An intelligence evaluation system based on intelligent algorithm

Abstract

The application provides an intelligence evaluation system based on an intelligent algorithm, which continuously records attribute coding and position coding of a preset geometric body through an operation coding construction module, and converts the attribute coding and the position coding into a displacement operation and an operation event sequence by an operation event modeling module, so that the system can restore the actual operation process of a target user instead of judging only according to a final result. An attribute interference identification module further identifies a failure behavior under a multi-dimensional attribute interference relationship and generates attribute interference features, and a task deviation analysis module generates task matching features in combination with a position change result, so as to distinguish whether an error is caused by attribute resolution deficiency, spatial matching deviation or rule understanding deviation. An intelligent evaluation model module obtains an attribute resolution capability evaluation value, a rule understanding capability evaluation value and a spatial matching capability evaluation value based on the above features, and outputs an intelligence evaluation result. Therefore, the refinement degree and the accuracy of intelligence evaluation can be improved, and subsequent targeted training is facilitated.

Description

Technical Field

[0001] This application relates to the field of intelligent assessment technology, specifically to an intelligence assessment system based on intelligent algorithms. Background Technology

[0002] Currently, in assessing children's cognitive abilities, rule comprehension, spatial matching, and operational coordination, geometric manipulatives with multi-dimensional attributes such as color, shape, and material are often used to guide users in completing tasks such as identification, placement, shape matching, and attribute classification. For example, users need to place geometric objects with specific colors, shapes, or materials into corresponding grooves or containers according to task requirements. Since the same geometric object usually contains multiple identifiable attributes, users not only need to determine whether the target geometric object meets the current task requirements during the operation, but also need to complete the selection, movement, and placement under interference conditions of similar colors, shapes, or materials. Therefore, this type of operational task can reflect the user's attribute discrimination, spatial matching, and rule execution to a certain extent.

[0003] Current assessment methods typically rely on teachers, trainers, or parents observing the user's actions and judging the results based on whether the user completed the task, whether the placement was correct, and the time taken. While this method can roughly assess the user's task completion, it struggles to accurately record every selection, displacement, misplacement, and correction during the continuous operation, and it also makes it difficult to determine the specific reasons for errors. For example, a user might mistakenly select a geometric shape with a different shape because they are the same color, or ignore material differences because they are the same shape. They might be able to match the groove but fail to correctly execute the subsequent attribute classification rules. Judging solely based on the final placement result or total time is insufficient to distinguish whether the error stems from insufficient attribute discrimination, spatial matching deviation, or rule misunderstanding.

[0004] Furthermore, existing evaluation methods based on intelligent interaction primarily focus on recording general interactive behaviors such as clicking, dragging, and selecting, and statistically analyzing operation paths, error counts, or completion times. These methods typically fail to fully integrate the multi-dimensional attributes of the physical geometry, such as color, shape, and material, nor do they model the continuous displacement process of the geometry between its original position, groove position, container position, and corrected position. Therefore, existing methods struggle to identify user errors arising from multi-dimensional attribute interference based on continuous operation data, and also fail to incorporate execution biases in shape matching rules and attribute classification rules to provide a more refined and interpretable assessment of user intelligence. Summary of the Invention

[0005] In view of the above problems, this application provides an intelligence assessment system based on intelligent algorithms, which can improve the differentiation of intelligence assessment results in attribute discrimination ability, spatial matching ability and rule understanding ability.

[0006] To achieve the above objectives, this application provides an intelligence assessment system based on intelligent algorithms, comprising:

[0007] An operation coding construction module is used to acquire continuous operation coding data generated when a target user operates the teaching aid; the continuous operation coding data includes at least the attribute coding and position coding of a preset geometry.

[0008] The operation event modeling module is used to extract the displacement operations performed by the user on the preset geometry based on the continuous operation encoding data, and generate an operation event sequence based on the displacement operations;

[0009] The attribute interference identification module is used to identify user erroneous behaviors under multidimensional attribute interference relationships based on the changes in the attribute codes and position codes of adjacent operation events in the operation event sequence, and to generate attribute interference features.

[0010] The task deviation analysis module is used to determine the target user's execution deviation of the geometric shape matching rules and attribute classification rules based on the attribute interference features and the position change results of the preset geometry, and to generate task matching features.

[0011] The intelligent evaluation model module is used to input the attribute interference features and the task matching features into a preset intelligence evaluation model to obtain the target user's attribute discrimination ability evaluation value, rule understanding ability evaluation value and spatial matching ability evaluation value, and output the target user's intelligence evaluation result based on the attribute discrimination ability evaluation value, the rule understanding ability evaluation value and the spatial matching ability evaluation value.

[0012] Unlike existing technologies, the above-mentioned technical solution, by setting up an operation coding construction module, an operation event modeling module, an attribute interference identification module, a task deviation analysis module, and an intelligent evaluation model module, enables the continuous recording of attribute codes and position codes generated by the target user when operating the teaching aid, and further transforms them into displacement operations and operation event sequences. Thus, the system does not make a rough judgment based solely on the final placement result or completion time, but can reconstruct the actual operation process of the target user in the evaluation task. Furthermore, the attribute interference identification module, based on changes in attribute codes and position codes in adjacent operation events, identifies the target user's erroneous behaviors under multi-dimensional attribute interference relationships such as color, shape, and material, and generates attribute interference features, thereby reflecting the degree to which the target user is affected by different attribute interferences. The task deviation analysis module, combining attribute interference features and the position change results of the geometric object, determines the target user's execution deviation of the geometric shape matching rules and attribute classification rules, and generates task matching features, enabling the system to distinguish whether the error originates from insufficient attribute discrimination, spatial matching deviation, or rule comprehension deviation. The intelligent assessment model module then inputs attribute interference features and task matching features into a preset intelligence assessment model to obtain attribute discrimination ability assessment values, rule comprehension ability assessment values, and spatial matching ability assessment values, and outputs intelligence assessment results accordingly. Therefore, this application can improve the detail and accuracy of intelligence assessment results, enabling the assessment results to not only reflect whether the target user has completed the task, but also to reflect their specific performance in different ability dimensions, facilitating subsequent targeted training or intervention.

[0013] The above description of the invention is merely an overview of the technical solution of this application. In order to enable those skilled in the art to better understand the technical solution of this application and to implement it based on the description and drawings, and to make the above-mentioned objectives and other objectives, features and advantages of this application easier to understand, the following description is provided in conjunction with the specific embodiments and drawings of this application. Attached Figure Description

[0014] The accompanying drawings are only used to illustrate the principles, implementation methods, applications, features, and effects of specific embodiments of the present invention and other related contents, and should not be considered as limitations on this application.

[0015] In the accompanying drawings of the instruction manual:

[0016] Figure 1 This is an architecture diagram of an intelligence assessment system based on intelligent algorithms, as shown in the embodiment. Detailed Implementation

[0017] To illustrate the possible application scenarios, technical principles, implementable specific solutions, and achievable objectives and effects of this application in detail, the following description, in conjunction with the listed specific embodiments and accompanying drawings, provides a detailed explanation. The embodiments described herein are merely illustrative of the technical solutions of this application and are therefore intended to limit the scope of protection of this application.

[0018] Please refer to Figure 1 An intelligence assessment system based on intelligent algorithms includes:

[0019] An operation coding construction module is used to acquire continuous operation coding data generated when a target user operates the teaching aid; the continuous operation coding data includes at least the attribute coding and position coding of a preset geometry.

[0020] The operation event modeling module is used to extract the displacement operations performed by the user on the preset geometry based on the continuous operation encoding data, and generate an operation event sequence based on the displacement operations;

[0021] The attribute interference identification module is used to identify user erroneous behaviors under multidimensional attribute interference relationships based on the changes in the attribute codes and position codes of adjacent operation events in the operation event sequence, and to generate attribute interference features.

[0022] The task deviation analysis module is used to determine the target user's execution deviation of the geometric shape matching rules and attribute classification rules based on the attribute interference features and the position change results of the preset geometry, and to generate task matching features.

[0023] The intelligent evaluation model module is used to input the attribute interference features and the task matching features into a preset intelligence evaluation model to obtain the target user's attribute discrimination ability evaluation value, rule understanding ability evaluation value and spatial matching ability evaluation value, and output the target user's intelligence evaluation result based on the attribute discrimination ability evaluation value, the rule understanding ability evaluation value and the spatial matching ability evaluation value.

[0024] The teaching aids may include multiple preset geometric shapes, grooves for placing the preset geometric shapes, and plates for categorizing the preset geometric shapes. The preset geometric shapes can be multi-dimensional objects with various attributes, such as shapes with color, shape, and material attributes simultaneously. Attribute encoding can be used to represent at least one or more attributes of the preset geometric shape, including color, shape, and material. Position encoding can be used to represent the location of the preset geometric shape, such as its original location, groove location, plate location, non-target location, or target location. Continuous operation encoding data can be generated according to the chronological order of the target user's operations, enabling the system to obtain the target user's continuous operations on the preset geometric shapes.

[0025] Through the above method, the system can extract displacement operations from the target user's operation of the teaching aids and organize these operations into a sequence of operation events, rather than simply recording the final result. Since the sequence of operation events retains the attribute changes and positional changes of the preset geometric objects, it can further identify whether the target user makes erroneous behaviors under multi-dimensional attribute interference. For example, when the target user needs to match shapes, they may mistakenly select geometric objects because they are the same color; when the target user needs to classify by color or material, they may place the wrong plate because it is the same shape. By transforming such erroneous behaviors into attribute interference features, the system can reflect the degree to which the target user is affected by different attributes under conditions of multiple attributes coexisting. Further combining the positional change results to determine the target user's deviation from shape matching rules and attribute classification rules can form task matching features, enabling the intelligence assessment results to simultaneously reflect attribute discrimination ability, spatial matching ability, and rule understanding ability. Therefore, this application can improve the process-oriented, refined, and targeted nature of intelligence assessment, avoiding rough judgments based solely on completion time or final accuracy.

[0026] In some implementations, the operation code construction module is used to generate single operation codes according to the attribute codes and position codes of a preset geometry, and sort the single operation codes according to the generation time of the single operation codes to obtain the continuous operation code data.

[0027] Specifically, each time a target user completes an operation on a preset geometry, the system generates a single operation code based on the attribute code corresponding to that preset geometry and the current position code. A single operation code may include the geometry identifier, attribute code, position code, and generation time. The generation time can be the time when the system collects the operation code, or the data collection time after the target user triggers the operation. The system sorts multiple single operation codes according to their generation time to obtain continuous operation code data.

[0028] In this way, each operation can be recorded in a structured manner, and there is a clear temporal sequence between multiple individual operation codes. This avoids the problem of obtaining only scattered operation data and being unable to reconstruct the operation process, thus providing a reliable data foundation for subsequent displacement operation extraction, operation event sequence generation, and error behavior identification. Since the individual operation code includes both attribute and position codes, it is possible to simultaneously analyze which geometry the target user selected and where that geometry was moved, thereby enhancing the interpretability of the intelligence assessment results.

[0029] In some implementations, the operation event modeling module is used to determine the displacement start point and displacement end point of the preset geometry based on the position encoding changes of the same preset geometry in adjacent consecutive operation encoding data;

[0030] When the displacement start point and the displacement end point correspond to the original position and the preset groove position respectively, the displacement operation is determined to be a shape matching displacement.

[0031] When the displacement start point and the displacement end point correspond to the preset groove position and the preset plate position respectively, the displacement operation is determined to be an attribute classification displacement.

[0032] When the displacement start point corresponds to a non-target position and the displacement end point corresponds to a target position, the displacement operation is determined to be a correction displacement.

[0033] Specifically, the system can determine the position from which a preset geometry has moved based on the positional encoding changes of the same preset geometry in adjacent consecutive operation encoding data. If the preset geometry moves from its original position to a preset groove position, it indicates that the target user is matching the geometry shape with the groove position, and this can be identified as a shape matching displacement. If the preset geometry moves from a preset groove position to a preset disk position, it indicates that the target user is classifying the geometry into the corresponding disk according to the attribute classification rules in the evaluation task, and this can be identified as an attribute classification displacement. If the preset geometry moves from a non-target position to the target position, it indicates that the target user may have corrected a previous erroneous operation, and this can be identified as a correction displacement.

[0034] Using the above method, ordinary positional changes can be further categorized into shape-matching displacement, attribute-classification displacement, and corrective displacement, allowing different types of displacement operations to correspond to different capability assessment meanings. Shape-matching displacement reflects the target user's ability to match geometric shapes and the spatial position of grooves; attribute-classification displacement reflects the target user's understanding of the classification rules for attributes such as color, shape, and material; and corrective displacement reflects the target user's responsiveness in adjusting after discovering errors. Therefore, subsequent assessments no longer simply determine whether the user moved the geometry, but can analyze different dimensions of intellectual ability based on different displacement types.

[0035] In some implementations, the operation event modeling module is further configured to associate the data acquisition time of the displacement operation, the displacement start point, the displacement end point, the attribute code of the preset geometry, and the corresponding continuous operation code data to generate a single operation event, and to sort multiple single operation events according to the data acquisition time to generate the operation event sequence.

[0036] Specifically, after determining that a preset geometry has undergone a displacement operation, the system associates the type of the displacement operation, the starting point of the displacement, the ending point of the displacement, the attribute code of the preset geometry, and the corresponding data acquisition time to form a single operation event. A single operation event can be understood as a complete analyzable operation unit. Subsequently, the system sorts multiple single operation events according to the data acquisition time to form an operation event sequence.

[0037] Through the above method, continuous operation coded data is further organized into operation event sequences with event meaning. Compared with directly analyzing the raw coded data, operation event sequences can more clearly express what displacement operations the target user performed at different times, which preset geometry was moved, from where to where, and what attributes the geometry has. This provides a foundation for analyzing attribute and position code changes between adjacent operation events, making subsequent identification of erroneous behaviors under multi-dimensional attribute interference relationships more accurate.

[0038] In some implementations, the attribute interference identification module is used to compare the attribute code of the preset geometry corresponding to the current operation event with the target attribute requirements of the preset evaluation task;

[0039] When the preset geometry and the target attribute requirement have at least one class of the same attribute and at least one class of different attributes, the preset geometry is identified as an attribute similarity interference object.

[0040] When the position encoding change of the similar interference object in the current operation event does not meet the target position requirements of the preset evaluation task, the current operation event is identified as an erroneous behavior under the multidimensional attribute interference relationship.

[0041] Specifically, the preset evaluation task may require the target user to select or place a geometry that meets specific attribute requirements, such as placing a red circular geometry, a plastic geometry, or a geometry of a specified shape. The system compares the attribute code of the preset geometry corresponding to the current operation event with the target attribute requirements. If the preset geometry and the target attribute requirements have at least one class of the same attributes and at least one class of different attributes, it indicates that the preset geometry and the target requirements have some similarity, but do not completely meet the target requirements. In this case, the system identifies the preset geometry as an attribute similarity interference object. If the position code change of the attribute similarity interference object does not meet the target position requirements, the current operation event can be identified as an erroneous behavior under multi-dimensional attribute interference relationships.

[0042] By employing the above methods, the system can identify errors made by target users that are not random, but rather errors influenced by attribute similarity. For example, if the target task requires selecting a red circular geometric shape, and the target user selects a red square geometric shape, then that shape is distracting due to its similar color; similarly, if the target task requires selecting a plastic triangular geometric shape, and the target user selects a rubber triangular geometric shape, then that shape is distracting due to its similar shape. Identifying such errors as mistakes arising from multi-dimensional attribute interference relationships more accurately reflects the target user's ability to distinguish and resist interference under multi-attribute information.

[0043] In some implementations, the attribute interference identification module is further configured to determine the type of interference attribute that causes the erroneous behavior based on the attribute types that are the same as and different from the target attribute in the attribute-similar interference objects.

[0044] Statistically analyze the number of erroneous behaviors, the number of consecutive erroneous behaviors, and the correction interval between the erroneous behavior and the target displacement operation for the same type of interference attribute.

[0045] Based on the number of erroneous behaviors, the number of consecutive occurrences of the erroneous behaviors, and the correction interval, attribute interference features are generated to characterize the degree of influence of the target user on the corresponding interference attribute type.

[0046] Specifically, once the system identifies a preset geometric shape as an object with similar attributes causing interference, it can further analyze which attributes of that geometry are the same as the target attribute requirements and which are different. For example, if the target requirement is a red circular plastic geometry, but the user selects a red square plastic geometry, then the shape difference can be identified as a significant interference attribute type causing erroneous behavior. The system counts the number of erroneous behaviors corresponding to the same interference attribute type, and further analyzes whether similar erroneous behaviors occur consecutively, as well as the correction interval between the occurrence of the erroneous behavior and the target user's execution of the target displacement operation.

[0047] In this way, attribute interference features go beyond simply recording how many times a user made a mistake. They characterize the type of attribute interference the target user experiences, the frequency of interference, whether the interference is continuous, and the speed of correction. The number of errors reflects the degree of interference, the number of consecutive errors reflects the persistence of interference, and the correction interval reflects the reaction time required for the target user to detect and correct the error. The resulting attribute interference features more accurately reflect the target user's attribute discrimination ability and cognitive stability.

[0048] In some implementations, the task deviation analysis module is used to determine whether the preset geometry is placed in the groove position corresponding to its shape based on the position change result of the preset geometry and the preset shape code, so as to obtain the shape matching deviation;

[0049] Based on the position change results and attribute codes of the preset geometry, it is determined whether the preset geometry is placed in the disk position corresponding to the attribute classification rule, and the attribute classification deviation is obtained.

[0050] Based on the attribute interference characteristics, determine the interference attribute types corresponding to the shape matching deviation and the attribute classification deviation; generate the execution deviation based on the interference attribute types, the shape matching deviation, and the attribute classification deviation.

[0051] Specifically, the system can determine whether a preset geometry has been placed in a groove position based on its positional changes, and further determine whether the groove position corresponds to the geometry's shape based on a preset shape code. If the geometry is not placed in the groove position corresponding to its shape, a shape matching deviation is obtained. The system can also determine whether a preset geometry has been placed in a plate position based on its positional changes, and determine whether the plate position is correct based on attribute codes and attribute classification rules. If the geometry is not placed in the plate position corresponding to the attribute classification rules, an attribute classification deviation is obtained.

[0052] Furthermore, the system combines attribute interference features to determine which types of interference attributes affect shape matching deviation and attribute classification deviation, and generates execution deviation based on interference attribute type, shape matching deviation, and attribute classification deviation.

[0053] Using the above method, the system can map the groove positions and plate positions to different task rules. The groove position focuses more on geometric shape matching, while the plate position focuses more on understanding attribute classification rules. By combining shape matching deviation and attribute classification deviation with attribute interference features, it can be determined whether the target user's error is related to specific attribute interference. For example, a user may be able to correctly identify the shape, but be interfered with by color or material when classifying attributes; or they may be affected by shape similarity when matching grooves. This execution deviation can reflect the target user's ability to execute different rules more accurately than simply the number of misplacements.

[0054] In some implementations, the task deviation analysis module is further configured to label the shape matching deviation and the attribute classification deviation with attribute type based on the attribute interference characteristics;

[0055] When the same attribute type is associated with both the shape matching deviation and the attribute classification deviation, the deviation weight corresponding to this attribute type is increased; the task matching feature is generated based on the deviation weight.

[0056] Specifically, after obtaining shape matching deviations and attribute classification deviations, the system labels these deviations by attribute type based on attribute interference features. For example, a shape matching deviation is labeled as shape interference, and an attribute classification deviation is labeled as material interference or color interference. If the same attribute type appears in both shape matching deviations and attribute classification deviations, it indicates that the target user is affected by this attribute type under different task rules, and the system increases the deviation weight corresponding to this attribute type. Subsequently, task matching features are generated based on the deviation weights.

[0057] In this way, task matching features can reflect common sources of deviation in different rule-based tasks. If a certain attribute type only appears sporadically in a single task, its impact on the evaluation results is low; if the same attribute type affects both shape matching and attribute classification, it indicates that this attribute type may constitute a significant weakness in the target user's abilities. By increasing the weight of the corresponding deviations, the intelligence assessment results can more sensitively reflect key sources of error and improve the assessment results' ability to identify weak abilities.

[0058] In some implementations, the intelligent evaluation model module is used to determine the attribute discrimination ability evaluation value of the target user based on the attribute interference features; determine the rule understanding ability evaluation value and spatial matching ability evaluation value of the target user based on the task matching features; and output the intelligence evaluation result according to the attribute discrimination ability evaluation value, the rule understanding ability evaluation value and the spatial matching ability evaluation value.

[0059] Specifically, attribute interference features can be used to determine the target user's ability to distinguish different geometric attributes in scenarios with multiple attributes, thus obtaining an attribute discrimination ability assessment value. The content related to attribute classification rules in task matching features can be used to determine the target user's understanding of task rules, thus obtaining a rule comprehension ability assessment value. The content related to groove shape matching in task matching features can be used to determine the target user's grasp of the correspondence between geometric objects and spatial positions, thus obtaining a spatial matching ability assessment value. The system can generate an intelligence assessment result for the target user based on these multiple assessment values.

[0060] In this way, intelligence assessment results can be broken down into multiple meaningful ability dimensions, rather than simply outputting a single score. The attribute discrimination ability assessment value corresponds to the target user's ability to distinguish attributes such as color, shape, and material; the rule comprehension ability assessment value corresponds to the target user's ability to understand and execute attribute classification rules; and the spatial matching ability assessment value corresponds to the target user's ability to recognize the spatial relationship between shapes and grooves. These multiple ability dimensions together form the intelligence assessment result, which is beneficial for subsequent training courses or intervention programs to adjust for weaknesses.

[0061] In some embodiments, the system further includes a corrective response analysis module, which is used to determine, in the sequence of operational events, a subsequent displacement operation of a preset geometry corresponding to the erroneous behavior after the erroneous position.

[0062] Based on the termination position code of the subsequent displacement operation and the target position requirement of the preset evaluation task, it is determined whether the subsequent displacement operation is a valid correction operation.

[0063] Based on the event interval between the erroneous behavior and the effective corrective action, as well as the changes in attribute encoding and location encoding corresponding to the effective corrective action, a corrective response feature is generated;

[0064] The intelligent assessment model module is also used to correct the intelligence assessment results of the target user based on the corrective response characteristics.

[0065] Specifically, when the system identifies a erroneous behavior, it does not immediately output it as an error result. Instead, it continues to search the operation event sequence for subsequent displacement operations of the same preset geometry after the erroneous location. If the termination position code of this subsequent displacement operation meets the target position requirements of the preset evaluation task, the system can determine that the subsequent displacement operation is a valid corrective operation. The system further calculates the event interval between the erroneous behavior and the valid corrective operation, and combines the attribute code and position code changes corresponding to the valid corrective operation to generate corrective response features. Corrective response features can characterize the target user's ability to discover errors, understand target locations, and complete corrections.

[0066] Through the above methods, the system can distinguish the error types of target users. If a target user makes a mistake but can move the preset geometry to the target position within a short time interval, it indicates that they have good self-correction and rule backtracking abilities. If a target user fails to correct their behavior for a long time, or fails to reach the target position after multiple corrections, it indicates that they may have more significant difficulties in rule understanding, spatial matching, or attribute discrimination. Therefore, modifying the intelligence assessment results based on the characteristics of corrective responses can avoid simply treating all errors as the same, making the assessment results more consistent with the target user's actual cognitive process.

[0067] Example 1:

[0068] This embodiment provides a specific application of an intelligence assessment system based on intelligent algorithms. The teaching aid may include multiple preset geometric shapes, multiple preset groove positions, and multiple preset plate positions. The preset geometric shapes may have various attributes such as color, shape, and material. For example, colors may include red, yellow, and green; shapes may include squares, circles, and triangles; and materials may include metal, plastic, and rubber. The system can set attribute codes for each preset geometric shape, including color codes, shape codes, and material codes. The system can also set position codes for the original placement area, groove area, plate area, non-target positions, and target positions in the teaching aid, enabling the identification and recording of positional changes of the preset geometric shapes.

[0069] Once the target user begins the evaluation task, the system presents preset evaluation tasks. These preset tasks may include a groove shape matching task and a plate attribute classification task. The groove shape matching task may require the target user to place geometric shapes into the groove positions corresponding to their shapes. For example, placing a square geometric shape into a square groove, a round geometric shape into a round groove, and a triangular geometric shape into a triangular groove. The plate attribute classification task may require the target user to place geometric shapes into corresponding plates according to one or more attributes such as color, shape, or material. For example, placing a red geometric shape into the first plate, a plastic geometric shape into the second plate, or a round plastic geometric shape into the third plate.

[0070] During the user's operation of the teaching aid, the operation coding construction module continuously acquires the attribute codes and position codes of preset geometries. Whenever a change in the position of a preset geometries is detected, the operation coding construction module generates a single operation code based on the attribute and position codes of that geometries and records the data acquisition time of that single operation code. A single operation code can include geometries identifiers, color codes, shape codes, material codes, position codes, and generation times. The operation coding construction module sorts multiple single operation codes according to their generation times to obtain continuous operation code data.

[0071] In this way, the system can record the target user's continuous operation process throughout the entire evaluation task. Compared with only recording the final placement result, the continuous operation coding data can reflect which geometry the target user selected at different times, what attributes the geometry has, and from which position the geometry was moved to which position, thus providing a basis for subsequent operation event modeling and error source analysis.

[0072] Example 2

[0073] Based on Example 1, the operation event modeling module receives continuous operation code data output by the operation code construction module, and determines the displacement start point and displacement end point of the preset geometry according to the position code change of the same preset geometry in adjacent continuous operation code data.

[0074] When the displacement start point and displacement end point correspond to the original position and the preset groove position, respectively, the operation event modeling module determines the displacement operation as a shape matching displacement. The shape matching displacement is used to characterize the operation process of the target user moving the preset geometry from the original position to the groove position. This operation process can reflect the target user's judgment on the correspondence between the geometry shape and the spatial position of the groove.

[0075] When the displacement start point and displacement end point correspond to the preset groove position and preset plate position respectively, the operation event modeling module determines the displacement operation as attribute classification displacement. Attribute classification displacement is used to characterize the operation process in which the target user, after completing or attempting to complete shape matching, further places the geometry into the corresponding plate position according to attribute classification rules such as color, shape, and material. This operation process can reflect the target user's understanding and execution of the attribute classification rules.

[0076] When the displacement start point corresponds to a non-target position and the displacement end point corresponds to the target position, the operation event modeling module identifies the displacement operation as a corrective displacement. The corrective displacement characterizes whether the target user can reposition the preset geometry to the target position after an operation that results in misplacement, incorrect selection, or failure to meet task requirements. This displacement operation can be used for subsequent analysis of the target user's corrective response capability.

[0077] Furthermore, the operation event modeling module correlates the displacement operation, displacement start point, displacement end point, attribute codes of the preset geometry, and the data acquisition time of the corresponding continuous operation codes to generate individual operation events. An individual operation event can represent a complete analyzable operation. For example, one individual operation event could represent the target user moving a red square plastic geometry from its original position to a square groove at a certain time. Another individual operation event could represent the target user moving the red square plastic geometry from the square groove to a plate used to receive the red geometry. The operation event modeling module sorts multiple individual operation events according to their data acquisition time to generate an operation event sequence.

[0078] Through the above processing, continuous operation encoded data is converted into an event sequence with behavioral meaning. This operation event sequence not only preserves the temporal order but also retains geometric attributes, positional changes, and displacement types, enabling the system to further analyze whether the target user's operations meet the preset evaluation task requirements and whether errors are related to interference from a certain type of attribute.

[0079] Example 3

[0080] Based on Example 2, the attribute interference identification module identifies the erroneous behavior of the target user under multi-dimensional attribute interference relationship according to the operation event sequence.

[0081] Specifically, the attribute interference recognition module compares the attribute code of the preset geometry corresponding to the current operation event with the target attribute requirements of the preset evaluation task. The target attribute requirements may include one or more of the following: color requirements, shape requirements, and material requirements. For example, the preset evaluation task may require the target user to select a red circular plastic geometry, or it may require the target user to select all plastic geometry, or it may require the target user to place the circular geometry into the corresponding groove and then place it into the corresponding plate according to color.

[0082] When the preset geometry corresponding to the current operation event shares at least one type of attribute with the target attribute requirement and also shares at least one type of attribute difference, the attribute interference identification module identifies the preset geometry as an attribute-similar interference object. Attribute-similar interference objects are not entirely unrelated to the target requirement; rather, they share some attributes with the target requirement but differ from it in others, thus easily interfering with the target user. For example, if the task requires selecting a red circular plastic geometry, and the target user selects a red square plastic geometry, this geometry shares the same color and material as the target requirement but differs in shape, and can be identified as an attribute-similar interference object. Similarly, if the task requires selecting a plastic triangular geometry, and the target user selects a rubber triangular geometry, this geometry shares the same shape as the target requirement but differs in material, and can also be identified as an attribute-similar interference object.

[0083] When the positional encoding change of an object with similar attributes in the current operation event does not meet the target position requirements of the preset evaluation task, the attribute interference identification module identifies the current operation event as an erroneous behavior under multi-dimensional attribute interference relationships. This erroneous behavior not only indicates that the target user's operation result is incorrect, but also indicates that the error is related to similar interference between attributes such as color, shape, and material.

[0084] The attribute interference identification module further determines the type of interfering attribute that causes the erroneous behavior based on the attribute types that are the same as and different from those of the target attribute among the interference objects with similar attributes. For example, if the target user repeatedly selects geometric objects with the same color but different shapes, then the shape can be determined as the interfering attribute type corresponding to the erroneous behavior. If the target user repeatedly selects geometric objects with the same shape but different materials, then the material can be determined as the interfering attribute type corresponding to the erroneous behavior.

[0085] The attribute interference identification module counts the number of erroneous behaviors, the number of consecutive occurrences of erroneous behaviors, and the correction interval between the erroneous behavior and the target displacement operation for the same interference attribute type. It then generates attribute interference features based on these factors. The number of erroneous behaviors reflects the frequency with which the target user is affected by the interference attribute type; the number of consecutive occurrences reflects whether the target user is continuously affected by the same attribute type; and the correction interval reflects the operation interval or time interval required for the target user to recover from the erroneous state to the target operation state.

[0086] Through the above methods, the attribute interference identification module can further distinguish ordinary errors into errors caused by color interference, shape interference, material interference, or interference from multiple attributes, enabling the system to more accurately assess the attribute discrimination stability of the target user in scenarios where multiple attributes coexist.

[0087] Example 4

[0088] Based on Example 3, the task deviation analysis module determines the target user's execution deviation of the geometric shape matching rules and attribute classification rules according to the attribute interference characteristics and the position change results of the preset geometry.

[0089] Specifically, the task deviation analysis module determines whether a preset geometry is placed in a groove corresponding to its shape based on the positional changes of the preset geometry and the preset shape code, thus obtaining a shape matching deviation. For example, if a circular geometry is placed in a square groove, or a triangular geometry is placed in a circular groove, then a shape matching deviation is confirmed. The shape matching deviation reflects the target user's recognition of the correspondence between the geometry's shape and the groove's position.

[0090] The task deviation analysis module also determines whether the preset geometry was placed in the corresponding disk position according to the positional changes and attribute codes of the preset geometry, thus obtaining the attribute classification deviation. For example, if the preset evaluation task requires placing a plastic geometry in the second disk, but the target user places a rubber geometry of the same color in the second disk, then an attribute classification deviation can be identified. Attribute classification deviation reflects the target user's understanding and execution of attribute classification rules such as color, shape, and material.

[0091] The task deviation analysis module further determines the types of interfering attributes corresponding to shape matching deviation and attribute classification deviation based on attribute interference characteristics, and generates execution deviation based on the interfering attribute type, shape matching deviation, and attribute classification deviation. Therefore, execution deviation includes not only whether the target user placed the item incorrectly, but also which type of attribute interference the incorrect placement is related to, and whether the error occurred in the shape matching task or the attribute classification task.

[0092] When generating task matching features, the task deviation analysis module labels shape matching deviation and attribute classification deviation by attribute interference features. When the same attribute type is associated with both shape matching deviation and attribute classification deviation, the task deviation analysis module increases the deviation weight corresponding to this attribute type and generates task matching features based on the deviation weight. For example, if a target user makes an error in the groove shape matching task due to shape similarity, and also makes an error in the plate attribute classification task due to shape similarity, it indicates that the target user is affected by the shape attribute under different task rules, and the task deviation analysis module increases the deviation weight corresponding to the shape attribute. As another example, if a target user is only occasionally affected by color interference in a single classification task, and does not repeatedly exhibit the same deviation in other tasks, then the deviation weight corresponding to the color attribute can be lower than that of interference attribute types that simultaneously affect multiple task stages.

[0093] In this way, task matching features can reveal common sources of deviation across different tasks. Compared to simply counting misplacements, these features can more accurately reflect the target user's weaknesses in rule understanding, spatial matching, and attribute transformation.

[0094] Example 5

[0095] Based on Example 4, the intelligent evaluation model module receives the attribute interference features output by the attribute interference identification module and the task matching features output by the task deviation analysis module, and inputs the attribute interference features and task matching features into the preset intelligence evaluation model.

[0096] The pre-defined intelligence assessment model can be a weighted fusion-based assessment model or a machine learning assessment model trained with training samples. Training samples may include historical users' attribute interference features, task matching features, corrective response features, and corresponding human assessment results or standard assessment results. Human assessment results or standard assessment results can be provided by teachers, trainers, assessors, or existing scale results, and are used to enable the pre-defined intelligence assessment model to learn the correspondence between different operational characteristics and different ability performances.

[0097] In one implementation, the intelligent evaluation model module determines the attribute discrimination ability assessment value of the target user based on attribute interference characteristics. Attribute interference characteristics reflect the degree to which the target user is affected by different attributes when multiple attributes such as color, shape, and material coexist. When the number of erroneous behaviors corresponding to a certain type of interference attribute is high, the number of consecutive erroneous behaviors is high, and the correction interval is long, it indicates that the target user's discrimination stability for this type of attribute is weak, and the intelligent evaluation model module correspondingly lowers the attribute discrimination ability assessment value. When the number of erroneous behaviors corresponding to a certain type of interference attribute is low, or when errors are corrected quickly, it indicates that the target user has a good discrimination and adjustment ability for this type of attribute, and the intelligent evaluation model module can correspondingly increase or maintain the attribute discrimination ability assessment value.

[0098] In one implementation, the intelligent evaluation model module determines the target user's rule comprehension ability assessment value and spatial matching ability assessment value based on task matching features. The content related to attribute classification deviation in the task matching features can be used to determine the rule comprehension ability assessment value, and the content related to shape matching deviation in the task matching features can be used to determine the spatial matching ability assessment value. When the target user repeatedly places a geometric object into a plate that does not conform to the attribute classification rules in the attribute classification task, the intelligent evaluation model module can lower the rule comprehension ability assessment value. When the target user repeatedly places a geometric object into an unmatched groove position in the shape matching task, the intelligent evaluation model module can lower the spatial matching ability assessment value.

[0099] When the same type of interference attribute is associated with both shape matching bias and attribute classification bias, the intelligent assessment model module can increase the bias weight of that interference attribute type in the preset intelligence assessment model, making this type of bias have a more significant impact on the intelligence assessment results. This avoids treating occasional errors and persistent ability biases equally, allowing the assessment results to more accurately reflect the target user's main weaknesses.

[0100] The intelligent assessment model module outputs the target user's intelligence assessment results based on attribute discrimination ability assessment scores, rule comprehension ability assessment scores, and spatial matching ability assessment scores. The intelligence assessment results can be a comprehensive score or include multiple sub-assessment results. Sub-assessment results can include attribute discrimination ability results, rule comprehension ability results, spatial matching ability results, and corrective response ability results. In this way, the intelligence assessment results are no longer just a single completion result judgment, but rather reflect the target user's performance across different cognitive ability dimensions.

[0101] Example 6

[0102] Based on Embodiment 5, the system may further include a corrective response analysis module. This module is used to determine, within the sequence of operational events, the subsequent displacement operations of the preset geometry corresponding to the erroneous behavior after the erroneous position, and to determine, based on these subsequent displacement operations, whether the target user has completed effective correction.

[0103] Specifically, when the attribute interference identification module detects a certain erroneous behavior, the corrective response analysis module searches the operation event sequence for the subsequent displacement operation of the preset geometry corresponding to the erroneous behavior after the erroneous position. Based on the termination position code of the subsequent displacement operation and the target position requirement of the preset evaluation task, the corrective response analysis module determines whether the subsequent displacement operation is a valid corrective operation. If the subsequent displacement operation causes the preset geometry to reach the target position, the corrective response analysis module determines the subsequent displacement operation as a valid corrective operation. If the subsequent displacement operation still fails to cause the preset geometry to reach the target position, it can be determined that the subsequent displacement operation is not a valid corrective operation, or it can be recorded as an incompletely corrected subsequent operation.

[0104] The corrective response analysis module generates corrective response features based on the event interval between the erroneous behavior and the effective corrective action, as well as the changes in the attribute and location codes corresponding to the effective corrective action. The event interval can be the number of operational events separating the erroneous behavior and the effective corrective action, or the time interval between the occurrence of the erroneous behavior and the occurrence of the effective corrective action. Corrective response features characterize the target user's ability to detect errors, understand the target location, and complete the correction.

[0105] The intelligent assessment model module can also modify the target user's intelligence assessment results based on corrective response characteristics. If the target user makes a mistake but can move the preset geometry to the target position within a short time interval, it indicates that the target user has good self-correction and rule backtracking abilities, and the intelligent assessment model module can reduce the adverse impact of this mistake on the intelligence assessment results. If the target user cannot correct the mistake for a long time, or fails to reach the target position after multiple subsequent displacement operations, it indicates that the target user may have more significant difficulties in rule understanding, spatial matching, or attribute discrimination, and the intelligent assessment model module can increase the degree of impact of this mistake on the intelligence assessment results.

[0106] By introducing corrective response features, the system can distinguish between occasional errors, quickly correctable errors, and persistent errors, making the intelligence assessment results more consistent with the target user's actual operating process and cognitive performance.

[0107] Example 7:

[0108] Based on Example 5 or Example 6, the intelligent assessment model module can determine the attribute interference feature value, task matching feature value, corrective response feature value, and intelligence assessment result in the following manner.

[0109] In some implementations, the intelligent evaluation model module can calculate the attribute interference feature value corresponding to the k-th type of interference attribute based on the number of erroneous behaviors, the number of consecutive occurrences of erroneous behaviors, the correction interval, and the effective correction ratio.

[0110]

[0111] in, This represents the attribute interference feature value corresponding to the kth type of interference attribute; the kth type of interference attribute can be color, shape, or material. E represents the number of erroneous behaviors corresponding to the k-th type of interference attribute; E represents the total number of erroneous behaviors. This represents the number of consecutive occurrences of the error behavior corresponding to the k-th type of interference attribute; L represents the total number of consecutive occurrences of all error behaviors. This represents the correction interval corresponding to the k-th type of interference attribute. This indicates the preset correction interval reference value; This represents the effective correction ratio corresponding to the kth type of interference attribute. , , , σ represents the preset weighting coefficients; σ represents the normalization function.

[0112] Wherein, the effective correction ratio It can be represented as:

[0113]

[0114] in, This represents the number of effective corrective actions performed to reach the target location after a erroneous action, under the k-th type of interference attribute. This represents the total number of erroneous behaviors under the k-th type of interference attribute.

[0115] Through the above methods, attribute interference features are eliminated. It is not determined solely by the number of erroneous actions, but rather by a combination of the number of erroneous actions, the number of consecutive occurrences of erroneous actions, the correction interval, and the effective correction rate. This allows for the differentiation between occasional errors, continuous errors, and errors that can be quickly corrected, enabling attribute interference features to more accurately characterize the degree to which the target user is affected by the corresponding interference attribute type.

[0116] In some implementations, the intelligent evaluation model module can calculate the task matching feature value corresponding to the k-th type of interference attribute based on shape matching deviation, attribute classification deviation, and attribute interference feature value:

[0117]

[0118] in, This represents the task matching feature value corresponding to the k-th type of interference attribute. This represents the shape matching deviation corresponding to the k-th type of interference attribute. This represents the attribute classification deviation corresponding to the k-th type of interference attribute. · This refers to the association terms formed when the same type of interference attribute is simultaneously associated with shape matching deviation and attribute classification deviation. and This is the preset adjustment coefficient.

[0119] Among them, shape matching deviation It can be determined based on whether a preset geometry is placed in a groove position corresponding to its shape; attribute classification deviation. This can be determined based on whether a preset geometry is placed in the disk position corresponding to the attribute classification rule. When the same interfering attribute type is associated with both shape matching deviation and attribute classification deviation... This amplifies the impact of the deviation corresponding to this type of interference attribute, thereby increasing the corresponding deviation weight for this attribute type.

[0120] In some implementations, the intelligent evaluation model module can calculate the correction response characteristic value corresponding to the k-th type of interference attribute based on the effective correction ratio, the number of subsequent displacement operations that did not reach the target position, and the correction interval.

[0121]

[0122] in, This represents the characteristic value of the corrective response corresponding to the k-th type of interference attribute. This represents the number of subsequent displacement operations performed after a erroneous action that has not yet reached the target position, under the k-th type of interference attribute. This represents the total number of erroneous behaviors under the k-th type of interference attribute. , , This is the preset adjustment coefficient.

[0123] Using the above method, when the target user can quickly reach the target location through subsequent displacement operations after an erroneous action, A higher value corresponds to a lower corrective response characteristic value Ck, indicating that this type of error has a smaller adverse impact on the intelligence assessment results. When the target user makes a mistake and there are still multiple subsequent displacement operations that fail to reach the target position, or when the correction interval is long, the corrective response characteristic value Ck increases accordingly, indicating that this type of error has a larger adverse impact on the intelligence assessment results.

[0124] In some implementations, the intelligent assessment model module can determine the intelligence assessment result of the target user based on attribute interference feature values, task matching feature values, and corrective response feature values:

[0125]

[0126] Where Z represents the intelligence assessment result of the target user; This represents the joint action term between task matching features and corrective response features; This represents the maximum value among the attribute interference feature values ​​corresponding to each interference attribute type; , , , σ represents the preset weighting coefficients; σ represents the normalization function.

[0127] Using the above formula, the intelligence assessment result Z can simultaneously reflect the degree to which the target user makes erroneous behaviors under multidimensional attribute interference, the degree of deviation in the execution of geometric shape matching rules and attribute classification rules, and the corrective response after the erroneous behavior. Compared with outputting assessment results solely based on the final accuracy rate or total time spent, the above calculation method enables the intelligence assessment results to more accurately distinguish the target user's attribute discrimination ability, spatial matching ability, rule understanding ability, and corrective response ability.

[0128] Finally, it should be noted that although the above embodiments have been described in the text and drawings of this application, this should not limit the scope of patent protection of this application. Any technical solutions that are based on the essential concept of this application and utilize the content described in the text and drawings of this application, resulting in equivalent structural or procedural substitutions or modifications, as well as the direct or indirect application of the technical solutions of the above embodiments to other related technical fields, are all included within the scope of patent protection of this application.

Claims

1. An intelligence assessment system based on intelligent algorithms, characterized in that, include: The operation coding construction module is used to obtain continuous operation coding data generated when the target user operates the teaching aid; The continuous operation encoded data includes at least the attribute encoding and position encoding of the preset geometry; The operation event modeling module is used to extract the displacement operations performed by the user on the preset geometry based on the continuous operation encoding data, and generate an operation event sequence based on the displacement operations; The attribute interference identification module is used to identify user erroneous behaviors under multidimensional attribute interference relationships based on the changes in the attribute codes and position codes of adjacent operation events in the operation event sequence, and to generate attribute interference features. The task deviation analysis module is used to determine the target user's execution deviation of the geometric shape matching rules and attribute classification rules based on the attribute interference features and the position change results of the preset geometry, and to generate task matching features. The intelligent evaluation model module is used to input the attribute interference features and the task matching features into a preset intelligence evaluation model to obtain the target user's attribute discrimination ability evaluation value, rule understanding ability evaluation value and spatial matching ability evaluation value, and output the target user's intelligence evaluation result based on the attribute discrimination ability evaluation value, the rule understanding ability evaluation value and the spatial matching ability evaluation value.

2. The intelligence assessment system based on intelligent algorithms according to claim 1, characterized in that, The operation code construction module is used to generate single operation codes according to the attribute codes and position codes of the preset geometry, and sort the single operation codes according to the generation time of the single operation codes to obtain the continuous operation code data.

3. The intelligence assessment system based on intelligent algorithms according to claim 1 or 2, characterized in that, The operation event modeling module is used to determine the displacement start point and displacement end point of the preset geometry based on the position encoding changes of the same preset geometry in adjacent continuous operation encoding data; When the displacement start point and the displacement end point correspond to the original position and the preset groove position respectively, the displacement operation is determined to be a shape matching displacement. When the displacement start point and the displacement end point correspond to the preset groove position and the preset plate position respectively, the displacement operation is determined to be an attribute classification displacement. When the displacement start point corresponds to a non-target position and the displacement end point corresponds to a target position, the displacement operation is determined to be a correction displacement.

4. The intelligence assessment system based on intelligent algorithms according to claim 3, characterized in that, The operation event modeling module is also used to associate the data acquisition time of the displacement operation, the displacement start point, the displacement end point, the attribute code of the preset geometry and the corresponding continuous operation code data to generate a single operation event, and to sort multiple single operation events according to the data acquisition time to generate the operation event sequence.

5. The intelligence assessment system based on intelligent algorithms according to claim 1, characterized in that, The attribute interference identification module is used to compare the attribute code of the preset geometry corresponding to the current operation event with the target attribute requirements of the preset evaluation task; When the preset geometry and the target attribute requirement have at least one class of the same attribute and at least one class of different attributes, the preset geometry is identified as an attribute similarity interference object. When the position encoding change of the similar interference object in the current operation event does not meet the target position requirements of the preset evaluation task, the current operation event is identified as an erroneous behavior under the multidimensional attribute interference relationship.

6. The intelligence assessment system based on intelligent algorithms according to claim 5, characterized in that, The attribute interference identification module is further configured to determine the type of interference attribute that causes the erroneous behavior based on the attribute types that are the same as and different from the target attribute in the attribute-similar interference objects. Statistically analyze the number of erroneous behaviors, the number of consecutive erroneous behaviors, and the correction interval between the erroneous behavior and the target displacement operation for the same type of interference attribute. Based on the number of erroneous behaviors, the number of consecutive occurrences of the erroneous behaviors, and the correction interval, attribute interference features are generated to characterize the degree of influence of the target user on the corresponding interference attribute type.

7. The intelligence assessment system based on intelligent algorithms according to claim 1, characterized in that, The task deviation analysis module is used to determine whether the preset geometry is placed in the groove position corresponding to its shape based on the position change result of the preset geometry and the preset shape code, and to obtain the shape matching deviation. Based on the position change results and attribute codes of the preset geometry, it is determined whether the preset geometry is placed in the disk position corresponding to the attribute classification rule, and the attribute classification deviation is obtained. Based on the attribute interference characteristics, determine the interference attribute types corresponding to the shape matching deviation and the attribute classification deviation; generate the execution deviation based on the interference attribute types, the shape matching deviation, and the attribute classification deviation.

8. The intelligence assessment system based on intelligent algorithms according to claim 7, characterized in that, The task deviation analysis module is also used to mark the shape matching deviation and the attribute classification deviation by attribute type according to the attribute interference characteristics; When the same attribute type is associated with both the shape matching deviation and the attribute classification deviation, the deviation weight corresponding to this attribute type is increased; the task matching feature is generated based on the deviation weight.

9. The intelligence assessment system based on intelligent algorithms according to claim 1, characterized in that, The intelligent evaluation model module is used to determine the attribute discrimination ability evaluation value of the target user based on the attribute interference features; determine the rule understanding ability evaluation value and spatial matching ability evaluation value of the target user based on the task matching features; and output the intelligence evaluation result according to the attribute discrimination ability evaluation value, the rule understanding ability evaluation value and the spatial matching ability evaluation value.

10. The intelligence assessment system based on intelligent algorithms according to claim 1, characterized in that, The system also includes a corrective response analysis module, which is used to determine, in the sequence of operation events, the subsequent displacement operation of the preset geometry corresponding to the erroneous behavior after the erroneous position; Based on the termination position code of the subsequent displacement operation and the target position requirement of the preset evaluation task, it is determined whether the subsequent displacement operation is a valid correction operation. Based on the event interval between the erroneous behavior and the effective corrective action, as well as the changes in attribute encoding and location encoding corresponding to the effective corrective action, a corrective response feature is generated; The intelligent assessment model module is also used to correct the intelligence assessment results of the target user based on the corrective response characteristics.

Images