A learning state tracking method and system based on a visual knowledge map

Abstract

This invention discloses a learning status tracking method and system based on a visualized knowledge map. The method includes: acquiring the hierarchical knowledge structure of the learning content; generating a visualized knowledge map based on the hierarchical knowledge structure; acquiring user learning behavior data for knowledge nodes, wherein the learning behavior data and map nodes are mapped to each other; calculating the node status information of the map nodes based on the learning behavior data, wherein the node status information is used to characterize the user's mastery or learning progress of the corresponding knowledge node; rendering the visualized knowledge map in the display interface, and dynamically adjusting the visual presentation attributes of the map nodes according to the node status information, so as to uniformly present the knowledge structure and learning status in the visualized knowledge map. By mapping user learning behavior data to nodes of the visualized knowledge map and dynamically adjusting the visual presentation attributes, the unified presentation and dynamic tracking of the knowledge structure and learning status on the same spatial carrier are achieved.

Description

Technical Field

[0001] This invention relates to the field of knowledge graph visualization technology, and in particular to a learning state tracking method and system based on a visualized knowledge map. Background Technology

[0002] In the field of educational informatization and online learning, learning support systems are widely used to help learners manage learning content and track their learning progress. In existing technologies, learning content is often organized using hierarchical structures such as chapters, modules, or directory trees, while user learning behavior data, including learning records, practice results, and mastery levels, is often stored independently in a backend database. This separation of knowledge structure and learning behavior data makes it difficult for learners to intuitively map their progress information into the overall knowledge system framework when checking their learning status.

[0003] Existing learning progress feedback largely relies on abstract forms such as numerical statistics, progress bars, or completion rates, lacking spatial and immersive visualization methods. This prevents learners from simultaneously perceiving the knowledge structure and learning status within a single view, making it difficult to quickly identify weak areas or understand the overall learning distribution. Furthermore, the lack of a unified spatial framework to integrate knowledge hierarchy with dynamic learning states leads to a fragmented learning process, delayed feedback mechanisms, and hinders learners' ability to comprehensively grasp and specifically adjust their knowledge acquisition. Therefore, how to construct a method that can uniformly and spatially present and dynamically track hierarchical knowledge structures and user learning status has become a pressing issue in the current technological field. Summary of the Invention

[0004] In view of this, this invention proposes a learning state tracking method and system based on a visualized knowledge map, which can achieve unified presentation and dynamic tracking of knowledge structure and learning state on the same spatial carrier. This invention provides the following technical solution: A learning state tracking method based on a visualized knowledge map includes: Obtain the hierarchical knowledge structure of the learning content, wherein the hierarchical knowledge structure includes multiple knowledge nodes and the hierarchical relationship between the nodes; A visual knowledge map is generated based on the hierarchical knowledge structure. The visual knowledge map includes map nodes corresponding to the knowledge nodes and connection relationships representing the hierarchical relationships. Acquire user learning behavior data for the knowledge nodes, and the learning behavior data is mapped to the map nodes; The node status information of the map node is calculated based on the learning behavior data. The node status information is used to characterize the user's mastery or learning progress of the corresponding knowledge node. The visualized knowledge map is rendered in the display interface, and the visual presentation attributes of the map nodes are dynamically adjusted according to the node status information, so as to uniformly present the knowledge structure and learning status in the visualized knowledge map.

[0005] Optionally, generating a visual knowledge map based on the hierarchical knowledge structure includes: The hierarchical knowledge structure is converted into a star-map knowledge map, and the map nodes in the star-map knowledge map are used to calculate their positions using a force-directed layout algorithm. The force-guided layout algorithm applies a variety of constraint force fields, including inter-node repulsive force to maintain the dispersed distribution of nodes, connection tension to maintain hierarchical relationships, collision detection force to reduce label overlap, and central centripetal force to stabilize the overall structure at the center of the viewport. After each frame iteration, the node coordinates are updated and the nodes and connections are re-rendered to form a stable and interactive star map layout.

[0006] Optionally, obtaining the user's learning behavior data for the knowledge node includes: Maintain the user's daily learning volume and historical cumulative learning volume for the knowledge nodes respectively; When a user is detected to be engaging in learning activities, it is determined whether the most recent learning date is the current day. If it is the current day, the learning amount for that day is accumulated; if it is not the current day, the learning amount for that day is reset and the most recent learning date is updated, while the historical cumulative learning amount is accumulated.

[0007] Optionally, the learning behavior data includes incorrect answer data, and the method further includes: The incorrect question data is bound and stored with the corresponding knowledge nodes. The incorrect question data includes at least the question reference, answer information, correctness result and answer time. The incorrect question data is used to generate multidimensional structured tags, which include at least knowledge path tags, error type tags, and difficulty tags. In the visualized knowledge map, the corresponding map nodes are aggregated and positioned according to the distribution density of the incorrect question data to form a spatialized error distribution view.

[0008] Optionally, calculating the node state information of the map node based on the learning behavior data includes: Construct a learning memory that evolves across time, wherein the learning memory includes at least node mastery, stability, forgetting risk, and error profile parameters; When learning, answering, or reviewing events occur, the parameters in the learning memory are updated and their evolution is calculated. It supports summarizing the learning memory states of leaf nodes upwards to generate summary states of upper-level classification nodes for overall state evaluation.

[0009] Optionally, the method further includes: Based on the node status information and the learning memory, personalized review recommendation information is generated, including review priority, a set of recommended incorrect questions, and a review path; The review recommendation information is mapped to the corresponding map nodes in the visualized knowledge map, and then visualized through highlighting, recommendation halo, or navigation prompts to align the recommendation results with the knowledge space.

[0010] Optionally, the method further includes: Record the results after the user completes the review task corresponding to the review recommendation information. The results include at least whether the error correction was successful, the time taken, and the type of error caused by the error if the error recurs. The results are written back to the learning memory to dynamically adjust the mastery, stability, and forgetting risk of the corresponding knowledge nodes; The recommendation strategy is adaptively optimized based on the adjusted parameters, with nodes that have been consolidated being downweighted and nodes that repeatedly make mistakes being upweighted.

[0011] Optionally, dynamically adjusting the visual presentation attributes of the map nodes based on the node status information includes: A learning marker for the day is overlaid on the outside of the map node to visually distinguish immediate learning outcomes from historical cumulative status.

[0012] This invention further discloses a learning state tracking system based on a visualized knowledge map, comprising: The structure acquisition module is used to acquire the hierarchical knowledge structure of the learning content, wherein the hierarchical knowledge structure includes multiple knowledge nodes and the hierarchical relationship between the nodes; A map generation module is used to generate a visual knowledge map based on the hierarchical knowledge structure. The visual knowledge map includes map nodes corresponding to the knowledge nodes and connection relationships representing the hierarchical relationships. The data acquisition module is used to acquire user learning behavior data for the knowledge nodes, and the learning behavior data is mapped to the map nodes; The status calculation module is used to calculate the node status information of the map node based on the learning behavior data. The node status information is used to characterize the user's mastery or learning progress of the corresponding knowledge node. The rendering and display module is used to render the visualized knowledge map in the display interface and dynamically adjust the visual presentation attributes of the map nodes according to the node status information, so as to uniformly present the knowledge structure and learning status in the visualized knowledge map.

[0013] The present invention further discloses a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method of any one of claims 1-8.

[0014] According to the technical solution of the present invention, by acquiring the hierarchical knowledge structure of the learning content and generating a visualized knowledge map containing map nodes and connections, a mapping relationship is established between the user's learning behavior data for knowledge nodes and the map nodes. Then, based on the learning behavior data, node status information representing the degree of mastery or learning progress is calculated, and the visual presentation attributes of the map nodes are dynamically adjusted in the display interface according to the node status information. This achieves a unified presentation and dynamic tracking of knowledge structure and learning status on the same spatial carrier, enabling learners to intuitively perceive the integration of knowledge context and their own learning progress in the same view. This effectively solves the problems of the separation of knowledge structure and learning behavior, the single form of progress feedback, and the lack of spatial presentation in the prior art. It helps learners quickly locate weak areas and grasp the overall knowledge mastery, significantly improving the efficiency of visual expression and targeted adjustment capabilities of the learning process. Attached Figure Description

[0015] For illustrative and not limiting purposes, the present invention will now be described in conjunction with embodiments and accompanying drawings, wherein: Figure 1 This is a flowchart illustrating the learning state tracking method based on a visual knowledge map in an embodiment of the present invention. Figure 2 This is a schematic diagram of the constituent modules of the learning state tracking system based on a visual knowledge map in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of the electronic device in an embodiment of the present invention. Detailed Implementation

[0016] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present application.

[0017] It should be noted that, where there is no conflict, the embodiments and features of the embodiments in this application can be combined with each other. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0018] refer to Figure 1 This embodiment discloses a learning state tracking method based on a visualized knowledge map, including: S100: Obtain the hierarchical knowledge structure of the learning content, wherein the hierarchical knowledge structure includes multiple knowledge nodes and the hierarchical relationship between the nodes.

[0019] During the service startup phase, a knowledge point classification system for the learning subjects is loaded from a pre-configured structured configuration file. This knowledge point classification system constitutes the hierarchical knowledge structure of the learning content. The structured configuration file is organized in a multi-level hierarchical structure, with the top level set as several knowledge category nodes. Each category node recursively contains sub-category nodes or leaf learning units to express the hierarchical relationship and knowledge path between the knowledge points in the learning content.

[0020] Specifically, each node entry in the hierarchical knowledge structure includes at least a unique node identifier, a display name, feature keywords, and descriptive text. The unique node identifier is used to uniquely distinguish each knowledge node internally; the display name is used for visual display on the client interface; the feature keywords are used to assist in the retrieval and matching of knowledge nodes; and the descriptive text is used to provide a semantic description of the knowledge node.

[0021] After parsing the structured configuration file, the following three basic data structures are generated to complete the acquisition and initialization of the hierarchical knowledge structure: First, a hierarchical data structure is generated with "Learning Subjects" as the root node. This data preserves the parent-child node relationships in the original configuration and is used to determine the node connection relationships and hierarchical topology required for subsequent star map construction.

[0022] Secondly, a mapping table is generated between node identifiers and knowledge paths. The knowledge path is encoded using a combination of "parent node identifier / ... / current node identifier". This mapping table is used to uniformly associate user learning data, incorrect question data, and learning memory data in subsequent steps, ensuring that learning behavior data can be accurately mapped to specific knowledge nodes in the hierarchical knowledge structure.

[0023] Third, generate a list of category nodes. This list is used to determine the statistical summary criteria and the visual mapping rules for different category regions in the star map. For example, when assigning theme colors to different top-level categories, the inheritance relationship of subordinate nodes is determined based on this list.

[0024] By using the above methods, static configuration information is transformed into a unified structural foundation that can directly carry learning status and behavioral data, realizing the digital acquisition of hierarchical knowledge structures and laying a data foundation for the subsequent generation of visualized knowledge maps based on this structure.

[0025] S200: Generate a visual knowledge map based on the hierarchical knowledge structure. The visual knowledge map includes map nodes corresponding to the knowledge nodes and connection relationships representing the hierarchical relationships.

[0026] Using the hierarchical structure data generated in step S100, it is converted into a node array and a connection array required for visualizing the knowledge map. Specifically, each knowledge node in the hierarchical knowledge structure is mapped to a map node in the visual knowledge map, and the parent-child hierarchical relationship between knowledge nodes is mapped to the connection relationship between map nodes, thereby constructing the basic topology of the knowledge map.

[0027] In this embodiment, the visualized knowledge map is specifically represented as a star-map-style knowledge map. After receiving the knowledge structure data, the client initializes the physics simulation engine to perform layout calculations for the map nodes. The layout calculation employs a force-directed layout algorithm, applying various constraint force fields in the star-map space to determine the spatial coordinates of each map node. First, apply repulsive forces between nodes to maintain their dispersed distribution. The root node has the maximum repulsive strength to keep the overall structure centered, while the other nodes have moderate repulsive strengths. Second, apply connecting tension to maintain the hierarchical relationship. The connection distance between upper-level nodes is set to be longer to expand the main branches, while the connection distance between lower-level nodes is dynamically adjusted according to the length of the node text. Third, apply collision detection forces by setting a protection radius for each map node. The protection radius of leaf nodes is dynamically calculated based on their text length to reduce label overlap. Finally, apply a central centripetal force to stabilize the overall star map structure at the center of the viewport.

[0028] The physics simulation engine updates node coordinates and re-renders nodes and connections after each frame iteration, thus forming a stable and interactive star map layout. During the rendering phase, the visual effects are uniformly mapped according to the node's category and state: the root node is rendered as a large node with a glow effect, representing the core knowledge of the learning subject; each top-level category node is assigned a preset theme color, and its subordinate nodes inherit this theme color to form clear visual partitions. For illuminated map nodes, the corresponding theme color is used for filling, and the brightness or glow intensity is adjusted according to the learning depth mapping value; for unilluminated map nodes, only the outline is drawn and the display is weakened.

[0029] In this way, the abstract hierarchical knowledge structure is transformed into a visual knowledge map with spatial positional relationships and visual state expression, providing a unified spatial carrier for the mapping and state tracking of subsequent learning behavior data.

[0030] S300: Obtain user learning behavior data for the knowledge node, wherein the learning behavior data is mapped to the map node.

[0031] The learning behavior data mainly includes user learning statistics, incorrect answer data, and learning memory data. A mapping relationship between learning behavior data and map nodes is established through the mapping table between node identifiers and knowledge paths generated in step S100. Specifically, the knowledge path serves as a unique index key, binding dynamically generated user behavior data to specific knowledge nodes in the hierarchical knowledge structure, and then mapping it to the corresponding map nodes in the visualized knowledge map.

[0032] First, regarding the acquisition and storage of user learning statistics, a dedicated data table is established in the database for user learning statistics. Each record uses a combination of user identifier and knowledge path as a unique constraint to ensure a one-to-one correspondence between learning behavior and specific knowledge nodes. When a user completes the learning behavior of a question, the target storage path is determined based on the knowledge path corresponding to the question, and the following update logic is executed: It checks whether a learning statistics record already exists on the path; if it exists, it checks whether the most recent learning date in the record is the current day. If it is the current day, the learning volume for that day is accumulated; if not, the learning volume for that day is reset and the most recent learning date is updated, while the cumulative learning volume is also accumulated; if it does not exist, a new learning statistics record is created, the cumulative learning volume and the learning volume for the current day are initialized, and the first learning time and the most recent learning date are recorded. For learning behaviors that can only be located to the upper-level classification node but cannot be accurately located to the leaf learning unit, placeholders are used as the final-level node identifiers for storage. The placeholder records do not participate in the rendering of leaf nodes but participate in the statistical summary of upper-level nodes to ensure the continuity of learning behavior in the overall structure of the star map.

[0033] Secondly, regarding the acquisition and binding of incorrect question data, the correctness of the answer is judged after the learner submits the answer. If it is judged to be incorrect, the incorrect question collection and archiving process is triggered. Incorrect questions are strongly bound to corresponding knowledge paths, and the recorded information includes at least the user identifier, question reference or summary, node path, answer time, time consumed, answer result, number of errors, and question source. In the incorrect question classification stage, using the knowledge path as the main axis, incorrect questions are mapped to corresponding nodes in the star map, and further error cause labels and feature labels are generated. The error cause labels include at least the types of conceptual misunderstanding deviation, misuse of formulas or theorems, broken reasoning chain, calculation error, and omission of conditions in question reading, and may be accompanied by a confidence level description. Through the above method, incorrect question data is structurally associated with specific map nodes, forming a spatialized error distribution basis.

[0034] Secondly, regarding the acquisition and evolution of learning memory data, a learning memory data structure is constructed for each user to model their learning status across knowledge nodes over time. This learning memory data includes at least parameters such as node mastery, stability, forgetting risk, error profile, and review history. When learning events, answering events, or review events occur, the learning memory is dynamically updated and its evolution is calculated. For example, when an answer is correct, the node mastery is increased and the forgetting risk is reduced; when an answer is incorrect, the mastery is reduced and the error profile is updated. It also supports summarizing the memory states of leaf nodes upwards to generate memory summaries for higher-level classification nodes.

[0035] By using the knowledge path as a unified index, scattered learning statistics, incorrect answer records, and memory parameters are aggregated into a learning behavior data set that is mapped to map nodes. This data set not only records the user's historical cumulative learning, but also distinguishes the learning outcomes of the day, and integrates the distribution of incorrect answers and long-term evolution status, providing a complete data foundation for subsequent calculation of node status information and dynamic adjustment of visual presentation attributes.

[0036] S400: Calculate the node status information of the map node based on the learning behavior data. The node status information is used to characterize the user's mastery or learning progress of the corresponding knowledge node.

[0037] Specifically, a hierarchical calculation method is used to perform state calculations on leaf nodes and upper-level classification nodes respectively, and the calculation results are written into a node state mapping table. The node state mapping table uses the node path as the key to store the display state of each map node.

[0038] First, for each leaf node, query the user's learning statistics records across all leaf node paths. For each record, determine if its most recent learning date is the current day to verify the validity of the learning status for that day. Based on the cumulative learning amount, calculate the learning depth mapping value for the corresponding node. This learning depth mapping value is calculated using a non-linear growth function: when the cumulative learning amount is zero, the mapping value is zero; when the cumulative learning amount is one, the mapping value is a preset base value; as the cumulative learning amount increases, the mapping value gradually increases according to a logarithmic curve and tends to saturate, to avoid achieving maximum display effect with only a small amount of learning.

[0039] Secondly, state evolution calculations are performed using learning memory data. The learning memory includes at least node mastery, stability, forgetting risk, and error profile parameters. During learning, answering, or review events, the parameters in the learning memory are updated and their evolution is calculated. Specifically, when an answer is correct, node mastery is increased and forgetting risk is reduced; when an answer is incorrect, mastery is reduced and the error profile is updated; when review and error correction are successful, stability is increased and forgetting risk is reduced. The mastery level representation value in the node state information is determined based on a combination of node mastery and stability parameters in the learning memory.

[0040] Secondly, for upper-level classification nodes, statistical summaries are performed on each node through grouping and aggregation to obtain the cumulative learning volume, daily learning volume, and earliest first learning time of its subordinate nodes, and the learning depth mapping value of the corresponding upper-level node is calculated accordingly. It supports summarizing the learning memory states of leaf nodes upwards to generate memory summaries of upper-level classification nodes for overall state evaluation.

[0041] Finally, the above calculation results are combined to generate the final node status information. The node status information includes the node's lit-up status, the daily learning tag, and a visual mapping value reflecting the learning depth or mastery level. Specifically, the node lit-up status is determined based on whether the cumulative learning amount is greater than zero; the daily learning tag is determined based on whether the most recent learning date is the current day; and the visual mapping value is determined based on the learning depth mapping value and the mastery parameter in the learning memory.

[0042] Through the above methods, the transformation from raw learning behavior data to visualized node status information was completed, realizing a quantitative representation of the user's mastery of knowledge nodes or learning progress, and providing accurate data support for dynamically adjusting the visual presentation attributes of map nodes in the display interface.

[0043] Furthermore, when a user enters the Star Map interface or a preset learning cycle begins, a personalized review recommendation generation process is triggered. This generation process uses Star Map nodes as the basic unit, comprehensively calculating node priority scores based on multiple factors. Specifically, it reads the node status information and learning memory data of each map node. These multiple factors include at least the node's illuminated status, mastery score, stability score, forgetting risk value, error density, and error concentration.

[0044] The system calculates weighted averages for each factor based on a pre-defined weighting model, generating node priority scores. For example, nodes with low mastery and high risk of forgetting are assigned higher weights, as are nodes with high error density and concentrated error causes. The system then uses these node priority scores to identify a set of weak nodes that require priority review, thus creating a review path.

[0045] Subsequently, for the set of weak points, related incorrect question data is retrieved from the incorrect question database to generate a set of recommended incorrect questions. This set is sorted by error type or difficulty level. Finally, the generated review priority, recommended incorrect question set, and review path are encapsulated into a recommendation information object for subsequent rendering module calls and mapping to corresponding map nodes in the visualized knowledge map.

[0046] S500: Render the visualized knowledge map in the display interface, and dynamically adjust the visual presentation attributes of the map nodes according to the node status information, so as to uniformly present the knowledge structure and learning status in the visualized knowledge map.

[0047] Specifically, after receiving the knowledge structure data and the calculated node status information, the client starts the rendering engine to draw the visualized knowledge map on the display interface. The rendering process includes two stages: layout calculation and visual mapping, which aims to transform abstract node status information into an intuitive spatial representation through dynamically changing visual presentation attributes.

[0048] First, during the layout calculation phase, the client initializes the physics simulation engine to perform force-guided layout calculations on the map nodes in the visualized knowledge map. Specifically, the physics simulation engine applies various constraint force fields in the star map space: including repulsive forces between nodes to maintain their dispersed distribution, with the root node's repulsive strength set to the maximum to keep the overall structure centered, and other nodes using moderate repulsive strengths; connection pulls to maintain hierarchical relationships, with upper-level nodes having longer connection distances to expand main branches, and lower-level nodes having connection distances dynamically adjusted according to the node's text length; collision detection forces to reduce label overlap, setting a protection radius for each map node, with the leaf node's protection radius dynamically calculated based on its text length; and a centripetal force to stabilize the overall structure at the viewport center. After each frame iteration, the physics simulation engine updates the node coordinates and re-renders the nodes and connection relationships to form a stable and interactive star map layout.

[0049] Secondly, during the visual mapping phase, the visual presentation attributes of the map nodes are dynamically adjusted based on the node state information. Specific mapping rules include: First, the root node is always rendered as a large node with a glowing effect, used to represent the core knowledge of the subject being studied. Secondly, each top-level category node is assigned a preset theme color, and its subordinate nodes inherit this theme color to form clear visual partitions; Third, for map nodes that are indicated as lit in the node status information (i.e., the cumulative learning amount is greater than zero), the corresponding theme color is used to fill them, and the brightness or glow intensity is adjusted according to the learning depth mapping value in the node status information. The learning depth mapping value is calculated using a non-linear growth function to avoid achieving the maximum display effect with only a small amount of learning. Fourth, for map nodes whose status information indicates they are not lit (i.e., the cumulative learning amount is zero), only the outline of the border is drawn and the display is weakened; Fifth, for map nodes whose node status information includes the learning mark for the day, a prominent learning mark for the day is superimposed on the outside of the map node to visually distinguish between the immediate learning results and the historical cumulative status.

[0050] In addition, the display interface includes a learning statistics overview card and interactive controls in a fixed area. The learning statistics overview card is used to summarize the overall learning status in the visualized knowledge map, including at least statistical indicators such as the number of lit nodes, cumulative learning volume, and daily learning volume, and is consistent with the status of map nodes. The interactive controls include zoom in, zoom out, and adaptive view buttons. The client listens for window size change events and dynamically adjusts the canvas size and re-executes physical simulation calculations when the viewport size changes, to ensure the stability and readability of the star map display.

[0051] Through the above methods, the rendering of the visualized knowledge map was completed in the display interface. The visual presentation attributes such as color, brightness, glow, outline and markers of the map nodes were dynamically adjusted according to the node status information. This achieved a unified presentation of knowledge structure and learning status on the same spatial carrier, enabling learners to intuitively perceive the integration of knowledge context and their own learning progress.

[0052] After a user completes the review task corresponding to the recommended review information, review result data is collected. The review result data includes at least information such as whether the error correction was successful, the review time, the type of error caused by a subsequent error, and whether the user has re-entered the review pool.

[0053] The review results data are written back to the learning memory and recommendation module. Specifically, if the user successfully corrects an error, the mastery and stability parameters of the corresponding knowledge node are dynamically increased, and the forgetting risk value is reduced; if the user makes another mistake, the mastery parameter is reduced, the error cause profile is updated, and the node is marked as a high-frequency error-prone node.

[0054] The recommendation strategy is adaptively optimized based on the adjusted parameters. Specific strategies include: reducing the weight of well-established nodes to decrease their frequency of appearance in subsequent recommendations; and increasing the weight of repeatedly incorrect nodes, raising their recommendation priority and adjusting the difficulty or depth of the incorrect question recommendation set. Through this closed-loop feedback mechanism, the recommendation strategy is iteratively optimized, ensuring that subsequent recommendations better align with the learner's actual learning trajectory.

[0055] refer to Figure 2 This embodiment further discloses a learning state tracking system based on a visualized knowledge map, including: The structure acquisition module 21 is configured to acquire the hierarchical knowledge structure of the learning content. This hierarchical knowledge structure includes multiple knowledge nodes and hierarchical relationships between nodes. Specifically, the structure acquisition module corresponds to some functions of the knowledge structure configuration and loading module and the knowledge graph conversion module in the technical disclosure document. It is configured to read the knowledge point classification system of the learning subject from the structured configuration file, parse and generate hierarchical structure data with the learning subject as the root node, and generate a mapping table between node identifiers and knowledge paths. This mapping table is used to subsequently and uniformly associate user learning data, incorrect question data, and learning memory data, ensuring that learning behavior data can be accurately mapped to specific knowledge nodes in the hierarchical knowledge structure.

[0056] The map generation module 22 is configured to generate a visual knowledge map based on the hierarchical knowledge structure. The visual knowledge map includes map nodes corresponding to the knowledge nodes and connection relationships representing the hierarchical relationships. Specifically, the map generation module corresponds to some functions of the knowledge graph conversion module and the knowledge map layout and rendering module in the technical disclosure document. It is configured to convert the hierarchical knowledge structure into the node array and connection relationship array required for a star-map-style knowledge map, and initialize the physics simulation engine to perform layout calculations for the map nodes. The layout calculation employs a force-directed layout algorithm, applying various constraint force fields such as repulsive forces between nodes, connection tension forces, collision detection forces, and centripetal forces. The node coordinates are updated after each frame iteration to form a stable and interactive star-map layout.

[0057] The data acquisition module 23 is configured to acquire user learning behavior data for the knowledge nodes, wherein the learning behavior data is mapped to the map nodes. Specifically, the data acquisition module corresponds to the integrated functions of the user learning statistics storage module, learning statistics update module, incorrect question collection and archiving module, incorrect question classification and error cause identification module, and learning memory storage module in the technical disclosure document. It is configured to: maintain user learning behavior statistics records on each knowledge path, recording the daily learning volume and historical cumulative learning volume respectively; collect incorrect question information and bind it to the corresponding knowledge path for storage, generating multi-dimensional structured labels including knowledge path labels, error cause type labels, and difficulty labels; maintain a learning memory data structure that evolves over time, wherein the memory includes at least node mastery, stability, forgetting risk, and error cause profile parameters. The data acquisition module aggregates the aforementioned scattered data into a learning behavior data set that is mapped to the map nodes, using the knowledge path as a unique index key.

[0058] The state calculation module 24 is configured to calculate the node state information of the map nodes based on the learning behavior data. This node state information characterizes the user's mastery level or learning progress of the corresponding knowledge nodes. Specifically, the state calculation module corresponds to the functions of the node state calculation module and the memory update and evolution module in the technical disclosure document. It is configured to use a hierarchical calculation method to perform state calculations on leaf nodes and upper-level classification nodes separately. Specifically, this includes: calculating a learning depth mapping value based on cumulative learning volume, calculating a mastery level representation value based on learning memory parameters, and determining the learning tag for the current day based on the most recent learning date. The state calculation module also supports summarizing the learning memory states of leaf nodes upwards to generate a memory summary of upper-level classification nodes for overall state evaluation.

[0059] The rendering and display module 25 is configured to render the visualized knowledge map in the display interface and dynamically adjust the visual presentation attributes of the map nodes according to the node status information, so as to uniformly present the knowledge structure and learning status in the visualized knowledge map. Specifically, the rendering and display module corresponds to some functions of the client star map display and interaction module and the knowledge map layout and rendering module in the technical disclosure document. It is configured to: uniformly map the visual effects according to the node's category and node status, including the root node glow effect, category theme color inheritance, and color filling of the node's lit state; adjust the node brightness or glow intensity according to the learning depth mapping value, and use a non-linear growth function to avoid achieving the maximum display effect with a small amount of learning; overlay the learning mark of the day on the outside of the node with the learning behavior of the day, so as to visually distinguish the immediate learning results from the historical cumulative status; provide interactive controls such as zoom, pan, drag and adaptive view, listen for window size change events and dynamically adjust the canvas size.

[0060] To further support personalized review and adaptive optimization, the system may also include the following extended modules: The personalized review recommendation module 26 is configured to generate personalized review recommendation information based on the node status information and the learning memory. It is configured to calculate a node priority score by comprehensively considering factors such as node activation status, mastery score, forgetting risk value, and error density, filter out a set of weak nodes that require priority review, generate a review path and a set of recommended errors, and map the review recommendation information to corresponding map nodes in the visualized knowledge map. This visualization is achieved through highlighting, recommendation halos, or navigation prompts, aligning the recommendation results with the knowledge space.

[0061] The feedback optimization module 27 is configured to record the results after the user completes the review task corresponding to the review recommendation information, and write the results back to the learning memory. It is configured to dynamically adjust the mastery, stability, and forgetting risk of corresponding knowledge nodes based on the review results (such as whether error correction was successful, time spent, and the type of error caused by repeated mistakes). Based on the adjusted parameters, the recommendation strategy is adaptively optimized, with nodes that have been consolidated receiving lower weights and nodes that repeatedly make mistakes receiving higher weights, thus achieving iterative optimization of the recommendation strategy.

[0062] Figure 3 A schematic diagram of the physical structure of an electronic device provided in an embodiment of the present invention, such as... Figure 3 As shown, the electronic device 50 includes: a processor 501, a memory 502, and a bus 503; The processor 501 and the memory 502 communicate with each other via the bus 503; the processor 501 is used to call the program instructions in the memory 502 to execute the methods provided in the above-described embodiments.

[0063] This embodiment provides a non-transitory computer-readable storage medium that stores computer instructions that cause a computer to execute the methods provided in the above-described embodiments.

[0064] Those skilled in the art will understand that all or part of the steps of the above-described method implementation can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above-described method implementation. The aforementioned storage medium includes various storage media capable of storing program code, such as ROM, RAM, magnetic disk, or optical disk.

[0065] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0066] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of each embodiment or some parts of the embodiments.

[0067] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A learning state tracking method based on a visualized knowledge map, characterized in that, include: Obtain the hierarchical knowledge structure of the learning content, wherein the hierarchical knowledge structure includes multiple knowledge nodes and the hierarchical relationship between the nodes; A visual knowledge map is generated based on the hierarchical knowledge structure. The visual knowledge map includes map nodes corresponding to the knowledge nodes and connection relationships representing the hierarchical relationships. Acquire user learning behavior data for the knowledge nodes, and the learning behavior data is mapped to the map nodes; The node status information of the map node is calculated based on the learning behavior data. The node status information is used to characterize the user's mastery or learning progress of the corresponding knowledge node. The visualized knowledge map is rendered in the display interface, and the visual presentation attributes of the map nodes are dynamically adjusted according to the node status information, so as to uniformly present the knowledge structure and learning status in the visualized knowledge map.

2. The learning state tracking method according to claim 1, characterized by, The generation of a visual knowledge map based on the hierarchical knowledge structure includes: The hierarchical knowledge structure is converted into a star-map knowledge map, and the map nodes in the star-map knowledge map are used to calculate their positions using a force-directed layout algorithm. The force-guided layout algorithm applies a variety of constraint force fields, including inter-node repulsive force to maintain the dispersed distribution of nodes, connection tension to maintain hierarchical relationships, collision detection force to reduce label overlap, and central centripetal force to stabilize the overall structure at the center of the viewport. After each frame iteration, the node coordinates are updated and the nodes and connections are re-rendered to form a stable and interactive star map layout.

3. The learning state tracking method according to claim 1, characterized in that, The acquisition of user learning behavior data for the knowledge nodes includes: Maintain the user's daily learning volume and historical cumulative learning volume for the knowledge nodes respectively; When a user is detected to be engaging in learning activities, it is determined whether the most recent learning date is the current day. If it is the current day, the learning amount for that day is accumulated; if it is not the current day, the learning amount for that day is reset and the most recent learning date is updated, while the historical cumulative learning amount is accumulated.

4. The learning state tracking method according to claim 1, characterized in that, The learning behavior data includes incorrect answer data, and the method further includes: The incorrect question data is bound and stored with the corresponding knowledge nodes. The incorrect question data includes at least the question reference, answer information, correctness result and answer time. The incorrect question data is used to generate multidimensional structured tags, which include at least knowledge path tags, error type tags, and difficulty tags. In the visualized knowledge map, the corresponding map nodes are aggregated and positioned according to the distribution density of the incorrect question data to form a spatialized error distribution view.

5. The learning state tracking method according to claim 1, characterized in that, The calculation of the node state information of the map node based on the learning behavior data includes: Construct a learning memory that evolves across time, wherein the learning memory includes at least node mastery, stability, forgetting risk, and error profile parameters; When learning, answering, or reviewing events occur, the parameters in the learning memory are updated and their evolution is calculated. It supports summarizing the learning memory states of leaf nodes upwards to generate summary states of upper-level classification nodes for overall state evaluation.

6. The learning state tracking method according to claim 5, characterized in that, The method further includes: Based on the node status information and the learning memory, personalized review recommendation information is generated, including review priority, a set of recommended incorrect questions, and a review path; The review recommendation information is mapped to the corresponding map nodes in the visualized knowledge map, and then visualized through highlighting, recommendation halo, or navigation prompts to align the recommendation results with the knowledge space.

7. The learning state tracking method according to claim 6, characterized in that, The method further includes: Record the results after the user completes the review task corresponding to the review recommendation information. The results include at least whether the error correction was successful, the time taken, and the type of error caused by the error if the error recurs. The results are written back to the learning memory to dynamically adjust the mastery, stability, and forgetting risk of the corresponding knowledge nodes; The recommendation strategy is adaptively optimized based on the adjusted parameters, with nodes that have been consolidated being downweighted and nodes that repeatedly make mistakes being upweighted.

8. The learning state tracking method according to claim 3, characterized in that, The step of dynamically adjusting the visual presentation attributes of the map nodes based on the node status information includes: A learning marker for the day is overlaid on the outside of the map node to visually distinguish immediate learning outcomes from historical cumulative status.

9. A learning state tracking system based on a visual knowledge map, characterized in that, include: The structure acquisition module is used to acquire the hierarchical knowledge structure of the learning content, wherein the hierarchical knowledge structure includes multiple knowledge nodes and the hierarchical relationship between the nodes; A map generation module is used to generate a visual knowledge map based on the hierarchical knowledge structure. The visual knowledge map includes map nodes corresponding to the knowledge nodes and connection relationships representing the hierarchical relationships. The data acquisition module is used to acquire user learning behavior data for the knowledge nodes, and the learning behavior data is mapped to the map nodes; The status calculation module is used to calculate the node status information of the map node based on the learning behavior data. The node status information is used to characterize the user's mastery or learning progress of the corresponding knowledge node. The rendering and display module is used to render the visualized knowledge map in the display interface and dynamically adjust the visual presentation attributes of the map nodes according to the node status information, so as to uniformly present the knowledge structure and learning status in the visualized knowledge map.

10. A computer-readable storage medium, characterized in that, The storage medium stores a computer program that, when executed by a processor, implements the method of any one of claims 1-8.

Images