An education course resource intelligent recommendation and teaching effect evaluation method and system

Abstract

This invention relates to the field of educational management technology, specifically to a method and system for intelligent recommendation of educational course resources and evaluation of teaching effectiveness. The method includes: acquiring initial selection data of students regarding educational course resources, wherein the initial selection data includes the identifier of the course resource selected by the student and selection time information; and filtering similar selection trajectory groups from the historical learning record database of the educational platform based on the initial selection data, wherein the historical learning record database includes multiple learning trajectory groups, and each learning trajectory group includes the identifier of historically selected course resources with a similar selection order and the corresponding learning performance. This invention filters course resources similar to students' learning trajectories based on their initial selection data and performs personalized optimization sorting using student group characteristics and learning goal tendencies. This ensures that course recommendations better meet the learning needs of each student, thus improving the accuracy of recommendations and avoiding overly generic recommendation schemes.

Description

Technical Field

[0001] This invention relates to the field of educational management technology, specifically to a method and system for intelligent recommendation of educational curriculum resources and evaluation of teaching effectiveness. Background Technology

[0002] Currently, traditional methods typically rely on a uniform recommendation model or broad course-based recommendations without fully considering students' individual needs, learning trajectories, and learning preferences. These methods often recommend courses suitable for all students, ignoring their different learning goals and progress. Moreover, they usually fix a certain recommendation strategy and cannot adjust in real time according to changes in students' learning progress, learning patterns, or external environment. This can lead to recommended course resources that do not match students' current learning status, affecting learning outcomes.

[0003] Furthermore, traditional methods often lack teaching effectiveness prediction models, making it impossible to assess the potential teaching effect of each course. Consequently, they cannot effectively predict the contribution of courses to students' knowledge acquisition and skill improvement. Therefore, students may waste time on courses with insignificant effects. Moreover, recommendations are usually based on certain fixed parameters, making it difficult to refine the ranking according to students' specific needs. This results in course recommendations that may be too general and fail to accurately match students' needs, thus affecting learning outcomes. Summary of the Invention

[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for intelligent recommendation of educational course resources and evaluation of teaching effectiveness, comprising:

[0005] Obtain students' initial selection data for educational course resources, wherein the initial selection data includes the course resource identifier selected by the student and the selection time information;

[0006] Based on the initial selection data, similar selection trajectory groups are selected from the historical learning record database of the education platform. The historical learning record database includes multiple learning trajectory groups, and each learning trajectory group includes the identifier of the historical selected course resources with similar selection order and the corresponding learning performance.

[0007] The historical selected course resources in the similar selection trajectory group are input into a pre-trained resource feature extraction model to obtain the core teaching point information of each historical selected course resource. The core teaching point information is used to characterize the content that plays a role in students' knowledge acquisition and skill improvement.

[0008] The core teaching points are input into a pre-trained teaching effectiveness prediction model to obtain a teaching effectiveness prediction value corresponding to each core teaching point. The teaching effectiveness prediction value is used to characterize the potential effect of the core teaching points on student learning.

[0009] Based on the predicted teaching effectiveness, the historically selected course resources in the selected similar selection trajectory groups are optimized and sorted to obtain an optimized course resource recommendation sequence. Courses in the optimized course resource recommendation sequence that are different from the course resources in the students' initial selection data are selected as courses to be recommended, and the detailed teaching information corresponding to the courses to be recommended is used as recommendation details for recommendation.

[0010] Preferably, the training method for the resource feature extraction model includes:

[0011] To obtain historical education curriculum resources and students' learning performance on these historical education curriculum resources in history;

[0012] The history education curriculum resources are input into the resource feature extraction model. The resource feature extraction model is used to determine the core teaching points of history in the history education curriculum resources. At the same time, the key teaching links in the history education curriculum resources are determined based on the students’ academic performance.

[0013] The resource feature extraction model is trained with the optimization objective of minimizing the discrepancy between the core teaching points of history and the teaching content corresponding to the key teaching links in history education curriculum resources.

[0014] Preferably, before inputting the historically selected course resources from the similar selection trajectory group into a pre-trained resource feature extraction model to obtain the core teaching point information of each historically selected course resource, the method further includes:

[0015] Process the historical selected course resources in the filtered similar selection trajectory groups to obtain a set of course segments of a specified duration;

[0016] Inputting the historically selected course resources from the similar selection trajectory group into a pre-trained resource feature extraction model yields the core teaching points information for each historically selected course resource, including:

[0017] For each teaching segment in each historical selected course resource, the teaching depth index corresponding to each course type in the teaching segment is determined by the resource feature extraction model.

[0018] For each teaching segment in each historical selected course resource, the teaching importance of the teaching segment is determined based on the teaching depth index corresponding to each course type at the teaching segment and the preset teaching weight corresponding to each course type.

[0019] If the teaching importance corresponding to the teaching segment is greater than the preset importance threshold, then the core teaching points are determined based on the teaching content corresponding to the teaching segment.

[0020] Preferably, before inputting the historical selected course resources from the similar selection trajectory group into the pre-trained resource feature extraction model, the method further includes:

[0021] Determine the student group characteristics corresponding to the historically selected course resources in the filtered similar selection trajectory groups;

[0022] Input the historical selected course resources from the similar selection trajectory group into the pre-trained resource feature extraction model, including:

[0023] The core teaching points and student group characteristics are input into the teaching effectiveness prediction model, and the teaching effectiveness prediction model determines the teaching effectiveness prediction value for each core teaching point for the student group.

[0024] Preferably, the historically selected course resources in the selected similar selection trajectory groups are optimized and sorted according to the predicted teaching effectiveness values, including:

[0025] For each core teaching point information, if the predicted teaching effectiveness value corresponding to the core teaching point information is lower than the preset effectiveness threshold, then several core teaching point information located before the teaching segment corresponding to the core teaching point information and several core teaching point information located after the teaching segment corresponding to the core teaching point information are used as the context teaching point information corresponding to the core teaching point information.

[0026] Based on the contextual teaching points, the teaching content at the corresponding teaching stages is optimized and adjusted, and the historical selected course resources in the selected similar selection trajectory groups are reordered based on the optimized and adjusted teaching content.

[0027] Preferably, the training method for the teaching effectiveness prediction model includes:

[0028] To obtain historical education curriculum resources and students' learning performance on historical education curriculum resources;

[0029] By inputting history education curriculum resources into the resource feature extraction model, the core teaching points of history in the history education curriculum resources are extracted.

[0030] Input the core teaching points of history into the teaching effectiveness prediction model to determine the teaching effectiveness prediction value corresponding to each core teaching point of history.

[0031] Based on the order of predicted teaching effectiveness values ​​corresponding to each core teaching point in history from high to low, a specified number of core teaching points are selected as the first core point group. At the same time, based on the order of the actual improvement in students' learning effectiveness in history from high to low, a specified number of core teaching points are selected as the second core point group.

[0032] The teaching effectiveness prediction model is trained with the optimization objective of minimizing the difference between the first core point group and the second core point group.

[0033] Preferably, after obtaining students' initial selection data regarding educational course resources, the method further includes:

[0034] Determine students' learning goal preferences based on the initial selection data;

[0035] Based on the learning objectives, identify the supplementary course resources that need to be learned.

[0036] The course resources to be supplemented are compared with the course resources in the initial selection data to obtain potential course resources to be recommended;

[0037] If the potential course resources to be recommended include complex course resources with multiple teaching characteristics, then the step of filtering similar selection trajectory groups from the historical learning record database of the education platform based on the initial selection data is executed;

[0038] If all potential course resources to be recommended are simple course resources with a single teaching characteristic, then the learning trajectory groups that include key content corresponding to the learning objectives are identified from the historical learning record database, and the course resources in the learning trajectory groups with the highest usage frequency are recommended.

[0039] Preferably, before filtering similar selection trajectory groups from the historical learning record database of the education platform based on the initial selection data, the method further includes:

[0040] Obtain at least one of the following identifying features from the initial selection data: selection operation duration, selection operation frequency, time interval between two adjacent selection operations, and total selection time of course resources in the initial selection data;

[0041] Student identification is determined based on identification features and initial selection data;

[0042] Based on the learning history of the education platform, the learning status of the education platform on the current day, and the student's identity, different course recommendation strategies are adopted to recommend courses.

[0043] Preferably, different course recommendation strategies are adopted based on the historical learning data of the education platform, the learning data of the current day, and the student's identity identifier, including:

[0044] If historical learning data indicates that the number of times a course was studied on the education platform within a preset time period was a single study, then the course recommendation strategy adopted is: to select similar selection trajectory groups from the historical learning record database of the education platform based on the initial selection data;

[0045] If historical learning data shows that the course was studied multiple times within a preset time period, and the student's learning data and student identity indicate that the same student studied the course multiple times, then the course recommendation strategy adopted is: to recommend a learning trajectory group consisting of the course resources corresponding to the student's identity and the current time from the historical learning record database of the education platform.

[0046] If historical learning data shows that courses on the education platform have been studied multiple times within a preset time period, and the learning data and student identification on the day indicate that multiple students have studied the courses multiple times, then the course recommendation strategy adopted is: to select recommended trajectory groups from the historical learning record database of the education platform based on the initial selection data and course difficulty.

[0047] An intelligent recommendation system for educational curriculum resources and a teaching effectiveness evaluation system, applicable to the aforementioned intelligent recommendation system for educational curriculum resources and a teaching effectiveness evaluation system, comprising:

[0048] The data acquisition unit is used to acquire the initial selection data of students for educational course resources, wherein the initial selection data includes the course resource identifier selected by the student and the selection time information;

[0049] A similar grouping unit is used to filter similar selection trajectory groups from the historical learning record library of the education platform based on the initial selection data. The historical learning record library includes multiple learning trajectory groups, and each learning trajectory group includes the identifier of the historical selected course resource with a similar selection order and the corresponding learning score.

[0050] The resource extraction unit is used to input the historical selected course resources in the similar selection trajectory group into the pre-trained resource feature extraction model to obtain the core teaching point information in each historical selected course resource. The core teaching point information is used to characterize the content that plays a role in students' knowledge acquisition and skill improvement.

[0051] The effectiveness prediction unit is used to input the core teaching point information into a pre-trained teaching effectiveness prediction model to obtain the teaching effectiveness prediction value corresponding to each core teaching point information. The teaching effectiveness prediction value is used to characterize the potential effect of the core teaching point information on student learning.

[0052] The resource recommendation unit is used to optimize and sort the historically selected course resources in the selected similar selection trajectory group based on the predicted teaching effectiveness value, so as to obtain an optimized course resource recommendation sequence. Courses in the optimized course resource recommendation sequence that are different from the course resources in the students' initial selection data are selected as courses to be recommended, and the detailed teaching information corresponding to the courses to be recommended is used as recommendation details for recommendation.

[0053] Compared with the prior art, the beneficial effects of the present invention are:

[0054] (1) This invention filters out course resources similar to students’ learning trajectories through students’ initial selection data, and optimizes the ranking by using student group characteristics and learning goal tendencies. This ensures that the course recommendations are more in line with each student’s learning needs. This not only improves the accuracy of the recommendations, but also avoids overly general recommendation schemes. Moreover, by introducing a teaching effectiveness prediction model, the teaching effectiveness of each core teaching point is predicted, which can quantify the potential role of course resources in students’ knowledge acquisition and skill improvement, thereby optimizing the recommendation process and helping students obtain more efficient learning resources. This evaluation mechanism improves the relevance and practicality of course content.

[0055] (2) This invention reorders historical course resources by predicting teaching effectiveness, which can ensure that the recommended course resources match the students' learning goals and needs. In particular, the optimization and adjustment of teaching links improves the depth and relevance of teaching content. This approach helps to avoid students repeatedly learning course resources that are not suitable for their current level, thereby improving learning efficiency. Moreover, through the analysis of learning trajectory and student identity, it can intelligently adapt to the learning patterns and preferences of different students and adopt different course recommendation strategies for different historical learning situations, student status and other factors, so that the recommendation system is more flexible and efficient and can dynamically adjust the recommendation strategy according to the students' learning progress and learning environment. Attached Figure Description

[0056] Figure 1 This is a schematic flowchart of the overall method in one embodiment of the present invention;

[0057] Figure 2 This is a schematic diagram of the overall system architecture in one embodiment of the present invention.

[0058] In the diagram: 1. Data acquisition unit; 2. Similar grouping unit; 3. Resource extraction unit; 4. Effectiveness prediction unit; 5. Resource recommendation unit. Detailed Implementation

[0059] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0060] Example 1, please refer to Figure 1 This invention provides a technical solution: a method for intelligent recommendation of educational course resources and evaluation of teaching effectiveness, comprising:

[0061] S1. Obtain the student's initial selection data for educational course resources, wherein the initial selection data includes the course resource identifier selected by the student and the selection time information;

[0062] S2. Based on the initial selection data, filter out similar selection trajectory groups from the historical learning record database of the education platform. The historical learning record database includes multiple learning trajectory groups, and each learning trajectory group includes the identifier of the historical selected course resources with similar selection order and the corresponding learning scores.

[0063] S3. Input the historical selected course resources in the similar selection trajectory group into the pre-trained resource feature extraction model to obtain the core teaching point information in each historical selected course resource. The core teaching point information is used to represent the content that plays a role in students' knowledge acquisition and skill improvement.

[0064] S4. Input the core teaching points information into the pre-trained teaching effectiveness prediction model to obtain the teaching effectiveness prediction value corresponding to each core teaching point information. The teaching effectiveness prediction value is used to characterize the potential role of the core teaching points information in student learning.

[0065] S5. Based on the predicted teaching effectiveness, optimize and sort the historically selected course resources in the selected similar selection trajectory groups to obtain an optimized course resource recommendation sequence. Courses in the optimized course resource recommendation sequence that are different from the course resources in the students' initial selection data are selected as courses to be recommended, and the detailed teaching information corresponding to the courses to be recommended is used as recommendation details for recommendation.

[0066] It should be noted that the initial selection data of students for educational course resources is collected; this data usually includes the identifier of the course resource selected by the student (i.e., course name or ID) and the selection time information; for example: suppose student Xiaoming selects the following three courses on the education platform: course A (basic mathematics course), course B (linear algebra course).

[0067] Based on the students' initial selection data (course identifier and selection time information), the system will filter out other students whose selection trajectories are similar to Xiaoming's from the historical learning record database. The historical learning record database contains multiple learning trajectory groups, each trajectory group including the student's course selection order and corresponding academic performance. For example, the following students' learning trajectories were selected from the historical records: Student A: selected course A and course B, with a grade of A; Student B: selected course A, course B, and course C, with a grade of B; Student C: selected course A, course B, and course D, with a grade of A. These students' learning trajectories are similar to Xiaoming's initial selections (course A and course B), and therefore they were selected as the similar selection trajectory group.

[0068] Historical course resources from similar selection trajectory groups are input into a pre-trained resource feature extraction model to extract core teaching points from each historical course resource. These core teaching points characterize the role of each course in students' knowledge acquisition and skill enhancement. For example, for course A (basic mathematics course), the extracted core teaching points may include: understanding basic mathematical concepts and mastering mathematical problem-solving techniques; for course B (linear algebra course), the extracted core teaching points may include: matrix operations, calculation and application of eigenvalues ​​and eigenvectors. These core teaching points will help evaluate the actual learning effectiveness of these courses for students.

[0069] Core teaching points extracted from historical curriculum resources are input into a pre-trained teaching effectiveness prediction model. This model evaluates the predicted teaching effectiveness value for each core teaching point; the predicted teaching effectiveness value represents the degree of potential impact of the core teaching point on student learning. For example, suppose the prediction model evaluates the following teaching effectiveness predicted values: for "Understanding basic mathematical concepts" in Course A, the predicted teaching effectiveness value is 0.85 (i.e., this point has high potential to improve students' basic mathematical abilities); for "Mastering mathematical problem-solving skills" in Course A, the predicted teaching effectiveness value is 0.75; for "Matrix operations" in Course B, the predicted teaching effectiveness value is 0.80; for "Calculation and application of eigenvalues ​​and eigenvectors" in Course B, the predicted teaching effectiveness value is 0.70. These predicted values ​​demonstrate the potential impact of each core teaching point on student learning outcomes.

[0070] Based on the predicted teaching effectiveness of each core teaching point, the system optimizes the ranking of historical course resources in similar selection trajectory groups and generates a recommended course list based on the optimized ranking. The system recommends courses that differ from the student's initial selection data and provides detailed teaching information for these courses as recommendation details. For example, suppose that after optimization, the recommended course sequence is: Course C (because it shows a high predicted teaching effectiveness value in the historical trajectory group, especially in the learning trajectories of students A and B); Course D (because it has a high degree of overlap with the core teaching points of courses A and B, and also has a high predicted effectiveness value). Therefore, the system will recommend Course C and Course D to Xiaoming and provide detailed teaching information for these courses (such as course content, learning objectives, learning paths, etc.) so that Xiaoming understands the advantages of the recommended courses.

[0071] In an optional embodiment, the training method for the resource feature extraction model includes:

[0072] To obtain historical education curriculum resources and students' learning performance on these historical education curriculum resources in history;

[0073] The history education curriculum resources are input into the resource feature extraction model. The resource feature extraction model is used to determine the core teaching points of history in the history education curriculum resources. At the same time, the key teaching links in the history education curriculum resources are determined based on the students’ academic performance.

[0074] The resource feature extraction model is trained with the optimization objective of minimizing the discrepancy between the core teaching points of history and the teaching content corresponding to the key teaching links in history education curriculum resources.

[0075] It should be noted that it is necessary to collect history education course resources and students' learning performance on these resources; these resources can be course content, learning modules, teaching videos, exercises, etc.; learning performance refers to evaluation indicators related to students' performance in these courses, such as scores, pass rates, etc.

[0076] The history education curriculum resources are input into the resource feature extraction model; the task of the resource feature extraction model is to extract relevant features from the original curriculum resources, especially the core teaching points; these points reflect the content of the curriculum resources that can best help students master knowledge.

[0077] After inputting resources, the model will use technologies such as deep learning and natural language processing (e.g., word embedding, text classification, topic modeling) to extract the core teaching points of the course; for example, for a mathematics course, the core teaching points may include: basic formulas, theorems, key steps, etc.

[0078] Use student performance to identify which teaching elements have a significant impact on student grades; these key elements may be specific learning content, skills, or tasks; for example, in a math course, a student's grades may be related to "understanding algebraic formulas," while in a history course, grades may be related to "memorizing historical events."

[0079] The goal is to optimize the resource feature extraction model by minimizing the discrepancy between the core historical teaching information and the key teaching links. This means that we hope the features learned by the model can match students' academic performance and learning outcomes as closely as possible, thereby reducing the difference between the model's predictions and actual learning performance.

[0080] The specific structure of a resource feature extraction model: The structure of a resource feature extraction model typically consists of the following main parts: The input data usually includes text descriptions of courses, academic performance, student behavior data, etc.; these data are preprocessed (such as word embedding, standardization, etc.) and then fed into the model; if the input data is text data (such as course content), it is usually converted into vector representations through embedding layers (such as Word2Vec, GloVe, BERT, etc.); these vector representations can capture the semantic information implicit in the text;

[0081] Convolutional Neural Networks (CNNs) or Recurrent Neural Networks (RNNs): Used to extract local or global features from input course resources; CNNs can be used to extract key information from course texts, while RNNs or Long Short-Term Memory Networks (LSTMs) are suitable for processing learning data with temporal dependencies (e.g., changes in students' grades at different points in the learning process); Transformer Models: In complex text analysis, Transformer structures (e.g., BERT or GPT) are often used to capture long-term dependencies and understand the semantic hierarchy of course content;

[0082] The key instructional component prediction layer analyzes the relationship between historical grades and course content to predict which instructional components have the greatest impact on student performance. This can be achieved through regression analysis, classification algorithms, or neural networks. The goal of this layer is to correlate key instructional components (such as specific knowledge points) in the course with student performance. During training, a loss function is used to calculate the difference between the model's predicted values ​​and actual performance. The objective is to minimize the deviation between core instructional information and key instructional components, typically measured using mean squared error (MSE) or cross-entropy loss. Optimization algorithms (such as SGD, Adam, etc.) are used to adjust the model's parameters to minimize the loss function. The output layer provides the core instructional information and corresponding instructional component scores for each course resource. These outputs can be used for further recommendations or evaluations.

[0083] In an optional embodiment, before inputting the historically selected course resources from the similar selection trajectory group into a pre-trained resource feature extraction model to obtain the core teaching points information of each historically selected course resource, the method further includes:

[0084] Process the historical selected course resources in the filtered similar selection trajectory groups to obtain a set of course segments of a specified duration;

[0085] Inputting the historically selected course resources from the similar selection trajectory group into a pre-trained resource feature extraction model yields the core teaching points information for each historically selected course resource, including:

[0086] For each teaching segment in each historical selected course resource, the teaching depth index corresponding to each course type in the teaching segment is determined by the resource feature extraction model.

[0087] For each teaching segment in each selected historical course resource, the teaching importance of the teaching segment is determined based on the teaching depth index corresponding to each course type and the preset teaching weight corresponding to each course type.

[0088] If the importance of a teaching segment is greater than the preset importance threshold, then the core teaching points are determined based on the teaching content corresponding to that segment.

[0089] It should be noted that by analyzing the historical student selection trajectories, similar learning paths or selection trajectory groups are screened out. These selection trajectories refer to the courses or modules that students have selected in their past learning process. Based on the students' historical learning data, course resources with similar learning trajectories are selected. A set of course segments of a specified duration is generated: the historically selected course resources are divided into multiple segments, each containing a certain amount of teaching content. The purpose of this step is to ensure that each course resource can be decomposed into multiple independent and easily analyzable units, thereby providing structured data for the feature extraction model.

[0090] The selected course segments are input into a pre-trained resource feature extraction model to obtain the core teaching points of each course segment. These points are the parts of the course that are most helpful for students to master the knowledge, and are usually extracted through NLP techniques (such as text mining and topic modeling).

[0091] For each teaching segment, the model calculates the teaching depth based on the course type (e.g., mathematics, history, linguistics, etc.) and the depth of the teaching content. This depth can be calculated by analyzing the complexity of the course content and the hierarchical structure of knowledge points using natural language processing models (e.g., BERT, LSTM, etc.). Teaching importance: Based on the teaching depth index of each segment and the preset teaching weights for each course type, the model calculates the importance of that segment. These teaching weights are typically preset by education experts and reflect the importance of different teaching segments to student learning.

[0092] When the importance of a certain teaching segment exceeds a preset importance threshold, the model will consider that teaching segment to be critical, and then determine the core teaching points based on the teaching content of that segment.

[0093] In an alternative embodiment, the method further includes, before inputting historically selected course resources from similar selection trajectory groups into a pre-trained resource feature extraction model:

[0094] Determine the student group characteristics corresponding to the historically selected course resources in the filtered similar selection trajectory groups;

[0095] Input the historical selected course resources from the similar selection trajectory group into the pre-trained resource feature extraction model, including:

[0096] The core teaching points and student group characteristics are input into the teaching effectiveness prediction model, and the model determines the predicted teaching effectiveness value for each core teaching point for the student group.

[0097] It's important to note that before inputting historically selected course resources into the resource feature extraction model, it's crucial to first identify and understand the student group characteristics associated with these courses. These characteristics include students' basic information, learning habits, interests, and academic background. These features will help better understand which courses are suitable for specific student groups. For example, student group characteristics may include: age (e.g., high school students, university students), learning style (e.g., visual learners, auditory learners, hands-on learners), subject interests (e.g., mathematics, physics, literature), academic level (e.g., excellent, average, poor), and past course selection records (e.g., types of courses chosen, subject areas studied, etc.).

[0098] Historical course resources refer to the course resources that students have previously taken. These resources may include: textbooks, video tutorials, exercise sets, online courses, etc. Each course resource corresponds to different core teaching points (for example, the core teaching point of a certain mathematics course may be "the fundamental theorem of calculus").

[0099] The resource feature extraction model is a trained model used to extract useful information from historical curriculum resources. This information helps to better assess the applicability and effectiveness of the curriculum content for students. Before inputting the information into the model, the following needs to be input: core teaching points (such as key knowledge points of the curriculum content) and student group characteristics (such as age, learning style, academic level, etc.).

[0100] A teaching effectiveness prediction model is a model that predicts the expected learning outcomes of students when learning a particular course. This model predicts the effectiveness of each teaching point for a specific student group based on input core teaching points and student group characteristics. Core teaching points refer to the important knowledge points in the course (e.g., basic formulas in algebra, laws of mechanics in physics). Student group characteristics, as mentioned earlier, include students' personal information and learning progress. By inputting this information, the teaching effectiveness prediction model can derive the predicted effectiveness value for each teaching point. For example, a certain teaching point may be less effective for students with lower grades but more effective for students with higher grades.

[0101] Specific example: Suppose there is a class composed of students from different backgrounds; the collected student group characteristics information includes: Student A: 16 years old, second year of high school, average grades, prefers visual learning; Student B: 18 years old, third year of high school, excellent grades, likes hands-on practice; Student C: 17 years old, third year of high school, poor grades, likes listening to lectures and taking notes;

[0102] Suppose that in the past semester, students selected the following course resources: Course Resource 1: Mathematics course, the core teaching point of which is "the fundamental theorem of calculus"; Course Resource 2: Physics course, the core teaching point of which is "the principles of mechanics"; Course Resource 3: Chemistry course, the core teaching point of which is "chemical reaction equations".

[0103] The selected course resources are input into the resource feature extraction model. The model analyzes the characteristics of the course content and extracts resource features based on the characteristics of the student group. For example, Course 1 (Fundamental Theorem of Calculus) may not be very effective for students with average or poor grades because these students may lack sufficient mathematical foundation; Course 2 (Principles of Mechanics) may be suitable for high-achieving students B who like hands-on learning, but less effective for students C whose learning style is more theoretical; Course 3 (Chemical Reaction Equations) may be suitable for all students, especially those C who like listening to explanations.

[0104] Core teaching points (such as "the fundamental theorem of calculus") and student group characteristics (such as "student A, average grades, prefers visual learning") are input into a teaching effectiveness prediction model. This model analyzes these inputs to predict the teaching effectiveness of each teaching point for a specific student group. The model is assumed to conclude that: for student A, the predicted effectiveness for "the fundamental theorem of calculus" is low because he is weak in mathematical foundations; for student B, the predicted effectiveness for "principles of mechanics" is high because he is good at hands-on activities and enjoys practical applications; and for student C, the predicted effectiveness for "chemical reaction equations" is high because he enjoys listening to explanations and taking notes.

[0105] Based on the predicted teaching effectiveness, the system can adjust the order of course resources to ensure that each student group can choose the course content that best suits them; for example, for student A, more basic mathematics courses or visual teaching content may be recommended; for student B, more experimental physics courses may be recommended; and for student C, lecture-based chemistry courses may be recommended.

[0106] In an optional embodiment, optimizing the ranking of historically selected course resources in the selected similar selection trajectory groups based on predicted teaching effectiveness includes:

[0107] For each core teaching point information, if the predicted teaching effectiveness value corresponding to the core teaching point information is lower than the preset effectiveness threshold, then several core teaching point information that are before the teaching link corresponding to the core teaching point information and several core teaching point information that are after the teaching link corresponding to the core teaching point information will be used as the context teaching point information corresponding to the core teaching point information.

[0108] Based on the contextual teaching points, the teaching content at the corresponding teaching stages is optimized and adjusted. Based on the optimized and adjusted teaching content, the historical selected course resources in the selected similar selection trajectory groups are reordered.

[0109] It should be noted that the teaching effectiveness prediction value is a score of the effectiveness of a teaching activity, course content or teaching segment based on historical data and model prediction; this score is used to evaluate the impact of a course or teaching segment on students' learning outcomes; core teaching points are very important knowledge points, skills points or teaching segments in the course; these are key parts of the course content and are crucial to students' learning objectives.

[0110] Suppose that the predicted effectiveness of a core teaching point in a course is lower than a preset threshold (e.g., the predicted effectiveness is lower than 70), it indicates that the teaching segment is not effective. Then, in order to improve the effectiveness of this part, the teacher needs to find contextual teaching points related to the core teaching point, that is: the teaching content before the core teaching point; the teaching content after the core teaching point; these contents constitute contextual teaching points, which are used to help students better understand the core teaching point.

[0111] Based on the contextual teaching points, the teaching content needs to be optimized and adjusted. For example, it may be necessary to supplement or adjust the way some knowledge points are explained; it may be necessary to add more exercises, case studies or interactive sessions; after the course content is adjusted, similar historical selection trajectories (i.e., similar course combinations previously selected by students) need to be reordered; specifically, the effectiveness of the course resources previously selected by students should be re-evaluated based on the optimized teaching content to ensure that the optimized content can better serve the students' learning needs.

[0112] Specific example: Suppose there is a course on mathematical analysis with several core teaching points, such as "the definition of limits", "the concept of derivatives", and "the application of integrals". In the evaluation process, suppose the predicted effectiveness value of the core teaching point "the concept of derivatives" is less than 70, indicating that it has little impact on students' learning outcomes. This value is below the threshold of 70, so it needs to be optimized.

[0113] The key teaching point of "the concept of derivative" may be located in the 5th lesson of the teaching process. In order to help students better understand derivatives, it is necessary to review the content before and after this teaching process: the content before may be "the definition of limit"; the content after may be "the rule of differential calculus" and "the fundamental theorem of integrals". These contents constitute the contextual teaching points of "the concept of derivative".

[0114] Based on these contextual teaching points, teachers may find that students have not fully grasped the "definition of limits," leading to difficulties in understanding the "concept of derivatives." Therefore, teachers may decide to add a review session on limits before explaining the "concept of derivatives," and use more interaction and concrete examples to help students understand the concept of limits. In addition, teachers may decide to use graphical methods to demonstrate the geometric meaning of derivatives, thereby helping students understand derivatives more intuitively.

[0115] After optimizing the teaching content, it is also necessary to sort similar course resources that students have historically chosen. For example, some students chose course resources containing "limits" and "derivatives". Now, after optimization, students' understanding of derivatives has improved, so teaching content that focuses more on the combination of limits and derivatives will be ranked higher among similar course resources. For example, some historical course resources contain more examples and interactive elements, and these courses may be more attractive in the optimized ranking because they are more in line with the adjusted teaching content.

[0116] In an optional embodiment, the training method for the teaching effectiveness prediction model includes:

[0117] To obtain historical education curriculum resources and students' learning performance on historical education curriculum resources;

[0118] By inputting history education curriculum resources into the resource feature extraction model, the core teaching points of history in the history education curriculum resources are extracted.

[0119] Input the core teaching points of history into the teaching effectiveness prediction model to determine the teaching effectiveness prediction value corresponding to each core teaching point of history.

[0120] Based on the order of predicted teaching effectiveness values ​​corresponding to each core teaching point in history from high to low, a specified number of core teaching points are selected as the first core point group. At the same time, based on the order of the actual improvement in students' learning effectiveness in history from high to low, a specified number of core teaching points are selected as the second core point group.

[0121] The teaching effectiveness prediction model is trained with the optimization objective of minimizing the difference between the first core point group and the second core point group.

[0122] It should be noted that some historical data needs to be obtained: Historical education curriculum resources: This refers to the course content that students have encountered in the past, including textbooks, videos, exercises, etc.; Student academic performance: This refers to the student's actual learning performance in these courses; it may be exam scores, homework scores, or other forms of assessment. For example, consider the following data: Course Resource 1: Mathematics course, the core teaching point is "The Fundamental Theorem of Calculus"; Course Resource 2: Physics course, the core teaching point is "Principles of Mechanics"; Course Resource 3: Chemistry course, the core teaching point is "Chemical Reaction Equations"; Student A and Student B received different grades in these courses: Student A: Mathematics 80, Physics 90, Chemistry 70; Student B: Mathematics 60, Physics 85, Chemistry 95.

[0123] The role of the resource feature extraction model is to extract core teaching points from historical course resources. These points may include important concepts, theorems, principles, etc. in the course. For example, the core teaching points of a mathematics course may be "the fundamental theorem of calculus", the core teaching points of a physics course may be "the principles of mechanics", and the core teaching points of a chemistry course may be "chemical reaction equations". These core teaching points will be extracted and become the input features of the model.

[0124] The core teaching points are input into the teaching effectiveness prediction model. The model will predict the teaching effectiveness of each teaching point based on historical data, that is, the degree of influence of the teaching point on different students. The predicted value will be a number, indicating the effectiveness of the teaching point. For example, suppose the model predicts the following teaching effectiveness values: "The Fundamental Theorem of Calculus" has a predicted teaching effectiveness value of 0.8 for student A and 0.6 for student B; "Principles of Mechanics" has a predicted teaching effectiveness value of 0.7 for student A and 0.9 for student B; "Chemical Reaction Equations" has a predicted teaching effectiveness value of 0.6 for student A and 0.8 for student B.

[0125] Based on the predicted teaching effectiveness, the core teaching points need to be divided into two groups: The first core point group: a specified number of teaching points are selected according to their predicted values ​​from highest to lowest; the second core point group: a specified number of teaching points are selected according to the actual improvement in students' learning effectiveness in history, from highest to lowest. For example, if the first two core points are selected: the first core point group: the teaching point with the highest predicted value is chosen, for example, "The Fundamental Theorem of Calculus" (highest predicted value for student A); the second core point group: the teaching point with the greatest improvement in students' actual learning effectiveness is chosen. For example, if student A's score improves from 60 to 80 and student B's score improves from 70 to 75, student A's improvement is greater, therefore "The Fundamental Theorem of Calculus" is selected in the second core point group.

[0126] The goal of training is to minimize the difference between the first and second core point groups; that is, to adjust the teaching effectiveness prediction model so that the teaching points selected in the first and second groups are as consistent as possible. This can be accomplished using optimization algorithms (such as gradient descent). Specifically, the difference between the first and second core point groups will be calculated, for example, by calculating the difference in the predicted effect of each point in the two groups, with the goal of minimizing this difference. The structure of the teaching effectiveness prediction model in this context may include the following parts: Input layer: receives core teaching point information and student group characteristic information from history education curriculum resources; Feature extraction layer: extracts features of curriculum resources (such as teaching content, difficulty, etc.) through deep learning or other feature extraction algorithms; Teaching effectiveness prediction layer: predicts the teaching effectiveness of each core teaching point based on student characteristics and resource characteristics; this may use algorithms such as regression analysis and neural networks for prediction; Optimization layer: continuously adjusts the model parameters through optimization algorithms (such as gradient descent) to minimize the difference between the first and second core point groups.

[0127] In an optional embodiment, after obtaining students' initial selection data for educational course resources, the method further includes:

[0128] Determine students' learning goal preferences based on the initial selection data;

[0129] Based on the learning objectives, identify the supplementary course resources that need to be learned.

[0130] The course resources to be supplemented are compared with the course resources in the initial selection data to obtain potential course resources to be recommended;

[0131] If the potential course resources to be recommended include complex course resources with multiple teaching characteristics, then the step of filtering similar selection trajectory groups from the historical learning record database of the education platform based on the initial selection data is executed;

[0132] If all potential course resources to be recommended are simple course resources with a single teaching characteristic, then the learning trajectory groups that include key content corresponding to the learning objectives are identified from the historical learning record database, and the course resources in the learning trajectory groups with the highest usage frequency are recommended.

[0133] It should be noted that the initial selection data of students is collected, that is, the courses or learning resources that students have selected in the past; this data may include the courses that students have selected within a certain period of time, or their learning preferences, study duration, etc.; for example: student A selected "Calculus", "Principles of Mechanics" and "English Writing"; student B selected "Introduction to Biology", "Physical Mechanics" and "Mathematical Analysis".

[0134] Based on the courses students choose, their learning goal tendencies can be analyzed. This refers to students' interests and needs in different fields, with varying levels of difficulty and knowledge points. For example, if student A chooses "Calculus" and "Principles of Mechanics," it may indicate that they tend to study the fundamental theories and applications of mathematics and physics. Specifically, student A's course selection shows a strong interest in mathematics and physics, so their learning goal might be to improve their mathematical and physical abilities; student B's course selection shows an interest in biology, so their learning goal might be to gain a deeper understanding of life sciences and biological principles.

[0135] Based on a student's learning goals, the system will recommend supplementary learning resources that the student may need. For example, if student A's learning goal is to improve their mathematical and physical abilities, the system may recommend courses such as "Advanced Calculus" or "Advanced Physics." Specific examples: Student A (with a mathematical and physical inclination) may need supplementary learning resources such as "Linear Algebra," "Classical Mechanics," or "Electromagnetism"; Student B (with a biological inclination) may need supplementary learning resources such as "Fundamentals of Genetics" or "Principles of Ecology."

[0136] By comparing the available course resources with students' historical selections, potential course resources that might be helpful to students can be identified. For example, if student A has already taken the course "Calculus", then courses to be added might include "Advanced Calculus" or "Advanced Mathematics". These resources will be marked as potential courses to be recommended. If student B has already taken the course "Introduction to Biology", then courses to be added might include "Fundamentals of Genetics" or "Biological Experimental Techniques".

[0137] If the potential recommended course resources are very complex and have multiple teaching characteristics (such as involving cross-disciplinary knowledge points or complex course content), then it is necessary to filter out similar learning trajectory groups from the historical learning record database of the education platform based on the student's initial selection data.

[0138] In an optional embodiment, before filtering similar selection trajectory groups from the historical learning record database of the education platform based on initial selection data, the method further includes:

[0139] Obtain at least one of the following identifying features from the initial selection data: selection operation duration, selection operation frequency, time interval between two adjacent selection operations, and total selection time of course resources in the initial selection data;

[0140] Student identification is determined based on identification features and initial selection data;

[0141] Based on the learning history of the education platform, the learning status of the education platform on the current day, and the student's identity, different course recommendation strategies are adopted to recommend courses.

[0142] It should be noted that the collection of students' behavioral characteristics when selecting course resources includes: Duration of selection operation: the length of time a student spends selecting a particular course resource; for example, a student might spend 10 minutes selecting a math course but only 2 minutes selecting a language course; Frequency of selection operation: how frequently a student selects course resources; for example, a student might select a new course every day, while another student might only select one once a week; Time interval between two adjacent selection operations: the time difference between two selections of course resources; for example, a student might select three courses in one day with a time interval between each selection. The timeframes are 1 hour, 30 minutes, and 2 hours respectively; the total selection time for course resources: that is, the total study time a student spends on a particular course; for example, student A spent 50 hours on the "Calculus" course, while student B spent 30 hours on the "Biology" course; specific examples: student A selects and studies math-related courses every week, with each selection lasting approximately 30 minutes, indicating a high frequency of selection; student B selects courses less frequently, only once a week, with each selection lasting only about 10 minutes; these behavioral characteristics can help the system identify students' learning habits and preferences;

[0143] By analyzing students' selection behavior characteristics, the system can infer students' identity identifiers. These identifiers represent not only students' personal information (such as names or student IDs), but also their learning styles and habits. Based on these behavioral characteristics, the system can create a "learning profile" for each student. For example, student A (who frequently selects courses, spends a relatively long time on each selection, and frequently studies math-related courses) may be identified as a "theory-oriented, diligent student"; student B (who selects courses infrequently and spends a relatively short time on each selection) may be identified as a "study-infrequently, short-term-study-preferred student."

[0144] Based on students' identification and their historical learning data, the system will select different course recommendation strategies; different types of students require different course recommendation methods:

[0145] (a) Recommendation strategy based on historical learning: The system will view the student's past learning records on the education platform to understand what the student has learned and their learning progress; for example, if student A has previously learned a lot of math-related content, the system may recommend some more advanced math courses.

[0146] (b) Recommendation strategy based on the day’s learning: If a student has already studied on the day, the system may adjust the recommendations based on the day’s learning; for example, if student B has already selected a longer course to study today (such as “Fundamentals of Physics”), the system may recommend advanced content related to that course, or recommend some simple review courses.

[0147] (c) Student-based recommendation strategy: The system will select different recommendation strategies for students with different learning styles. For example, for students who "emphasize theory and study diligently" (such as student A), the system may recommend some longer and more in-depth courses to help them improve further. For students who "study infrequently and prefer short study sessions" (such as student B), the system may recommend some shorter or more flexible courses to avoid making them feel overwhelmed by the course load. Specific examples: Student A: Assuming that student A has studied many advanced mathematics courses in the past and studies for a long time every day, the system may recommend some courses such as "Advanced Calculus" or "Algebraic Topology" to further improve their mathematical abilities. Student B: Student B's study habit is to study once a week, and each study session is relatively short. Therefore, the system may recommend some short courses such as "Introduction to Biology" or "Basic Ecology" to ensure that the student can learn efficiently in a short period of time.

[0148] In one optional embodiment, different course recommendation strategies are employed to recommend courses based on the educational platform's historical learning data, the platform's current learning data, and student identification, including:

[0149] If historical learning data indicates that the number of times a course was studied on the education platform within a preset time period was a single study, then the course recommendation strategy adopted is: to select similar selection trajectory groups from the historical learning record database of the education platform based on the initial selection data;

[0150] If historical learning data shows that the course was studied multiple times within a preset time period, and the student's learning data and student identity indicate that the same student studied the course multiple times, then the course recommendation strategy adopted is: to recommend a learning trajectory group consisting of the course resources corresponding to the student's identity and the current time from the historical learning record database of the education platform.

[0151] If historical learning data shows that courses on the education platform have been studied multiple times within a preset time period, and the learning data and student identification on the day indicate that multiple students have studied the courses multiple times, then the course recommendation strategy adopted is: to select recommended trajectory groups from the historical learning record database of the education platform based on the initial selection data and course difficulty.

[0152] It's important to note that if a student only studies one course once within a preset timeframe (e.g., the past week, month, etc.), the system will filter similar learning trajectories from the historical database based on the student's selection behavior (e.g., study time, frequency) to recommend courses. By analyzing the similarity of historical learning trajectories, the system can recommend courses that students may be interested in or related to. For example, suppose student A only studied "a basic English course" once in the past month, and the study time was short (e.g., 30 minutes). Based on this single study data, the system will filter the trajectories of other students who have studied similar basic English courses based on their selection behavior. For instance, if other students A, B, and C have also studied similar basic English courses and subsequently studied "English writing" or "English listening" courses, these courses might be recommended to student A. The reason for the recommendation: student A may have a weaker English foundation, and the learning content is relatively simple; the system infers from historical behavior that the student may need more similar courses or learning content that interests them.

[0153] If a student has repeatedly selected and studied a particular course within a certain period (e.g., the past month), and their learning behavior on that day indicates they are continuing to study the same type of course, the system will generate recommended courses based on the student's identity and learning trajectory. This strategy targets students who are actively engaged in learning and consistently delve into a specific subject or course. For example, suppose student B has studied "Advanced Mathematics" multiple times within a past period (e.g., the past month) and also selected a related course on that day. The system will generate a learning trajectory group containing that course based on student B's historical learning records. In this case, the system might recommend advanced courses related to "Advanced Mathematics," such as "Mathematical Analysis" or "Mathematical Modeling," because student B clearly intends to continue learning more in-depth content. The rationale for the recommendation is: student B has strong learning motivation and has already studied Advanced Mathematics multiple times; therefore, advanced content related to this course is recommended to help them further improve.

[0154] If multiple students repeatedly choose to study a particular course within the same time period, and the system identifies the student group's identity and learning behavior, the system will filter relevant learning trajectory groups for recommendation based on the course's difficulty level and the students' interests. This strategy is suitable for group recommendations, especially when learning frequency is high. For example, suppose that in the past month, multiple students (student C, student D, student E, etc.) frequently studied the "Introduction to Programming" course, and on the same day, these students also continued to choose "Advanced Programming" courses. The system will aggregate these students' learning trajectories and then recommend relevant learning resources based on these trajectories. For instance, the system might recommend more challenging courses based on course difficulty, such as "Data Structures and Algorithms" or "Fundamentals of Artificial Intelligence," which may be related to the learning trajectories of students C, D, and E and meet their learning needs.

[0155] Example 2, please refer to Figure 2 This invention provides a technical solution: an intelligent recommendation system for educational course resources and a teaching effectiveness evaluation system, applicable to the aforementioned intelligent recommendation system for educational course resources and a teaching effectiveness evaluation method, comprising:

[0156] Data acquisition unit 1 is used to acquire students' initial selection data for educational course resources, wherein the initial selection data includes the course resource identifier selected by the student and the selection time information;

[0157] Similar grouping unit 2 is used to filter similar selection trajectory groups from the historical learning record library of the education platform based on the initial selection data. The historical learning record library includes multiple learning trajectory groups, and each learning trajectory group includes the identifier of the historical selected course resources with similar selection order and the corresponding learning performance.

[0158] Resource extraction unit 3 is used to input the historical selected course resources in the similar selection trajectory group into the pre-trained resource feature extraction model to obtain the core teaching point information in each historical selected course resource. The core teaching point information is used to represent the content that plays a role in students' knowledge acquisition and skill improvement.

[0159] The effectiveness prediction unit 4 is used to input the core teaching point information into the pre-trained teaching effectiveness prediction model to obtain the teaching effectiveness prediction value corresponding to each core teaching point information. The teaching effectiveness prediction value is used to characterize the potential effect of the core teaching point information on student learning.

[0160] Resource recommendation unit 5 is used to optimize and sort the historically selected course resources in the selected similar selection trajectory group based on the predicted teaching effectiveness value, so as to obtain an optimized course resource recommendation sequence; courses in the optimized course resource recommendation sequence that are different from the course resources in the students' initial selection data are selected as courses to be recommended, and the detailed teaching information corresponding to the courses to be recommended is used as recommendation details for recommendation.

[0161] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A method for intelligent recommendation of educational curriculum resources and evaluation of teaching effectiveness, characterized in that, include: Obtain students' initial selection data for educational course resources, wherein the initial selection data includes the course resource identifier selected by the student and the selection time information; Based on the initial selection data, similar selection trajectory groups are selected from the historical learning record database of the education platform. The historical learning record database includes multiple learning trajectory groups, and each learning trajectory group includes the identifier of the historical selected course resources with similar selection order and the corresponding learning performance. The historical selected course resources in the similar selection trajectory group are input into a pre-trained resource feature extraction model to obtain the core teaching point information of each historical selected course resource. The core teaching point information is used to characterize the content that plays a role in students' knowledge acquisition and skill improvement. The core teaching points are input into a pre-trained teaching effectiveness prediction model to obtain a teaching effectiveness prediction value corresponding to each core teaching point. The teaching effectiveness prediction value is used to characterize the potential effect of the core teaching points on student learning. Based on the predicted teaching effectiveness, the historically selected course resources in the selected similar selection trajectory groups are optimized and sorted to obtain an optimized course resource recommendation sequence. Courses in the optimized course resource recommendation sequence that are different from the course resources in the students' initial selection data are selected as courses to be recommended, and the detailed teaching information corresponding to the courses to be recommended is used as recommendation details for recommendation. The training method for the teaching effectiveness prediction model includes: To obtain historical education curriculum resources and students' learning performance on historical education curriculum resources; By inputting history education curriculum resources into the resource feature extraction model, the core teaching points of history in the history education curriculum resources are extracted. Input the core teaching points of history into the teaching effectiveness prediction model to determine the teaching effectiveness prediction value corresponding to each core teaching point of history. Based on the order of predicted teaching effectiveness values ​​corresponding to each core teaching point in history from high to low, a specified number of core teaching points are selected as the first core point group. At the same time, based on the order of the actual improvement in students' learning effectiveness in history from high to low, a specified number of core teaching points are selected as the second core point group. The teaching effectiveness prediction model is trained with the optimization objective of minimizing the difference between the first core point group and the second core point group.

2. The method for intelligent recommendation of educational course resources and evaluation of teaching effectiveness according to claim 1, characterized in that, The training method for the resource feature extraction model includes: To obtain historical education curriculum resources and students' learning performance on these historical education curriculum resources in history; The history education curriculum resources are input into the resource feature extraction model. The resource feature extraction model is used to determine the core teaching points of history in the history education curriculum resources. At the same time, the key teaching links in the history education curriculum resources are determined based on the students’ academic performance. The resource feature extraction model is trained with the optimization objective of minimizing the discrepancy between the core teaching points of history and the teaching content corresponding to the key teaching links in history education curriculum resources.

3. The method for intelligent recommendation of educational course resources and evaluation of teaching effectiveness according to claim 2, characterized in that, Before inputting the historically selected course resources from the similar selection trajectory group into a pre-trained resource feature extraction model to obtain the core teaching point information of each historically selected course resource, the method further includes: The historical selected course resources in the filtered similar selection trajectory groups are processed to obtain a set of course segments of a specified duration; Inputting the historically selected course resources from the similar selection trajectory group into a pre-trained resource feature extraction model yields the core teaching points information for each historically selected course resource, including: For each teaching segment in each historical selected course resource, the teaching depth index corresponding to each course type in the teaching segment is determined by the resource feature extraction model. For each teaching segment in each historical selected course resource, the teaching importance of the teaching segment is determined based on the teaching depth index corresponding to each course type at the teaching segment and the preset teaching weight corresponding to each course type. If the teaching importance corresponding to the teaching segment is greater than the preset importance threshold, then the core teaching points are determined based on the teaching content corresponding to the teaching segment.

4. The method for intelligent recommendation of educational course resources and evaluation of teaching effectiveness according to claim 3, characterized in that, Before inputting historically selected course resources from similar selection trajectory groups into a pre-trained resource feature extraction model, the method further includes: Determine the student group characteristics corresponding to the historically selected course resources in the filtered similar selection trajectory groups; Input the historical selected course resources from the similar selection trajectory group into the pre-trained resource feature extraction model, including: The core teaching points and student group characteristics are input into the teaching effectiveness prediction model, and the teaching effectiveness prediction model determines the teaching effectiveness prediction value for each core teaching point for the student group.

5. The method for intelligent recommendation of educational course resources and evaluation of teaching effectiveness according to claim 4, characterized in that, Based on the predicted teaching effectiveness values, the historically selected course resources in the selected similar selection trajectory groups are optimized and sorted, including: For each core teaching point information, if the predicted teaching effectiveness value corresponding to the core teaching point information is lower than the preset effectiveness threshold, then several core teaching point information located before the teaching segment corresponding to the core teaching point information and several core teaching point information located after the teaching segment corresponding to the core teaching point information are used as the context teaching point information corresponding to the core teaching point information. Based on the contextual teaching points, the teaching content at the corresponding teaching stages is optimized and adjusted, and the historical selected course resources in the selected similar selection trajectory groups are reordered based on the optimized and adjusted teaching content.

6. The method for intelligent recommendation of educational course resources and evaluation of teaching effectiveness according to claim 5, characterized in that, After obtaining students' initial selection data regarding educational course resources, the method further includes: Determine students' learning goal preferences based on the initial selection data; Based on the learning objectives, identify the supplementary course resources that need to be learned. The course resources to be supplemented are compared with the course resources in the initial selection data to obtain potential course resources to be recommended; If the potential course resources to be recommended include complex course resources with multiple teaching characteristics, then the step of filtering similar selection trajectory groups from the historical learning record database of the education platform based on the initial selection data is executed; If all potential course resources to be recommended are simple course resources with a single teaching characteristic, then the learning trajectory groups that include key content corresponding to the learning objectives are identified from the historical learning record database, and the course resources in the learning trajectory groups with the highest usage frequency are recommended.

7. The method for intelligent recommendation of educational course resources and evaluation of teaching effectiveness according to claim 6, characterized in that, Before filtering similar selection trajectory groups from the historical learning record database of the education platform based on the initial selection data, the method further includes: Obtain at least one of the following identifying features from the initial selection data: selection operation duration, selection operation frequency, time interval between two adjacent selection operations, and total selection time of course resources in the initial selection data; Student identification is determined based on identification features and initial selection data; Based on the historical learning data of the education platform, the learning data of the day, and the student's identity, different course recommendation strategies are adopted to recommend courses.

8. The method for intelligent recommendation of educational course resources and evaluation of teaching effectiveness according to claim 7, characterized in that, Based on the educational platform's historical learning data, the current day's learning data, and student identification, different course recommendation strategies are employed for course recommendations, including: If historical learning data indicates that the number of times a course was studied on the education platform within a preset time period was a single study, then the course recommendation strategy adopted is: to select similar selection trajectory groups from the historical learning record database of the education platform based on the initial selection data; If historical learning data shows that the course was studied multiple times within a preset time period, and the student's learning data and student identity indicate that the same student studied the course multiple times, then the course recommendation strategy adopted is: to recommend a learning trajectory group consisting of the course resources corresponding to the student's identity and the current time from the historical learning record database of the education platform. If historical learning data shows that courses on the education platform have been studied multiple times within a preset time period, and the learning data and student identification on the day indicate that multiple students have studied the courses multiple times, then the course recommendation strategy adopted is: to select recommended trajectory groups from the historical learning record database of the education platform based on the initial selection data and course difficulty.

9. An intelligent recommendation system for educational curriculum resources and a teaching effectiveness evaluation system, applicable to the intelligent recommendation system for educational curriculum resources and a teaching effectiveness evaluation method as described in any one of claims 1-8, characterized in that, include: The data acquisition unit is used to acquire the initial selection data of students for educational course resources, wherein the initial selection data includes the course resource identifier selected by the student and the selection time information; A similar grouping unit is used to filter similar selection trajectory groups from the historical learning record library of the education platform based on the initial selection data. The historical learning record library includes multiple learning trajectory groups, and each learning trajectory group includes the identifier of the historical selected course resource with a similar selection order and the corresponding learning score. The resource extraction unit is used to input the historical selected course resources in the similar selection trajectory group into the pre-trained resource feature extraction model to obtain the core teaching point information in each historical selected course resource. The core teaching point information is used to characterize the content that plays a role in students' knowledge acquisition and skill improvement. The effectiveness prediction unit is used to input the core teaching point information into a pre-trained teaching effectiveness prediction model to obtain the teaching effectiveness prediction value corresponding to each core teaching point information. The teaching effectiveness prediction value is used to characterize the potential effect of the core teaching point information on student learning. The resource recommendation unit is used to optimize and sort the historically selected course resources in the selected similar selection trajectory group based on the predicted teaching effectiveness value, so as to obtain an optimized course resource recommendation sequence. Courses in the optimized course resource recommendation sequence that are different from the course resources in the students' initial selection data are selected as courses to be recommended, and the detailed teaching information corresponding to the courses to be recommended is used as recommendation details for recommendation.

Images