S33. Comprehensively calculate the learning preference and derive a dynamic evaluation model for the learning preference.
[0115] The content preference, resource preference and method preference are weighted and integrated to obtain the overall learning preference. : ;
[0116] Calculated by entropy weight method It represents the weights of content preference, resource preference and method preference. The calculation steps of entropy weight method are the same as those in S2.
[0117] In S4 of the present invention, the learning progress and preference are comprehensively considered to ensure that the learning path is both in line with the user's current level and can stimulate their interest; a personalized learning path model is constructed based on the user's learning progress and learning preference, to improve learning motivation and efficiency, and to help users learn at their own pace and preference.
[0118] Get the current learning progress of the hard of hearing person from the second learning progress tracking model , obtain the learning preference from the learning preference dynamic evaluation model , , ; Represents the weights of content preference, resource preference, and method preference.
[0119] Build a personalized learning path model: ;
[0120] in: is the personalized learning path score; It is a content personalization scoring function based on learning progress and preference; It is a resource personalization scoring function based on learning progress and preference; It is a method-based scoring function for personalization based on learning progress and preference.
[0121] The personalization score function is defined as a weighted combination of learning progress and preference: ; ; ;
[0122] in: is a constant between 0 and 1, used to balance the effects of learning progress and preference.
[0123] according to Score, select the highest-scoring learning content, resources, and methods, and build a personalized learning path.
[0124] In S5, according to Scoring, categorizing learning content, resources and methods The scores are arranged from high to low and are recommended for people with hearing loss. Embodiment 2
[0125] This embodiment is further limited on the basis of the first embodiment.
[0126] Specifically, in step S4, the output learning path planning content is incoherent, and the incoherence phenomenon is optimized by a path smoothing algorithm.
[0127] For all knowledge points, is the corresponding difficulty coefficient;
[0128] For continuous knowledge points and , calculate the difficulty difference: ;
[0129] Defining a smoothness indicator , used to measure the smoothness of the knowledge point sequence: ;
[0130] Combine the smoothness with the personalized learning path score to get a weighted smoothness score : ;
[0131] According to the weighted smoothness scores, the learning content, resources and methods are arranged from high to low and recommended to the hard of hearing. Embodiment 3
[0132] This embodiment is further limited on the basis of the first or second embodiment.
[0133] A sign language learning progress tracking and personalized recommendation management system, comprising:
[0134] User information management module, which collects and stores basic information of users and manages them according to their categories: deaf, hard of hearing, and normal people;
[0135] The learning behavior tracking module monitors and records the user's behavior data on the sign language learning platform in real time, and builds a unique learning progress tracking model for each user category;
[0136] The learning preference evaluation module analyzes user behavior data, identifies and evaluates user preferences for different sign language learning resources and methods, and updates the learning preference dynamic evaluation model in real time;
[0137] The personalized recommendation engine module generates personalized learning resource and learning path recommendations based on the user's learning preference evaluation results and current learning progress;
[0138] The learning path optimization module optimizes and adjusts the personalized learning path for users based on the output of the personalized recommendation engine module to ensure the consistency and suitability of the learning content;
[0139] The user experience optimization module collects user feedback information, optimizes system iteration and recommendation algorithms; continuously improves user interface design, interactive functions and auxiliary tools, provides barrier-free access support, and enhances user learning experience.
[0140] The above is based on the ideal embodiment of the present invention. Through the above description, relevant personnel can make various changes and modifications without departing from the technical concept of the present invention. The technical scope of the present invention is not limited to the content in the specification, and the technical scope must be determined according to the scope of the claims.
Claims
1. A method for tracking sign language learning progress and providing personalized recommendations, characterized in that: Includes steps: S1. Collect basic information of the hard of hearing sign language learner and generate a first learning progress tracking model; S2, real-time recording of the behavior data of the hard of hearing person in the sign language learning process, determining the degree of the hard of hearing person, modifying the first learning progress tracking model according to the degree of the hard of hearing, and obtaining the second learning progress tracking model to obtain the learning progress of the hard of hearing person; S3, identify the learning preferences of the hard of hearing students based on the behavioral data in S2, build a dynamic evaluation model for learning preferences, and evaluate the preferences of the hard of hearing students for different sign language learning content, resources, and methods; S4, according to the preference output in S3 and the learning progress of the corresponding hard of hearing person, a personalized learning path is planned for the hard of hearing person; S5. Push the personalized learning path planning to the corresponding hard of hearing person; In step S1, the step of generating a first learning progress tracking model includes: S11. Conduct correlation analysis using basic information of hard of hearing persons; S12. Assign weights to basic information that affects learning progress; S13, outputting a first learning progress tracking model; The basic information is used as the influencing factors for building the first learning progress tracking model. for Each influencing factor X j Calculate the Pearson correlation coefficient r with the learning progress Y j : in, X ij It is i The sample in j The value of the influencing factors, Y i For the i The learning progress of samples, For all samples in j The average value of the influencing factors is is the average learning progress of all samples, n is the sample size of hard of hearing subjects; Standardize the correlation coefficient S j : m is the total number of influencing factors; Calculate the entropy value of each influencing factor e j : in, g j =1- e j , f ij It is i The sample in j The value of the influencing factors, g j is the coefficient of variation, P ij It is i Samples for each influencing factor X j The ratio of Combined with the correlation adjustment, the coefficient of difference, g j ′ = g j ·| r j |; Use the adjusted coefficient of variation to calculate the weight of each indicator: After calculating the weights, the corresponding first learning progress tracking model is obtained: P To predict learning progress, β 0 is the intercept term; In step S2, V user The video volume level set for the user. V default The default video volume level of the platform. V scale is the total volume range, ΔV Set the volume difference between the user and the default volume; Convert the volume difference to a level relative to the total scale: When the user adjusts the volume during viewing, N adjustments Introduce an adjustment factor for the number of times the user adjusts the volume within a fixed time f adjust =log( N adjustments +1), and finally the total hearing loss score is calculated as: S = V relativw ·(1+ f adjust ); Adjusting the intercept term by hearing impairment β′ ,in k is a constant coefficient: β′ = β 0 + k·S ; Get the second learning progress tracking model: X j ′ for X j The value of the influencing factor after adding behavioral data; the behavioral data of the hard of hearing is added as a new influencing factor to S1 for calculation, and the corresponding weight of the second learning progress tracking model w j ′ ; In step S3, S31, converting the behavior data reflecting the learning content preference, learning resource preference and learning method preference into numerical form; S32, respectively calculating the preference of learning content preference, learning resource preference and learning method preference; Learning content preference: For each learning content including sign language vocabulary, sentences or dialogues, calculate the user's viewing frequency and duration; in, P content It means a Content preference of each learning content, b is the total amount of learning content, f(a) For users to watch a The frequency of learning content, d(a) For users to watch a The duration of the learning content; ∑ b f(b)·d(b) The product of the total frequency and duration of all content viewed by the user; Learning resource preference: For each learning resource, including video tutorials, interactive exercises, and graphic teaching materials, record the user's access and usage frequency: in, P resource It means a′ The resource preference of the learning resource, b′ is the total number of learning resources, u(a′) User access and use a′ frequency of resource generation; ∑ b′ u(b′) The total frequency of users accessing and using all resources; Learning method preference: data on users’ practice completion and video viewing habits: in, P method (a″) Indicates a″ Preference for learning method, b″ is The total number of learning methods, h(a″) Indicates that the user adopts a″ The frequency of the method, ∑ b″ h(b″) Indicates the total frequency of all methods adopted by users; S33, comprehensively calculating the learning preference and obtaining a dynamic evaluation model of the learning preference; The content preference, resource preference and method preference are weighted and integrated to obtain the overall learning preference. P overall : P overall = w 1 · P content + w 2 · P resource + w 3 · P method ; Calculated by entropy weight method w 1 , w 2 , w 3 The weights representing content preference, resource preference and method preference are calculated by the entropy weight method in the same way as in S2. Get the current learning progress of the hard of hearing person from the second learning progress tracking model P′ , obtain the learning preference from the learning preference dynamic evaluation model P content , P resource , P method ; w 1 , w 2 , w 3 represents the weights of content preference, resource preference, and method preference, Build a personalized learning path model: LP = w 1 · C ( P′ , P content )+ w 2 · R ( P′ , P resource )+ w 3 · M ( P′ , P method ); in: LP is the personalized learning path score; C ( P′ , P content ) is a content personalization scoring function based on learning progress and preference; R ( P′ , P resource ) is a resource personalization scoring function based on learning progress and preference; M ( P′ , P method ) is a method-based personalized scoring function based on learning progress and preference; The personalization score function is defined as a weighted combination of learning progress and preference: C ( P′ , P content )= α · P′ +(1- α ) P content ; C ( P′ , P resource )= α · P′ +(1- α ) P resource ; C ( P′ , P method )= α · P′ +(1- α ) P method ;in: α is a constant between 0 and 1.
2. The method for tracking sign language learning progress and providing personalized recommendations according to claim 1, characterized in that: Basic information of the hearing-impaired is collected during registration, including: age, hearing impairment level, use of assistive devices, background in learning sign language, and learning goals; hearing impairment levels include: severe hearing impairment, moderate hearing impairment, and mild hearing impairment.
3. The method for tracking sign language learning progress and providing personalized recommendations according to claim 1, characterized in that: In step S2, the behavior data includes the degree of hearing loss, which is determined based on the volume of the acquired video, and the volume is used to numerically describe the degree of hearing loss.
4. The method for tracking sign language learning progress and providing personalized recommendations according to claim 1, characterized in that: According to the LP scores, the learning contents, resources and methods are arranged from high to low according to the LP scores and recommended to the hard of hearing people.
5. A sign language learning progress tracking and personalized recommendation management system, based on a sign language learning progress tracking and personalized recommendation method according to any one of claims 1 to 4, characterized in that: include: User information management module, which collects and stores basic information of users and manages them according to their categories: deaf, hard of hearing, and normal people; The learning behavior tracking module monitors and records the user's behavior data on the sign language learning platform in real time, and builds a unique learning progress tracking model for each user category; The learning preference evaluation module analyzes user behavior data, identifies and evaluates user preferences for different sign language learning resources and methods, and updates the learning preference dynamic evaluation model in real time; The personalized recommendation engine module generates personalized learning resource and learning path recommendations based on the user's learning preference evaluation results and current learning progress; The learning path optimization module optimizes and adjusts the personalized learning path for users based on the output of the personalized recommendation engine module to ensure the consistency and suitability of the learning content; User experience optimization module, which collects user feedback and optimizes system iteration and recommendation algorithms; Continuously improve user interface design, interactive functions and assistive tools, provide barrier-free access support, and enhance user learning experience.
