User portrait method and device for driving training learning, electronic equipment and medium
By collecting and processing multi-source heterogeneous data, extracting two-dimensional features and constructing dynamic student profiles, the problems of insufficient accuracy and untimely updates in existing technologies are solved, enabling precise delivery of personalized learning resources and teaching intervention.
Patent Information
- Application Number
- CN202511963145.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-24
- Publication Date
- 2026-03-27
AI Technical Summary
Existing methods for profiling driving trainees suffer from inconsistent data structures from multiple sources, limited evaluation dimensions, and restricted server resources, resulting in inaccurate and untimely profiles that fail to meet personalized learning needs.
Collect heterogeneous data from multiple sources, generate student data profiles through data preprocessing, extract two-dimensional features, including learning stage features and weak subject features, construct student profiles containing learning stage tags and weak subject tags, and adopt a dynamic iterative update mechanism to ensure the accuracy and timeliness of the profiles.
It improves the accuracy of student profiles, supports the precision of personalized learning resource delivery, precise teaching intervention, and large-scale operational decision-making, and ensures the integrity and consistency of student data files.
Smart Images

Figure CN121743543A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of information recommendation technology, specifically to a user profiling method, device, electronic device, and medium for driver training. Background Technology
[0002] With the accelerated digital transformation of the motor vehicle driver training industry, traditional large-scale and standardized teaching management models are increasingly unable to meet the personalized learning needs of students and the management goals of refined driving school operations. Against this backdrop, student profiling, as a data model capable of structurally describing students' learning status, ability characteristics, and personal attributes, is becoming increasingly important. This model provides core data-driven decision support for personalized learning resource recommendations, precise teaching interventions, and the optimized allocation of training resources.
[0003] However, existing technical solutions related to driver training student assessment or profile building still have several obvious limitations and defects in practical applications. Current student profiling methods, limited by data sources and formats, often assess students based on a single, isolated dimension, leading to a one-sided and inaccurate identification of students' skill gaps. Furthermore, most existing solutions generate static snapshots of student profiles, which, due to limitations in server resources, cannot be updated in a timely manner to reflect students' daily learning behaviors, thus failing to accurately reflect the students' current status.
[0004] In summary, the existing methods suffer from technical defects such as inaccurate student profiles and untimely updates due to inconsistent data structures from multiple sources, a single evaluation dimension, and limited server resources. Summary of the Invention
[0005] In view of this, it is necessary to provide a user profiling method, device, electronic device and storage medium for driver training learning, in order to solve the technical problem that the existing methods are not accurate enough in terms of student profiles due to inconsistent data structures from multiple sources, single evaluation dimensions and limited server resources.
[0006] To address the aforementioned technical problems, in a first aspect, the present invention provides a user profiling method for driver training, comprising: Collect multi-source heterogeneous data of trainees and perform data preprocessing to obtain trainee data profiles. The multi-source heterogeneous data includes learning stage data, ability assessment data and basic attribute data. Based on the student data profile, the student's two-dimensional features are extracted, including: learning stage features are obtained by dividing the driving test subject pass status and learning progress completion rate, and weak subject features are obtained by quantifying the mastery of sub-test points under weak subjects based on a weighted scoring algorithm. By integrating the aforementioned dual-dimensional features and basic attribute data, a student profile is constructed that includes learning stage tags and weak subject tags.
[0007] In one possible implementation, the process of collecting multi-source heterogeneous data from trainees and performing data preprocessing to obtain a trainee data profile includes: Raw data is obtained from multiple independent driver training business systems through application programming interface calls or message queue synchronization. The raw data is cleaned, standardized, and linked and merged based on student identifiers to generate the student data profile.
[0008] In one possible implementation, the step of dividing the learning stage characteristics based on the driving test subject pass status and learning progress completion rate includes: Based on the subject information that the student has passed, the core learning stage is determined. The core learning stage includes the period before learning begins and the learning period for subjects one through four. Within the core learning phase, the progress is subdivided based on the completion rate of the current subject. The progress subdivision includes the introductory phase, the improvement phase, the sprint phase, and the retake preparation phase.
[0009] In one possible implementation, the quantification of mastery of specific test points under weak subjects based on a weighted scoring algorithm includes: Obtain the indicator values of the subdivided test points on multiple evaluation dimensions, including the error rate, the number of times the same type of question is repeatedly wrong, the percentage of points lost in the mock exam, and the redundancy of practice time. Based on the preset weight allocation, the index values on the multiple evaluation dimensions are weighted and calculated to obtain the weakness coefficient that represents the degree of mastery of the test points.
[0010] In one possible implementation, the process of fusing the dual-dimensional features and basic attribute data to construct a student profile including learning stage tags and weak subject tags includes: The learning stage characteristics, weak subject characteristics, and basic attribute data are used to generate a student profile containing individual student tags. Based on the individual labels, a clustering algorithm is used to classify trainees into subgroups with different label combinations, and these subgroups constitute the group profile part of the trainee profile.
[0011] In one possible implementation, the user profiling method for driver training further includes dynamically iteratively updating the student profile, wherein the dynamic iterative updating of the student profile includes: The learning stage tags and weak subject tags are updated based on a dual-mode update mechanism, which includes event-triggered and timed-triggered updates.
[0012] In one possible implementation, updating the learning stage label and the weak subject label based on the dual-mode update mechanism includes: In response to events such as students completing exercises or mock exams, the weakness coefficients of relevant test points are recalculated in real time, and the characteristics of the weak subjects are updated. According to a preset cycle, the latest progress data of all students is collected and processed periodically, and the learning stage characteristics are updated.
[0013] On the other hand, the present invention also provides a user profiling device for driver training, comprising: The data acquisition module is used to collect multi-source heterogeneous data of trainees and perform data preprocessing to obtain trainee data files. The multi-source heterogeneous data includes learning stage data, ability assessment data, and basic attribute data. The feature extraction module is used to extract two-dimensional features of students based on the student data files, including: learning stage features based on the pass status and learning progress completion rate of driving test subjects, and quantifying the mastery of sub-test points under weak subjects based on a weighted scoring algorithm to obtain weak subject features. The user profile module is used to integrate the dual-dimensional features and basic attribute data to construct a student profile that includes learning stage tags and weak subject tags.
[0014] Thirdly, the present invention also provides an electronic device, including a memory and a processor, wherein, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the user profiling method for driver training described in any of the above implementations.
[0015] Fourthly, the present invention also provides a computer-readable storage medium for storing a computer-readable program or instructions, which, when executed by a processor, can implement the steps in the user profiling method for driver training described in any of the above implementations.
[0016] The beneficial effects of this invention are as follows: The user profiling method for driver training provided by this invention obtains student data profiles by preprocessing heterogeneous data from multiple sources, ensuring the data integrity and consistency of the obtained student data profiles, which is conducive to improving the accuracy of subsequent feature extraction; furthermore, a two-dimensional feature extraction framework is used to extract learning stage features and weak subject features, thereby generating student profiles, overcoming the shortcomings of the single dimension of existing technologies, improving the accuracy of profiles, and at the same time, it is conducive to improving the accuracy of personalized learning resource push, precise teaching intervention, and large-scale operation decision-making. Attached Figure Description
[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 This is a schematic flowchart of an embodiment of the user profiling method for driver training provided by the present invention; Figure 2 For the present invention Figure 1 A schematic diagram of an embodiment of S101; Figure 3 For the present invention Figure 1 A schematic diagram of an embodiment of S102; Figure 4 For the present invention Figure 1 A schematic diagram of another embodiment of S102; Figure 5 For the present invention Figure 1 A schematic diagram of an embodiment of S103; Figure 6 This is a flowchart illustrating another embodiment of the user profiling method for driver training provided by the present invention. Figure 7 For the present invention Figure 6 A schematic diagram of an embodiment of S601; Figure 8 This is a schematic diagram of an embodiment of the user profiling device for driver training provided by the present invention; Figure 9 A schematic diagram of an embodiment of the electronic device provided by the present invention. Detailed Implementation
[0019] 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 a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.
[0020] In the description of the embodiments of the present invention, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.
[0021] The terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature.
[0022] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0023] This invention provides a user profiling method, apparatus, electronic device, and storage medium for driver training. The technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.
[0024] Figure 1 This is a schematic flowchart of an embodiment of the user profiling method for driver training provided by the present invention, as shown below. Figure 1 As shown, user profiling methods used for driver training include: S101. Collect multi-source heterogeneous data of trainees and perform data preprocessing to obtain trainee data profiles. The multi-source heterogeneous data includes learning stage data, ability assessment data, and basic attribute data.
[0025] The multi-source heterogeneous data includes learning stage data, competency assessment data, and basic attribute data. Specifically, it is collected from data sources such as student registration systems, online theoretical learning platforms, mock exam software, and compliant timed training terminals through a combination of application programming interface calls and message queue synchronization.
[0026] Among them, the learning stage data is used to record the student's learning progress, including but not limited to the enrolled subjects, subject pass status, and class hour progress; Among them, the ability assessment data is used to reflect their knowledge mastery and operational skills, including but not limited to mock exam scores, chapter exercise accuracy, error details, and operation trajectory and mistakes in Subject 2 / 3. Among them, basic attribute data is used to describe the inherent attributes of trainees, such as age and driving test type C1 / C2.
[0027] In this embodiment, a tiered synchronization strategy is adopted to address the differences in data real-time performance: events such as practice submissions and exam results are synchronized in seconds via a Kafka message queue; near real-time data such as study time is synchronized in batches every 5 minutes; and registration information is synchronized every day at midnight.
[0028] In the data preprocessing stage, outliers are identified and removed using techniques such as box plots. Missing values are filled using methods combined with business rules, and indicators with different dimensions are standardized using methods such as Min-Max normalization. Finally, standardized and structurally consistent student data profiles are generated using data fusion technology, with student identifiers as the key.
[0029] S102. Based on the student data profile, extract the student's two-dimensional features, including: learning stage features based on the pass status of driving test subjects and the completion rate of learning progress; and quantifying the mastery of subdivided test points under weak subjects based on a weighted scoring algorithm to obtain weak subject features.
[0030] Specifically, step S102 performs two-dimensional feature extraction. This step extracts two core feature dimensions representing the learner's status from the unified archive: learning stage features and weak subject features.
[0031] In a preferred embodiment, the learning stage characteristics are generated through a two-level judgment model. The first level divides students into one of five core stages based on the official driving test subject pass status: "not started learning", "subject one learning period", "subject two learning period", "subject three learning period" or "subject four learning period". The second level further subdivides each core stage into four states based on the current subject learning progress completion rate (such as the percentage of completed class hours): "entry stage" (0-30%), "improvement stage" (30-70%), "sprint stage" (70-100%) or "retake preparation stage".
[0032] It should be understood that the characteristics of the learning stage mainly come from the learning stage data and the basic attribute data.
[0033] In a preferred embodiment, the characteristics of weak subjects are generated through a multi-level quantification model. First, weak subjects are identified based on the overall accuracy rate and practice time. Then, within each subject, specific sub-test points are identified based on the distribution of incorrect answers (e.g., "Subject Two - Reversing into a Parking Space"). Finally, a weakness coefficient is calculated by integrating data from all levels of indicators. The higher the coefficient, the weaker the mastery. For example, if a student's weakness coefficient for the reversing into a parking space test point is 0.72, it is quantified as a significant weakness for that student.
[0034] It should be understood that the characteristics of the learning stage are mainly derived from ability assessment data and basic attribute data.
[0035] S103. Integrate dual-dimensional features and basic attribute data to construct a student profile that includes learning stage tags and weak subject tags.
[0036] Step S103 integrates the above features with basic attributes, transforming them into a directly applicable tag system. Specifically, it maps quantified feature values to semantic tags.
[0037] For example, learning stage characteristics may be mapped to the label "Improvement period for Subject 2", and test points with a weakness coefficient exceeding 0.6 may be mapped to the label "Weakness in reversing into a parking space". These labels, together with basic attribute labels such as "C1 driver's license" and "25-30 years old", constitute a multi-dimensional and structured individual profile of the learner.
[0038] To further uncover group patterns, this embodiment also vectorizes the tag data of all trainees and processes it using clustering algorithms such as K-Means to automatically divide trainee groups with similar feature combinations (such as "the group that is in the improvement period of driving test 2 and is weak in reversing parking"), forming group profiles to support differentiated allocation and operation of teaching resources.
[0039] As a preferred embodiment of this invention, when acquiring multi-source heterogeneous data, this embodiment also extracts data containing time points and data related to user learning habits from the multi-source heterogeneous data, thereby analyzing user preferences and habits based on the data related to time points and user learning habits, and setting user preference tags based on the obtained user preferences and habits, and integrating them with dual-dimensional features and basic attribute data to construct a student profile.
[0040] The time points include, but are not limited to, morning, noon, afternoon, evening, and late at night. By analyzing the time points, we can find out the times when students usually spend their time studying for driving tests. In subsequent recommendations, we can push driving test study services for the corresponding time periods based on the user's preferences.
[0041] Among them, data related to students' learning habits include, but are not limited to, the order in which users learn knowledge points and their preferred learning methods (such as watching videos before doing questions for subjects one and four, or doing questions before reviewing knowledge points, etc.). Based on this data, students' habits are analyzed to develop personalized learning service plans.
[0042] This embodiment obtains student data profiles by preprocessing heterogeneous data from multiple sources, ensuring the data integrity and consistency of the obtained student data profiles, which is conducive to improving the accuracy of subsequent feature extraction. Then, a two-dimensional feature extraction framework is used to extract learning stage features and weak subject features, thereby generating student profiles. This overcomes the shortcomings of the single dimension of existing technologies, improves the accuracy of profiles, and is conducive to improving the accuracy of personalized learning resource delivery, precise teaching intervention, and large-scale operation decision-making.
[0043] In some embodiments of the present invention, such as Figure 2 As shown, step S101 involves collecting multi-source heterogeneous data from trainees and performing data preprocessing to obtain trainee data files, including: S201. Obtain raw data from multiple independent driver training business systems through application programming interface calls or message queue synchronization; S202. Clean, standardize, and integrate the raw data based on student identifiers to generate student data profiles.
[0044] First, raw data is obtained from multiple independently deployed driver training business systems through application programming interface calls or message queue synchronization.
[0045] Specifically, a tiered synchronization approach is adopted for data with different timeliness requirements. For data that requires immediate response, such as practice submission records and mock exam scores, synchronization is achieved in seconds through message queues such as Kafka and is set as the highest priority. For near real-time data such as learning duration and login frequency, batch synchronization is scheduled to be performed once every 5 minutes. For non-real-time data with low change frequency, such as registration information and driving school qualifications, batch synchronization is scheduled at a fixed time every day (such as early morning).
[0046] It should be understood that the hierarchical synchronization method used in this embodiment is used to balance data timeliness and processing load, which is beneficial to improving the timeliness of data collection and processing as well as the efficiency of server resource utilization.
[0047] Subsequently, a preprocessing procedure is performed on the acquired raw data.
[0048] During the data cleaning phase, outlier detection algorithms are employed to identify and remove invalid records. These algorithms include, but are not limited to, box plots based on statistical distribution, standard deviation-based methods (such as the 3σ criterion), density-based clustering algorithms (Density-Based Spatial Clustering of Applications with Noise, DBSCAN), and model-based methods (such as Isolation Forest). These algorithms can be selected or combined based on the different distribution characteristics of the data and the needs of the business scenario to effectively identify and remove outlier data points caused by system errors, recording mistakes, or atypical behavior, thereby ensuring the accuracy and reliability of the dataset upon which subsequent data analysis relies.
[0049] For example, regarding the single practice session duration metric, by calculating the interquartile range of the data, records with values greater than the upper quartile plus three times the interquartile range (such as obvious erroneous data where the single practice session duration exceeds 300 minutes) are identified as outliers and removed. For data with missing values, the mean is used for imputation, combined with business rules for completion; for example, when a student's learning stage data is missing, the current stage can be inferred and completed based on their registered subject and recent practice records, using preset business rules.
[0050] The data standardization phase aims to transform heterogeneous data from different systems into a unified format and scale. For example, the inconsistent formats of "error records" from various practice systems are uniformly decomposed and reconstructed into a standard structure containing fields such as "question ID, test point category, number of errors, and error timestamp." Simultaneously, for continuous numerical indicators such as practice duration and answer accuracy, a standardized approach is adopted. Min-Max The normalization algorithm scales the value to the interval [0, 1]. The formula for this algorithm is: X norm =(XX min ) / (X max -X min ) ,in X The original value, X min and X max These are the minimum and maximum values for this indicator for trainees at the same stage, respectively, to eliminate the influence of different units of measurement.
[0051] Finally, in the data fusion stage, using student identifiers (such as student IDs) as the association key, cleaned and standardized data from various business systems such as registration, practice, and examination are integrated through data association technology to form a complete and consistent standardized student data archive.
[0052] The hierarchical synchronization strategy employed in this embodiment ensures rapid flow of highly time-sensitive data while avoiding instantaneous pressure on the system caused by non-real-time data, thus improving data acquisition efficiency and system resource utilization. Simultaneously, the cleaning and standardization processes effectively enhance the quality and consistency of the raw data, which is beneficial for improving the accuracy of subsequent vital sign extraction.
[0053] In some embodiments of the present invention, such as Figure 3 As shown, step S102 divides the learning stage characteristics based on the driving test subject pass status and learning progress completion rate, including: S301. Based on the subject information that the student has passed, determine the core learning stage to which the student belongs. The core learning stage includes the period before learning begins and the learning period for subjects one through four. S302. During the core learning phase, based on the completion rate of the current subject's learning progress, determine the progress sub-states, which include the introductory phase, the improvement phase, the sprint phase, and the retake preparation phase.
[0054] First, based on the student's official passing information in the vehicle management office's examination system, their core learning stage is determined. This method follows the standard procedures and phased characteristics of the driving test process, dividing the overall learning journey into five core stages: no start of learning, Subject 1 learning period, Subject 2 learning period, Subject 3 learning period, and Subject 4 learning period. For example, if a student has successfully passed Subject 1 but has not yet passed Subject 2, their core learning stage is determined to be the "Subject 2 learning period."
[0055] Subsequently, within the defined current core learning phase, the progress status of each student is further determined based on their completion rate within that phase.
[0056] The progress completion rate is calculated as the proportion of the cumulative completed learning content (such as class hours, number of practice questions, or number of mock exams) to the total preset learning content for that stage. Further, based on the range of this proportion, the progress is subdivided into the introductory stage, the improvement stage, the sprint stage, and the retake preparation stage.
[0057] In a specific example, when the progress completion rate is between 0% and 30%, it is the introductory phase; between 30% and 70%, it is the improvement phase; and when it exceeds 70%, it is the sprint phase.
[0058] In this embodiment, the retake preparation period is a special state. When a student fails the official exam in the current core stage, regardless of their historical practice progress, their progress status is updated to "retake preparation period" until they pass the next exam. For example, a student in the "Subject 2 learning period" whose Subject 2 related class hours and practice completion rate reaches 52% will have their progress status changed to "Subject 2 improvement period"; if the student takes the Subject 2 exam but fails, their status immediately changes to "Subject 2 retake preparation period".
[0059] It should be noted that, in order to further accurately characterize the learning dynamics, this embodiment also introduces the progress rate as a supplementary quantitative indicator. This supplementary quantitative indicator is determined by the calculation method of "(current progress percentage - initial progress percentage) / number of days spent in this stage". It is used to objectively identify whether the learning has fallen into a stagnation state. When the supplementary quantitative indicator is less than the preset progress measurement threshold, it is determined that the learning has fallen into a stagnation state, and early warning intervention will be carried out by reminding.
[0060] As a preferred embodiment, the learning stage characteristics can be specifically represented by three-dimensional labels or three-dimensional vectorization of three indicators: "core stage, progress completion rate, and progress advancement rate". For example, "Subject 2 improvement period, progress 52%, advancement rate is normal".
[0061] This embodiment extends the stage determination from a simple "pass or fail" binary state to a two-level division of core stages and subdivided stages, thereby enabling more precise positioning of a student's progress and status over a longer learning cycle. Furthermore, by introducing quantifiable progress completion rates and progress rates, the previously vague learning progress is transformed into measurable and comparable objective indicators, making the evaluation and comparison of learning efficiency among different students more accurate. The resulting learning stage characteristics (e.g., "Subject 2 improvement period, progress 52%, normal progress rate") provide a more accurate and dynamic decision-making basis for subsequent applications such as personalized learning resource delivery and risk student warnings.
[0062] In some embodiments of the present invention, such as Figure 4 As shown, step S102 quantifies the mastery of specific test points under weak subjects based on a weighted scoring algorithm, including: S401. Obtain the indicator values of the subdivided test points on multiple evaluation dimensions. The evaluation dimensions include the error rate, the number of repeated errors on the same type of question, the percentage of points lost in the mock exam, and the redundancy of practice time. S402. Based on the preset weight allocation, the indicator values on multiple evaluation dimensions are weighted and calculated to obtain the weakness coefficient representing the degree of mastery of the test points.
[0063] First, for the specific sub-test points identified, raw indicator values across multiple pre-set assessment dimensions are obtained from the student's data profile. These assessment dimensions are comprehensively selected to objectively reflect the student's mastery from different perspectives, including but not limited to: error rate, number of times similar questions are answered incorrectly repeatedly, percentage of points lost in mock exams, and redundancy in practice time.
[0064] Among them, the error rate refers to the percentage of incorrect questions in all exercises for that test point; Among them, the number of repeated errors on similar questions refers to the cumulative number of errors made on the same or highly similar questions; Among them, the percentage of points lost in mock exams refers to the proportion of points lost due to this test point in the total points lost in the official mock exams. Among them, practice time redundancy refers to the proportion of practice time spent on this test point that exceeds the average practice time of students at the same stage.
[0065] For example, for the test point "Subject 2 - Reversing into a parking space", all relevant data of students under this test point will be collected separately, and the specific values of the above four dimensions will be calculated.
[0066] Subsequently, the index values of the above dimensions are normalized and weighted according to the preset weight allocation, so as to synthesize a weak coefficient in a unified and continuous numerical form.
[0067] The preset weight allocation can be based on driving test teaching experience and data analysis. The specific weight allocation can be set according to the actual application needs, and there are no restrictions here.
[0068] In the preferred embodiment of this example, the preset weight allocation is as follows: the error rate accounts for 40% of the weight, as it directly reflects the general mastery level; the number of times the same type of question is repeatedly wrong accounts for 30% of the weight, which is used to highlight persistent errors; the proportion of points lost in mock exams accounts for 20% of the weight, which is used to emphasize the application ability in the exam environment; and the redundancy of practice time accounts for 10% of the weight, which is used to consider the learning efficiency factor.
[0069] Specifically, the weighted calculation can be expressed as: Weakness coefficient = Error rate × 0.4 + Number of repeated errors on similar questions × 0.3 + Percentage of points lost in mock exams × 0.2 + Redundancy of practice time × 0.1. Since different questions have different difficulties, the number of repeated errors on similar questions can be normalized in the actual calculation.
[0070] For example, a student's error rate on the "reverse parking" test was 45%, the normalized number of repeated errors was 0.3 (assuming a scale of 0-10), the mock exam error rate was 25%, and the practice time redundancy was 15%. Therefore, their weakness coefficient is calculated as: 0.45×0.4 + 0.3×0.3 + 0.25×0.2 + 0.15×0.1 = 0.335. This result is a value between 0 and 1; a higher value indicates a weaker grasp of the subject.
[0071] This embodiment overcomes the limitations of judging solely by the number of incorrect answers or a single test score by comprehensively considering four dimensions: error rate, repeated errors, test score loss, and practice efficiency. This allows for a more comprehensive assessment of students' weaknesses. Furthermore, by assigning differentiated weights to different dimensions, the calculated weakness coefficient more accurately and discriminatively quantifies students' true mastery levels. This provides a direct and reliable basis for personalized teaching interventions. For example, when the weakness coefficient for a particular test point exceeds a preset threshold of 0.6, it can be automatically identified as a "core weakness," thereby prioritizing the delivery of targeted practice resources to that student.
[0072] In some embodiments of the present invention, such as Figure 5 As shown, step S103 integrates dual-dimensional features and basic attribute data to construct a student profile that includes learning stage labels and weak subject labels, including: S501. Generate a student profile containing individual student tags by combining the learning stage characteristics, weak subject characteristics, and basic attribute data. S502. Based on individual labels, a clustering algorithm is used to classify trainees into subgroups with different label combinations. These subgroups constitute the group profile part of the trainee profile.
[0073] First, this embodiment integrates quantified learning stage characteristics, weak subject characteristics, and student's basic attribute data to generate a structured individual profile. This method maps the above features to specific tags, forming a multi-dimensional tag set to accurately describe the status of an individual student.
[0074] For example, a student's profile might be represented by: basic attribute tags (C1 driver's license, 25-30 years old), learning stage tags (improvement period for driving test part 2), weak subject tags (weakness in reversing into a parking space), and optional behavioral preference tags (evening study preference). The determination of the "weak subject" tag is directly related to whether its weakness coefficient exceeds a preset threshold of 0.6. In this way, abstract quantitative features are transformed into semantic tags that can be directly used for business logic judgment, thus achieving a precise individual characterization of the student.
[0075] Subsequently, based on the individual profiles, and using the aforementioned tagging system as the feature dimension, the tagging information of all trainees is transformed into machine-processable feature vectors. These feature vectors are then processed using a clustering algorithm such as K-Means. This clustering algorithm can automatically categorize a massive number of trainees into several subgroups with distinct common characteristics based on the similarity of trainees' stage tags and weakness tags. For example, the clustering results may generate group profiles such as "Driving Skills Improvement Period + Weakness in Reversing into a Parking Space" or "Driving Skills Final Preparation Period + Weakness in Lighting Operation."
[0076] This embodiment generates individual tags, completing the final transformation from multi-source heterogeneous data to a user model. This unifies the evaluation of students from scattered indicators into profiles with clear business meaning, providing objective data basis for subsequent personalized services for students. Simultaneously, generating group profiles based on clustering algorithms enables rapid identification of the distribution and common needs of different categories of students, thereby improving the accuracy of differentiated and batch-based development of teaching strategies, allocation of tutoring resources, and design of other personalized service recommendations.
[0077] In some embodiments of the present invention, such as Figure 6 As shown, the user profiling method for driver training includes dynamically iteratively updating the student profile. This dynamic iterative updating of the student profile includes: S601. Based on a dual-mode update mechanism, the learning stage tags and weak subject tags are updated. The dual-mode update mechanism includes event triggering and timed triggering.
[0078] In some embodiments of the present invention, such as Figure 7 As shown, based on the dual-mode update mechanism, updating the learning stage tags and weak subject tags includes: S701: In response to events such as students completing exercises or mock exams, the system recalculates the weakness coefficients of relevant test points in real time and updates the characteristics of weak subjects. S702. According to the preset cycle, collect and process the latest progress data of all students at regular intervals, and update the learning stage characteristics.
[0079] In this embodiment, the user profiling method for driver training includes dynamically iteratively updating the constructed student profiles to ensure they continuously reflect the students' latest learning status. This dynamic iteration function is implemented through a dual-mode update mechanism, specifically including an event-triggered real-time update mode and a timed batch update mode. The two work together to balance the system's real-time requirements and processing efficiency.
[0080] Specifically, the event-triggered pattern is a real-time update chain driven by key learning behaviors of learners. When a learner submits a practice exercise or completes a mock exam on an online practice platform, the event is instantly captured and an update message is generated via a message broker such as Kafka. The profile update service listens for such messages, and once a message is received, it triggers a specific computational task for that learner. This specific computational task extracts the detailed test point data associated with this behavior (e.g., the "reverse parking" test point to which the recently practiced question belongs), and recalculates the weakness coefficients of these test points in real time based on the latest wrong question records, error data, etc.
[0081] Subsequently, the newly calculated weakness coefficient is compared with a preset threshold (e.g., 0.6), and the label for the weak subject is adjusted in real time accordingly. For example, if a student's newly calculated weakness coefficient drops from 0.75 to 0.55 after completing a reverse parking practice session, the system will automatically remove the core weakness label for reverse parking from their profile. This process is usually completed within one minute of the event, achieving a minute-level response to changes in the student's skill gaps.
[0082] On the other hand, the timed trigger mode is a batch update task based on a fixed period, which is usually scheduled to be executed during periods of low system load (such as early morning each day). This task will start a batch processing job to periodically collect the latest learning progress data of all students in the past period (such as the past 24 hours), including the completion status of class hours, the cumulative practice time, and the latest subject pass status.
[0083] Based on this integrated data, the progress completion rate of each student in the current subject is reassessed, and the core learning stage and sub-level status are determined according to business rules to determine whether there have been changes, thereby completing the batch update of learning stage characteristics and corresponding tags.
[0084] For example, if the system detects that a student's cumulative practice progress for Subject 2 has reached 75% during the daily batch processing at midnight, it will automatically update their learning stage label from "Subject 2 Improvement Period" to "Subject 2 Sprint Period". This approach avoids the pressure on the business database caused by high-frequency real-time calculations, ensuring the stability and efficiency of data processing.
[0085] This embodiment, through the aforementioned dual-mode collaborative implementation method, effectively resolves the contradiction between data update timeliness and system computing resource consumption from a technical architecture perspective. This makes the final generated student profile a dynamic model that changes dynamically with the student's learning, ensuring that the profile data relied upon by the teaching system is always synchronized with the student's real status. This fundamentally avoids problems such as personalized recommendation failure and resource misallocation caused by profile lag, thereby effectively improving the accuracy and effectiveness of profile-driven services (such as precise push and risk warning).
[0086] To better implement the user profiling method for driver training in this embodiment of the invention, based on the user profiling method for driver training, correspondingly, as follows: Figure 8 As shown, this embodiment of the invention also provides a user profiling device for driver training, the user profiling device 800 for driver training includes: The data acquisition module 801 is used to collect multi-source heterogeneous data of trainees and perform data preprocessing to obtain trainee data profiles. The multi-source heterogeneous data includes learning stage data, ability assessment data, and basic attribute data. The feature extraction module 802 is used to extract two-dimensional features of students based on student data files, including: learning stage features based on the pass status and learning progress completion rate of driving test subjects, and quantification of the mastery of sub-test points under weak subjects based on weighted scoring algorithm to obtain weak subject features. User profile module 803 is used to integrate dual-dimensional features and basic attribute data to construct a student profile that includes learning stage tags and weak subject tags.
[0087] The user profiling device 800 for driver training provided in the above embodiments can implement the technical solutions described in the above embodiments of the user profiling method for driver training. The specific implementation principles of each module or unit can be found in the corresponding content in the above embodiments of the user profiling method for driver training, and will not be repeated here.
[0088] like Figure 9 As shown, the present invention also provides an electronic device 900. The electronic device 900 includes a processor 901, a memory 902, and a display 903. Figure 9 Only some components of the electronic device 900 are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.
[0089] In some embodiments, processor 901 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 902 or process data, such as the user profiling method for driver training in this invention.
[0090] In some embodiments, processor 901 may be a single server or a group of servers. The server group may be centralized or distributed. In some embodiments, processor 901 may be local or remote. In some embodiments, processor 901 may be implemented on a cloud platform. In one embodiment, the cloud platform may include a private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, intranet, multi-cloud, etc., or any combination thereof.
[0091] In some embodiments, memory 902 may be an internal storage unit of electronic device 900, such as a hard disk or memory of electronic device 900. In other embodiments, memory 902 may also be an external storage device of electronic device 900, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on electronic device 900.
[0092] Furthermore, the memory 902 may include both internal storage units of the electronic device 900 and external storage devices. The memory 902 is used to store application software and various types of data installed on the electronic device 900.
[0093] In some embodiments, display 903 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 903 is used to display information from electronic device 900 and to display a visual user interface. Components 901-903 of electronic device 900 communicate with each other via a system bus.
[0094] In one embodiment, when processor 901 executes a user profiling program for driver training stored in memory 902, the following steps can be performed: Collect multi-source heterogeneous data of trainees and perform data preprocessing to obtain trainee data profiles. The multi-source heterogeneous data includes learning stage data, ability assessment data, and basic attribute data. Based on student data profiles, two-dimensional features of students are extracted, including: learning stage features based on the pass status and learning progress completion rate of driving test subjects, and weak subject features by quantifying the mastery of sub-test points under weak subjects based on a weighted scoring algorithm. By integrating dual-dimensional features and basic attribute data, a student profile is constructed that includes learning stage tags and weak subject tags.
[0095] It should be understood that when the processor 901 executes the user profile program for driver training in the memory 902, in addition to the functions mentioned above, it can also perform other functions, as detailed in the description of the corresponding method embodiments above.
[0096] Furthermore, the embodiments of the present invention do not specifically limit the type of the electronic device 900 mentioned. The electronic device 900 can be a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, laptop computer, or other portable electronic device. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic device can also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, the electronic device 900 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).
[0097] Accordingly, this application also provides a computer-readable storage medium for storing computer-readable programs or instructions. When the programs or instructions are executed by a processor, they can implement the steps or functions of the user profiling method for driver training provided in the above-described method embodiments.
[0098] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.), and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.
[0099] The above provides a detailed description of the user profiling method, apparatus, electronic device, and storage medium for driver training provided by this invention. Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, those skilled in the art will recognize that there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.
Claims
1. A user profiling method for driver training, characterized in that, include: Collect multi-source heterogeneous data of trainees and perform data preprocessing to obtain trainee data profiles. The multi-source heterogeneous data includes learning stage data, ability assessment data and basic attribute data. Based on the student data profile, the student's two-dimensional features are extracted, including: learning stage features are obtained by dividing the driving test subject pass status and learning progress completion rate, and weak subject features are obtained by quantifying the mastery of sub-test points under weak subjects based on a weighted scoring algorithm. By integrating the aforementioned dual-dimensional features and basic attribute data, a student profile is constructed that includes learning stage tags and weak subject tags.
2. The user profiling method for driver training according to claim 1, characterized in that, The process of collecting heterogeneous data from multiple sources from trainees and performing data preprocessing to obtain trainee data files includes: Raw data is obtained from multiple independent driver training business systems through application programming interface calls or message queue synchronization. The raw data is cleaned, standardized, and linked and merged based on student identifiers to generate the student data profile.
3. The user profiling method for driver training according to claim 1, characterized in that, The learning stage characteristics are derived from the driving test subject pass status and learning progress completion rate, including: Based on the subject information that the student has passed, the core learning stage is determined. The core learning stage includes the period before learning begins and the learning period for subjects one through four. Within the core learning phase, the progress is subdivided based on the completion rate of the current subject. The progress subdivision includes the introductory phase, the improvement phase, the sprint phase, and the retake preparation phase.
4. The user profiling method for driver training according to claim 1, characterized in that, The method of quantifying the mastery of specific test points in weak subjects based on a weighted scoring algorithm includes: Obtain the indicator values of the subdivided test points on multiple evaluation dimensions, including the error rate, the number of times the same type of question is repeatedly wrong, the percentage of points lost in the mock exam, and the redundancy of practice time. Based on the preset weight allocation, the index values on the multiple evaluation dimensions are weighted and calculated to obtain the weakness coefficient that represents the degree of mastery of the test points.
5. The user profiling method for driver training according to claim 1, characterized in that, The process of integrating the dual-dimensional features and basic attribute data to construct a student profile that includes learning stage tags and weak subject tags includes: The learning stage characteristics, weak subject characteristics, and basic attribute data are used to generate a student profile containing individual student tags. Based on the individual labels, a clustering algorithm is used to classify trainees into subgroups with different label combinations, and these subgroups constitute the group profile part of the trainee profile.
6. The user profiling method for driver training according to claim 4, characterized in that, The user profiling method for driver training includes dynamically iteratively updating the student profile, wherein the dynamic iterative updating of the student profile includes: The learning stage tags and weak subject tags are updated based on a dual-mode update mechanism, which includes event-triggered and timed-triggered updates.
7. The user profiling method for driver training according to claim 6, characterized in that, The update of the learning stage tags and weak subject tags based on the dual-mode update mechanism includes: In response to events such as students completing exercises or mock exams, the weakness coefficients of relevant test points are recalculated in real time, and the characteristics of the weak subjects are updated. According to a preset cycle, the latest progress data of all students is collected and processed periodically, and the learning stage characteristics are updated.
8. A user profiling device for driver training, characterized in that, include: The data acquisition module is used to collect multi-source heterogeneous data of trainees and perform data preprocessing to obtain trainee data files. The multi-source heterogeneous data includes learning stage data, ability assessment data, and basic attribute data. The feature extraction module is used to extract two-dimensional features of students based on the student data files, including: learning stage features based on the pass status and learning progress completion rate of driving test subjects, and quantifying the mastery of sub-test points under weak subjects based on a weighted scoring algorithm to obtain weak subject features. The user profile module is used to integrate the dual-dimensional features and basic attribute data to construct a student profile that includes learning stage tags and weak subject tags.
9. An electronic device, characterized in that, Including memory and processor, among which, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the user profiling method for driver training as described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the user profiling method for driver training as described in any one of claims 1 to 7.
Citation Information
Cited By
Intelligent intervention method and system for multi-source data fusion
CN122332628A
