Ideological and political answering system and method based on mobile terminal

Abstract

The invention relates to the technical field of answering systems, in particular to an ideological and political answering system and method based on a mobile terminal. The system comprises a user login management unit, the user login management unit carries out user authentication based on Hash encryption, and a user inputs a user name and a password through a mobile terminal to log in an ideological and political answering system; the topic selection management unit manages and maintains topics in a question bank through a question bank management module; the judgment and feedback unit is responsible for judging and analyzing the answer result of the user and sending the judgment and analysis result to the topic selection management unit for an intelligent recommendation module so as to optimize a topic recommendation mechanism; and the user interaction unit is used for providing interaction input between a user and each unit. An autonomous incremental learning mechanism is adopted, learning efficiency and a time decay factor are introduced to adjust a learning progress updating algorithm, the actual learning state of a user is accurately reflected, and a question sorting set is dynamically adjusted according to the latest learning progress.

Description

Technical Field

[0001] This invention relates to the field of question-answering systems, and more specifically, to a political and ideological question-answering system and method based on a mobile terminal. Background Technology

[0002] In the current digital age, ideological and political education (hereinafter referred to as "ideological and political education") is actively exploring the deep integration with information technology in order to enhance the coverage, attractiveness, and effectiveness of education. Mobile-based ideological and political education quiz systems, as an emerging auxiliary teaching tool, provide learners with the convenience of theoretical learning and self-assessment anytime, anywhere. However, existing systems of this kind still have many limitations in their design and fail to fully meet the special requirements of ideological and political education. On the one hand, the content system of ideological and political education is vast and highly theoretical, with close connections between knowledge points, making it easy for learners to get lost in a massive question bank. Most existing systems adopt a single question-pushing mode based on chapter order or random push. This "one-size-fits-all" approach cannot specifically address learners' weaknesses in specific theoretical modules or knowledge points, resulting in a lack of personalized guidance and low learning efficiency.

[0003] On the other hand, ideological and political education emphasizes the unity of knowledge and action and continuous deepening, requiring learning tools to dynamically reflect learners' cognitive level and theoretical mastery. However, traditional question-and-answer systems often only engage in superficial interactions of right and wrong judgment, failing to build a dynamic model that tracks changes in learners' knowledge structure and reflects their learning trajectory, making it difficult to effectively guide the transition from "knowledge transmission" to "ability internalization." Furthermore, some systems attempting to introduce recommendation algorithms either rely solely on knowledge point matching, resulting in highly homogenized recommended content; or they simply use collaborative filtering, which, when faced with less frequently practiced and specialized non-popular test points in ideological and political education, suffers from poor recommendation performance due to sparse user behavior data, or even fails to generate effective recommendations. This directly impacts the comprehensiveness of learning and the stability of the system. Simultaneously, the annotation and classification of knowledge points in ideological and political question banks heavily rely on manual labor, which is not only inefficient but also difficult to guarantee in terms of professionalism and accuracy, restricting the depth and breadth of subsequent intelligent applications.

[0004] Therefore, how to break through existing technological bottlenecks and create an ideological and political education answering system that can deeply understand the content of ideological and political education, accurately assess individual differences among learners, and provide adaptive and intelligent question recommendations is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] The purpose of this invention is to provide a political and ideological education question-answering system and method based on mobile terminals to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, on the one hand, the present invention aims to provide a political and ideological education question-answering system based on a mobile terminal, comprising: One or more processors; A memory, wherein the memory stores instructions that can be executed by the one or more processors; When the instructions are executed by the one or more processors, the system enables: The topic selection management unit is used to recommend topics from the question bank to target users based on the user learning model; The question presentation and answering unit is used to present questions recommended by the question selection management unit to the target user and receive the target user's answer results; The evaluation and feedback unit is used to generate user performance data based on the answer results and update the user learning model based on the user performance data. The topic selection management unit is configured as follows: Based on the user learning model and question attributes, an initial question ranking set is determined; Based on the target user's historical answer data with other users, a collaborative filtering recommendation score is determined; Based on the question scores in the initial question ranking set and the collaborative filtering recommendation scores, a first fusion recommendation ranking set is generated, and the questions in the first fusion recommendation ranking set are recommended to the target user.

[0007] As a preferred embodiment of the present invention: the topic selection management unit is configured to, when determining the initial topic sorting set, as follows: For each question in the question bank, a first weighted score is calculated that combines the user's mastery of the knowledge points with the difficulty of the question itself. The questions are sorted based on the first weighted score to generate the initial question sorting set.

[0008] As a preferred embodiment of the present invention: the user learning model includes the target user's mastery level of preset knowledge points; the determination of the user's mastery level of knowledge points is configured as follows: Based on the user performance data generated by the evaluation and feedback unit, and combined with the historical mastery level, the current mastery level of each knowledge point is dynamically calculated through a learning progress update algorithm.

[0009] As a preferred embodiment of the present invention: the learning progress update algorithm is configured as follows: The updated knowledge point mastery level is generated by weighting and combining the previous knowledge point mastery level with the current knowledge point mastery level corresponding to the current answer performance. The weights of the weighted fusion are controlled by a hyperparameter used to balance historical progress and current progress.

[0010] As a preferred embodiment of the present invention: the topic selection management unit is configured to, when generating the first fused recommendation ranking set, as follows: Obtain the first weighted score for each question in the initial question sorting set; The target user's predicted score for each question is calculated using a collaborative filtering algorithm, and this score is used as the collaborative filtering recommendation score. The first weighted score and the predicted score are weighted and fused together to generate a first fused score; The questions are reordered based on the first fusion score to generate the first fusion recommended ranking set.

[0011] As a preferred embodiment of the present invention: the topic selection management unit is further configured as follows: Obtain the recent answering behavior characteristics of the target user, wherein the recent answering behavior characteristics include at least the answering frequency and the answering accuracy rate; Based on the recent answering behavior characteristics, the first fusion score is adjusted to generate a final recommendation score; The questions are re-ranked based on the final recommendation score to generate a final recommendation ranking set, and the questions in the final recommendation ranking set are recommended to the target user.

[0012] As a preferred embodiment of the present invention: the evaluation and feedback unit is configured as follows: The user's answers are matched with preset standard answers to generate user performance data that represents whether the answers are correct or not; The user performance data is linked to the knowledge points associated with the question in order to update the user's mastery of the corresponding knowledge points in the user learning model.

[0013] As a preferred embodiment of the present invention, the evaluation and feedback unit is further configured as follows: When updating the user's mastery of the knowledge points, user learning efficiency and time decay factors are introduced; The user learning efficiency is determined based on the user's historical answer data and answering speed, and the time decay factor is related to the time interval since the last learning, so as to achieve incremental learning optimization of the user learning model.

[0014] As a preferred embodiment of the present invention: the topic selection management unit further includes a question bank management module, which is configured as follows: Extract keywords from the question text to generate question vectors; Calculate the similarity between the question vector and the preset knowledge point vector; Based on the similarity, the questions are automatically categorized and associated with the most relevant knowledge points.

[0015] As a preferred embodiment of the present invention: the system further includes a user login management unit, which is configured as follows: Perform hash-encrypted authentication on user login credentials; After user authentication is successful, the system assigns corresponding system operation permissions to the target user based on the preset association between user roles and permissions, and dynamically controls the interface content and functions of the question presentation and answer unit.

[0016] On the other hand, the present invention also provides a method for answering ideological and political questions based on a mobile terminal, comprising the following steps: Based on user learning models, question attributes, and users' historical answer data, recommended questions are generated for a target user. The steps for generating recommended questions include: Based on the user learning model and question attributes, an initial question ranking set is determined; Based on the target user's historical answer data with other users, a collaborative filtering recommendation score is determined; Based on the question scores in the initial question ranking set and the collaborative filtering recommendation scores, a first fusion recommendation ranking set is generated. Present the questions from the first fused recommendation ranking set to the target user, and receive the target user's answer results; User performance data is generated based on the answer results, and the user learning model is updated based on the user performance data.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention constructs a user learning model that can be updated in real time to accurately quantify learners' mastery of each ideological and political knowledge point, thus breaking away from the traditional system's extensive question-assignment method. It can intelligently identify learners' weaknesses in the theoretical framework and, combined with the difficulty attributes of the questions themselves, recommend questions most suitable to their current cognitive level, achieving truly personalized tutoring and precise practice, greatly improving learning efficiency and confidence.

[0018] More importantly, this invention creatively designs a multi-dimensional hybrid recommendation mechanism that combines rigorous knowledge point content matching with flexible user group behavior analysis (collaborative filtering). This ensures the professionalism and relevance of the recommended content in ideological and political theory, while also injecting personalization and interest by mining the behavioral preferences of similar learners, effectively alleviating the problems of "cold start" and "information cocoon." This mechanism further incorporates learners' recent answer frequency, accuracy, and other real-time behavioral characteristics for dynamic fine-tuning, enabling the recommendation results to sensitively reflect the learner's latest learning status and ensuring that the recommended questions are both reinforcing and forward-looking. This dynamic closed loop of "answering-judging-feedback-model updating-re-recommendation" constructs a continuously optimized adaptive learning path for learners, guiding them to progress gradually and spirally in the ocean of ideological and political knowledge, realizing the educational ideal of individualized instruction.

[0019] Furthermore, this invention also realizes the intelligent and automated classification and management of the ideological and political question bank through natural language processing technology, which not only significantly reduces the operation and maintenance costs, but also ensures the accuracy and consistency of knowledge point annotation, providing a solid data foundation for the reliable operation of the entire intelligent recommendation system. Attached Figure Description

[0020] Figure 1 This is an overall flowchart of the present invention; The meanings of the labels in the diagram are as follows: 1. User Login Management Unit; 11. User Authentication Login Module; 12. Permission Management Module; 2. Topic Selection Management Unit; 21. Question Bank Management Module; 22. Intelligent Recommendation Module; 221. Learning Progress Tracking and Management Module; 222. Topic Filtering and Matching Module; 223. Recommended Learning Feedback Module; 3. Topic Presentation and Answering Unit; 31. Answering Module; 32. Display Module; 4. Evaluation and Feedback Unit; 41. Answer Evaluation Module; 42. Feedback Optimization Module; 5. User Interaction Unit. Detailed Implementation

[0021] 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.

[0022] Example 1: Please refer to Figure 1 As shown, a mobile terminal-based ideological and political education question-answering system is provided, including a user login management unit 1. The user login management unit 1 performs user authentication based on hash encryption. Users log in to the ideological and political education question-answering system by entering their username and password through the mobile terminal.

[0023] In this embodiment, the user login management unit 1 includes a user authentication login module 11 and a permission management module 12.

[0024] Among them, hash encryption is integrated into the user authentication login module 11 to realize user verification.

[0025] After logging into the system, users can be assigned different user permissions based on their identity category through the permission management module 12.

[0026] After logging in, users are assigned corresponding permissions based on their roles through the permission management module 12, which involves the following steps: a. After a user logs in, the system retrieves the user's role information. Based on the logged-in user ID, the system queries the user table to obtain the user's role ID. Then, the system queries the role table using the role ID to obtain the user's role name (such as "Student", "Teacher", or "Administrator").

[0027] Specifically, the user model code is as follows: class User(db.Model): id=db.Column(db.Integer,primary_key=True) username=db.Column(db.String(50),unique=True) password=db.Column(db.String(100)) role_id=db.Column(db.Integer,db.ForeignKey('role.id')) role=db.relationship('Role',backref=db.backref('users', lazy=True)).

[0028] b. Based on the user's role, query the permissions for that role; find the permission IDs that the role possesses in the role permission association table by using the role ID, and then query the permission table based on the permission IDs to obtain detailed permission information.

[0029] Specifically, the character model code is as follows: class Role(db.Model): id=db.Column(db.Integer,primary_key=True) name = db.Column(db.String(50)) permissions=db.relationship('Permission', secondary='role_permissions').

[0030] c. Dynamically control the pages and functions that users can access based on their roles and permissions; specifically, use front-end and back-end permission control, with the back-end determining whether an operation is allowed based on the user's role, and the front-end deciding which buttons or menus are visible based on the user's permissions.

[0031] Specifically, the permission model code is as follows: class Permission(db.Model): id=db.Column(db.Integer,primary_key=True) name=db.Column(db.String(100)); The code for the role-permission association table is as follows: class RolePermission(db.Model): id=db.Column(db.Integer, primary_key=True) role_id=db.Column(db.Integer,db.ForeignKey('role.id')) permission_id=db.Column(db.Integer, db.ForeignKey('permission.id')).

[0032] This embodiment also includes a question selection management unit 2. The question selection management unit 2 manages and maintains the questions in the question bank through the question bank management module 21, and uses the intelligent recommendation module 22 to filter and sort a set of questions suitable for users using a collaborative filtering algorithm. Considering that the predicted score and the ranking of recommended questions will be affected by the user's answering time and answering frequency, the selected set of questions is re-sorted to generate a new set of ranked questions.

[0033] The question bank management module 21 is used to manage and maintain all questions in the question bank, and to classify the questions based on a classification algorithm. The specific steps involved are as follows: Existing knowledge point set Each knowledge point Corresponding to a knowledge point vector (Composed of the TF-IDF keyword weights of all questions under this knowledge point), keywords are extracted from the question text using the TF-IDF algorithm to generate question vectors. Cosine similarity is used to calculate the similarity between the question and each knowledge point. ; Based on the calculated similarity, select the knowledge points most relevant to the question. The question is assigned to the relevant knowledge point using a classification algorithm. ; In the formula, Representation and title Relevant knowledge points; Indicates the first question in the question bank. One knowledge point; Indicates the topic With the question bank One knowledge point The similarity between them.

[0034] Furthermore, the intelligent recommendation module 22 includes a learning progress tracking and management module 221, a question filtering and matching module 222, and a recommended learning feedback module 223.

[0035] Among them, the learning progress tracking and management module 221 updates the user's learning progress based on the learning progress update algorithm. Based on user learning progress The question filtering and matching module 222 uses a knowledge point matching algorithm to select a set of questions from the question bank that match the current learning progress. The recommendation and learning feedback module 223 uses a collaborative filtering algorithm to select questions with high matching degree with similar users based on the user's current learning progress, and recalculates new weighted scores for the selected questions. Sort the set of questions Reorder the questions to obtain a new sorted set. It recommends questions that match the user's current learning progress.

[0036] In this embodiment, the intelligent recommendation module 22 filters a set of questions from the question bank. The specific steps involved are as follows: If the user's learning progress If a vector contains multiple knowledge points, then the mastery level of each knowledge point is: : but ; In the formula, The number representing the knowledge point; This indicates the total number of knowledge points.

[0037] How well did you grasp this knowledge point in this answer? After the user completes the question, the learning progress update algorithm is used to update the learning progress. ; The learning progress update algorithm is as follows: ; In the formula, This indicates that after completing a question, the user will... The level of mastery of each knowledge point after the update; This represents a hyperparameter that controls the balance between historical and current progress. It is a parameter between 0 and 1, used to control the relative weight of historical and current progress. The higher the value, the more importance is attached to historical progress; The lower the value, the more attention is paid to the current learning situation; This indicates the number of questions a user answered correctly after completing a given set of questions. This indicates the total number of questions a user has attempted to answer regarding this knowledge point; This indicates that the user had views on the first update before the update. The degree of mastery of each knowledge point.

[0038] Mastery of the updated knowledge points Different learning progress levels are determined by assigning different thresholds to them; ; In the formula, , , , ..., These all represent thresholds for dividing learning progress, and different thresholds are used to classify the degree of mastery of updated knowledge points. It is divided into different learning progress levels.

[0039] Known user learning progress The algorithm calculates the user's level of mastery by matching knowledge points with questions in the question bank. Related knowledge points Matching degree Select questions that match your current learning progress: ; In the formula, Indicates the user's learning progress questions in the question bank Related knowledge points The degree of matching; The title The set of related knowledge points; The summation symbol represents the number of knowledge points in the set. This indicates traversing the problem. Each related knowledge point ; It is the user's knowledge point The formula assesses the user's level of knowledge matching with the question by calculating the user's average level of mastery of the knowledge points involved in the question.

[0040] For each question, consider the knowledge matching degree. and the difficulty of the questions Calculate the weighted score : ; In the formula, Indicates the topic The weighted score takes into account both the degree to which the question matches the user's current knowledge level and the difficulty of the question itself. This indicates the difficulty level of the questions. The initial value is set by the administrator and will be updated based on the historical accuracy rate. ); and These are two hyperparameters, used as weighting coefficients for knowledge matching degree and question difficulty, respectively, to balance the two strategies of "recommending questions related to knowledge points that users have already mastered" and "recommending questions with a certain level of challenge".

[0041] Based on weighted scoring Questions that meet the screening criteria are divided into question sets. The algorithm sorts the questions from highest to lowest matching degree and selects the questions with the highest matching degree to recommend to users. The sorting algorithm is as follows: ; In the formula, This represents the sorted set of questions. This represents the set of questions.

[0042] Specifically, ; In the formula, This represents a sorting function whose input is a set of problems. and weighted scoring criteria ; in, , of which each This refers to a question.

[0043] Specifically, functions The calculation steps are as follows: Based on all the questions Values, from largest to smallest, for the set of problems. The questions are sorted.

[0044] The sorting function is expressed as: ; in, This indicates the question with the highest score; The second highest indicates the highest, and so on.

[0045] The sorted set of questions is the recommended question sorting set. .

[0046] Recommended Learning Feedback Module 223: Set of Questions Sort by Question This yields a new sorted set of questions. The specific steps involved are as follows: To find user groups similar to the target users, since highly similar user groups often share certain implicit preferences and behavioral patterns, collaborative filtering algorithms are used to filter out users who are similar to the target users. High matching degree individual users Extract users Completed problem set .

[0047] Calculate user similarity using the cosine similarity algorithm: ; In the formula, Indicates user A collection of questions that have been answered; Indicates user A collection of questions that have been answered; Indicates user For the question The rating; Indicates user For the question The rating; Indicates user and users A collection of questions that have all been answered; Indicates user and users The similarity.

[0048] The expression for rating prediction based on similar users is: ; In the formula, Indicates the predicted user For the question The rating; Indicates user The average score of all questions answered; Indicates similar users For the question The actual score; Indicates user Average rating; This indicates the target user. Most similar front Each of the users.

[0049] Calculate the user using the above formula. For the set of questions Each question Predicted score Collaborative filtering is used to assign personalized preference weights to the content recommendation results.

[0050] Next, sort the set of questions. Each question in The new weighted score, which incorporates collaborative filtering predictions, is calculated. In order to achieve Reordering.

[0051] Among them, the new weighted score The specific expression is: ; This is the normalized content recommendation score. It is derived from the previously calculated weighted score. The score combines the matching degree of knowledge points and the difficulty of the questions. In order to compare with the score of collaborative filtering on the same scale, its value is mapped to the [0,1] interval through min-max scaling. This is the normalized collaborative filtering prediction score. It is derived from the predicted score calculated based on user similarity. Similarly, it needs to be normalized to the [0,1] interval. It is a mixed weight coefficient between 0 and 1. It is used to balance the importance of content recommendation and collaborative filtering recommendation. When the value is close to 1, the recommendation results rely more on the user's knowledge model and the content of the question itself (i.e., );when When the value is close to 0, the recommendation results depend more on the behavioral preferences of similar users.

[0052] Based on the calculated new weighted score For the original set of questions Reorder the questions to generate a new sorted set. This method ensures that even if a question is not popular among similar user groups, it still has a chance to be recommended as long as it highly matches the user's knowledge model. This effectively avoids the risk of an empty recommendation list and improves the robustness of the system.

[0053]

[0054] Considering that predicted scores and the ranking of recommended questions are influenced by users' overall behavioral patterns, historical user behavior characteristics are incorporated into the revised weighted scores. Fine-tuning was performed to obtain the final recommended score. : ; In the formula, It is the behavior-adjusted recommendation score ultimately used to generate the ranked set of questions. It is a user behavior score that reflects the user's behavior. My recent overall learning status, and independent of the questions currently being graded. , These are the weighting coefficients of user behavior features, used to control the magnitude of the influence of the behavior moderating factor.

[0055] User behavior rating The calculation method is as follows:

[0056] in, It is the normalized frequency of recent user responses, used to measure user learning activity. It is the normalized recent overall accuracy rate of users, used to measure the user's current learning performance. and These are the weighting coefficients for activity level and accuracy, respectively.

[0057] Based on the final weighted score right Reorder the questions to generate a new sorted set. .

[0058] It also includes a question presentation and answer unit 3, which is used to provide users with question display and answer operation.

[0059] Specifically, the question presentation and answering unit 3 includes an answering module 31 and a display module 32.

[0060] The answer module 31 receives the answer input from the user. After the user completes the answer, the answer module 31 sends the answer result to the answer evaluation module 41.

[0061] Meanwhile, the user's answer results are displayed through the display module 32. Recommended questions obtained from the question selection management unit 2 are transmitted to the display module 32 via the user interaction unit 5. Based on the question ranking set filtered by the intelligent recommendation module 22, the display module 32 presents questions suitable for the user's current learning progress from the recommended question list, displaying them in priority order. The display module 32 can also dynamically adjust the interface content based on the user's role (e.g., student, teacher, administrator) and the answer progress. For example, teachers may have different permissions, such as viewing students' answer records and learning progress, while administrators can manage users, the question bank, etc.

[0062] This embodiment also includes an evaluation and feedback unit 4, which is responsible for evaluating and analyzing the user's answer results and sending the evaluation and analysis results to the topic management unit 2 as the intelligent recommendation module 22 to optimize the topic recommendation mechanism.

[0063] The evaluation and feedback unit 4 includes a question evaluation module 41 and a feedback optimization module 42.

[0064] In this embodiment, the answer evaluation module 41 uses an answer matching algorithm to output the answer evaluation result, and the feedback optimization module 42 analyzes the user's answer result based on the knowledge point mastery calculation algorithm, and provides the necessary input parameters for the learning progress update algorithm in the question selection management unit 2.

[0065] The specific steps involved in the answer matching algorithm to evaluate the answer result are as follows: Get the correct answers from the question bank Get the answers submitted by users ; Comparison using answer matching algorithm and Judge the answer submitted by the user Is it consistent with the correct answer? completely consistent; The specific answer matching algorithm is as follows: ; In the formula, 1 represents the answer submitted by the user. With the correct answer Completely identical; 0 indicates the user-submitted answer. Is it consistent with the correct answer? Inconsistent.

[0066] The knowledge point mastery calculation algorithm analyzes the user's answer results and provides necessary input parameters for the learning progress update algorithm in question selection management unit 2. The specific steps involved are as follows: Link the question to a set of knowledge points ; Based on the user's answer results ( Is it equal to This generates the user's answers to each related knowledge point in this quiz. Performance observations ; Specifically, the observed values This performance observation value This will be passed as a core parameter to the learning progress update algorithm in topic selection management unit 2, for updating the unified user learning progress model. .

[0067] The feedback optimization module 42 will assess the user's mastery of each knowledge point. A topic selection management unit 2 is introduced to enable autonomous incremental learning. During the incremental learning phase, user learning efficiency is incorporated, and the learning progress update algorithm is used to update the learning progress. Sort the set of questions Readjust.

[0068] Specifically, the process of autonomous incremental learning is as follows: Based on the existing learning progress update algorithm, user learning efficiency is introduced. and time decay factor : ; in, This indicates that the user is targeting specific knowledge points in the current question. Performance observations, ; user In the knowledge points Learning efficiency It was calculated based on historical answer data and answering speed.

[0069] In the formula, This indicates that during the self-incremental learning phase, after completing a question, the user... The level of mastery of each knowledge point after the update; A parameter representing the attenuation factor; It indicates the time since the last study session.

[0070] Among them, the user interaction unit 5 is used to provide interactive input between the user and each unit.

[0071] Specifically, user interaction unit 5 provides login and permission control input for login management unit 1, and provides answer input for question presentation and answering unit 3.

[0072] Specifically, the user interaction unit 5 collects user login information, such as username and password, through a graphical user interface (GUI). This information is then passed to the user authentication login module 11 in the user login management unit 1 for performing hash-based encryption verification.

[0073] Based on the user's identity category (e.g., student, teacher, administrator), the user interaction unit 5 will display corresponding menus or options, allowing the user to select the functions or pages they wish to access. This involves obtaining the user's role information from the access control module 12 and dynamically adjusting the visibility and operability of user interface elements accordingly.

[0074] When a user starts answering questions, the user interaction unit 5 provides an intuitive interface for the user to easily input answers, including controls such as radio buttons, checkboxes, and text boxes, depending on the type of question. The user interaction unit 5 is the direct interface between the user and the system, ensuring that the user can easily log in, select functions within their authorized scope, and effectively participate in the question-answering activity.

[0075] Example 2: The difference between Example 2 and Example 1 is that this example is a method for answering political and ideological questions based on a mobile terminal, including the following steps: Based on user learning models, question attributes, and users' historical answer data, recommended questions are generated for a target user. The steps for generating recommended questions include: Based on the user learning model and question attributes, an initial question ranking set is determined; Based on the target user's historical answer data with other users, a collaborative filtering recommendation score is determined; Based on the question scores in the initial question ranking set and the collaborative filtering recommendation scores, a first fusion recommendation ranking set is generated. Present the questions from the first fused recommendation ranking set to the target user, and receive the target user's answer results; User performance data is generated based on the answer results, and the user learning model is updated based on the user performance data.

[0076] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A mobile terminal-based ideological and political education question-answering system, comprising: One or more processors; A memory, wherein the memory stores instructions that can be executed by the one or more processors; The characteristic is that, when the instruction is executed by the one or more processors, it causes the system to implement: The topic selection management unit (2) is used to recommend topics to target users from the question bank based on the user learning model; The question presentation and answering unit (3) is used to present the questions recommended by the question selection management unit (2) to the target user and receive the answer results from the target user; The evaluation and feedback unit (4) is used to generate user performance data based on the answer results and update the user learning model based on the user performance data; The topic selection management unit (2) is configured as follows: Based on the user learning model and question attributes, an initial question ranking set is determined; Based on the target user's historical answer data with other users, a collaborative filtering recommendation score is determined; Based on the question scores in the initial question ranking set and the collaborative filtering recommendation scores, a first fusion recommendation ranking set is generated, and the questions in the first fusion recommendation ranking set are recommended to the target user.

2. The system according to claim 1, characterized in that, The topic selection management unit (2) is configured as follows when determining the initial topic sorting set: For each question in the question bank, a first weighted score is calculated that combines the user's mastery of the knowledge points with the difficulty of the question itself. The questions are sorted based on the first weighted score to generate the initial question sorting set.

3. The system according to claim 1, characterized in that, The user learning model includes the target user's mastery of preset knowledge points; the determination of the user's mastery of knowledge points is configured as follows: Based on the user performance data generated by the evaluation and feedback unit (4), and combined with the historical mastery level, the mastery level of each knowledge point is dynamically calculated through the learning progress update algorithm.

4. The system according to claim 3, characterized in that, The learning progress update algorithm is configured as follows: The updated knowledge point mastery level is generated by weighting and combining the previous knowledge point mastery level with the current knowledge point mastery level corresponding to the current answer performance. The weights of the weighted fusion are controlled by a hyperparameter used to balance historical progress and current progress.

5. The system according to claim 1, characterized in that, When generating the first fusion recommendation ranking set, the topic selection management unit (2) is configured as follows: Obtain the first weighted score for each question in the initial question sorting set; The target user's predicted score for each question is calculated using a collaborative filtering algorithm, and this score is used as the collaborative filtering recommendation score. The first weighted score and the predicted score are weighted and fused together to generate a first fused score; The questions are reordered based on the first fusion score to generate the first fusion recommended ranking set.

6. The system according to claim 5, characterized in that, The topic selection management unit (2) is also configured as follows: Obtain the recent answering behavior characteristics of the target user, wherein the recent answering behavior characteristics include at least the answering frequency and the answering accuracy rate; Based on the recent answering behavior characteristics, the first fusion score is adjusted to generate a final recommendation score; The questions are re-ranked based on the final recommendation score to generate a final recommendation ranking set, and the questions in the final recommendation ranking set are recommended to the target user.

7. The system according to claim 1, characterized in that, The evaluation and feedback unit (4) is configured as follows: The user's answers are matched with preset standard answers to generate user performance data that represents whether the answers are correct or not; The user performance data is linked to the knowledge points associated with the question in order to update the user's mastery of the corresponding knowledge points in the user learning model.

8. The system according to claim 3, characterized in that, The evaluation and feedback unit (4) is also configured as follows: When updating the user's mastery of the knowledge points, user learning efficiency and time decay factors are introduced; The user learning efficiency is determined based on the user's historical answer data and answering speed, and the time decay factor is related to the time interval since the last learning, so as to achieve incremental learning optimization of the user learning model.

9. The system according to claim 1, characterized in that, The topic selection management unit (2) further includes a question bank management module (21), which is configured as follows: Extract keywords from the question text to generate question vectors; Calculate the similarity between the question vector and the preset knowledge point vector; Based on the similarity, the questions are automatically categorized and associated with the most relevant knowledge points.

10. The system according to claim 1, characterized in that, The system also includes a user login management unit (1), which is configured as follows: Perform hash-encrypted authentication on user login credentials; After user authentication is successful, the system operation permissions are assigned to the target user according to the preset association between user roles and permissions, and the interface content and functions of the question presentation and answer unit (3) are dynamically controlled.

11. A method for answering ideological and political questions based on mobile terminals, characterized in that, Includes the following steps: Based on user learning models, question attributes, and users' historical answer data, recommended questions are generated for a target user. The steps for generating recommended questions include: Based on the user learning model and question attributes, an initial question ranking set is determined; Based on the target user's historical answer data with other users, a collaborative filtering recommendation score is determined; Based on the question scores in the initial question ranking set and the collaborative filtering recommendation scores, a first fusion recommendation ranking set is generated. Present the questions from the first fused recommendation ranking set to the target user, and receive the target user's answer results; User performance data is generated based on the answer results, and the user learning model is updated based on the user performance data.

Images