Integrated education management system based on end-to-cloud interconnection

Abstract

The application discloses a comprehensive education management system based on end-cloud interconnection and belongs to the technical field of education informatization. The system comprises the following: a data analysis module, which receives an examination data packet, extracts an identity code and a record field, and queries a history behavior label; a cloud side scheduling module, which obtains a basic information queue, recalls a past evaluation interval record from a local cache according to an education stage field and a region code; a time period discrete module, which performs time period discretization on the past evaluation interval record and generates an evaluation index quantization table; a strategy matching module, which matches the evaluation index with a correction target field, extracts a correction operation sequence from a cloud strategy library; and a tracking judgment module, which sets a tracking timer according to a terminal execution time stamp of the correction operation sequence, obtains a post-graduation state feedback field, and performs difference judgment with an expected state field. The system can realize fine-grained behavior evaluation and post-graduation closed-loop tracking.

Description

Technical Field

[0001] This invention relates to the field of educational informatization technology, specifically to a comprehensive educational management system based on edge-cloud interconnection. Background Technology

[0002] Existing integrated education management systems typically only collect students' exam scores and attendance data within the school. Assessments are limited to static records within a single school, failing to integrate cross-school historical behavioral tags and regionalized past assessment data. When generating intervention strategies, these systems often simply summarize behavioral performance over the entire semester or academic year into an average score or grade, ignoring the non-uniform distribution of students' problem behaviors over time. Due to the lack of analytical tools to discretize continuous assessment periods, it is difficult to identify densely populated areas of behavioral deterioration or high incidence. This results in matching corrective measures often based on broad overall impressions, disconnected from the actual fluctuations in students' behavior, leading to misaligned intervention timing and limited effectiveness. Furthermore, existing systems rely primarily on short-term observations by teachers within the school for post-correction effectiveness verification, lacking automated remote data collection channels for students' status after leaving the school environment. Once the corrective action sequence is completed, it is impossible to continuously obtain structured post-school behavioral feedback, let alone compare it with the expected state set in the corrective plan. This makes it impossible for education management departments to quantify the lasting effects and degree of deviation of corrective measures in off-campus settings, creating a management gap from assessment to intervention to off-campus verification. The above-mentioned solutions reveal several flaws: the assessment dimensions are singular and coarse-grained, failing to extract the sequential characteristics of behavioral events over time, resulting in a lack of ability to pinpoint high-concern periods when matching corrective goals; post-correction relies entirely on manual, fragmented on-campus tracking, lacking an automatic collection and differential assessment mechanism for post-school status, making it impossible to achieve a closed-loop intervention. There is an urgent need to address how to decompose students' historical behavioral tags and previous assessment records recalled by educational stage and region into time periods, constructing quantitative indicators reflecting behavioral density and interval characteristics based on behavioral counts within each discrete time period, in order to accurately match corrective strategies; simultaneously, a solution is needed that sets a timer trigger at the end of correction, automatically acquiring post-school feedback fields and comparing them quantitatively with the expected status, to capture off-campus behavioral changes and fill blind spots within the school. Summary of the Invention

[0003] The purpose of this invention is to provide a comprehensive education management system based on edge-cloud interconnection to overcome the problems of existing education management systems, such as coarse granularity in behavior assessment, neglect of behavioral temporal characteristics, and lack of automatic tracking and difference determination of post-rehabilitation status after leaving school, so as to achieve fine-grained time-segment quantitative assessment and off-campus closed-loop effect verification.

[0004] To achieve the above objectives, this invention provides the following technical solution: This invention provides a comprehensive education management system based on edge-cloud interconnection, including a data parsing module, a cloud-side scheduling module, a time-segmentation module, a strategy matching module, and a tracking and judgment module. The data parsing module receives review data packets submitted by the client, extracts the student's identification code and academic record fields by parsing the data structure of the review data packets, and queries historical behavior tags in the inter-school shared database based on the identification code, achieving rapid association of cross-school behavior information and providing a basis for multi-dimensional assessment. The cloud-side scheduling module obtains a queue of basic information for students under review, and retrieves past assessment interval records associated with the educational stage field and the region code of each student in the basic information queue from the local cache. It utilizes the joint index of educational stage and region to quickly locate assessment history that matches the student's growth environment, reducing scheduling latency. The time-segmentation module performs time-segmentation discretization processing on the retrieved past assessment interval records, obtaining multiple discrete assessment periods. Based on the behavior count within each discrete assessment period, it generates a quantitative table of assessment indicators for the students under review, converting continuous historical records into structured quantitative indicators, facilitating accurate location of problem periods. The strategy matching module matches each assessment indicator in the assessment indicator quantification table with a preset correction target field. Based on the successfully matched correction target field, it extracts the corresponding correction operation sequence from the cloud strategy library, realizing the automatic conversion from assessment indicators to correction strategies and avoiding subjective bias from manual intervention. The tracking and judgment module extracts the end execution timestamp from the correction operation sequence, sets a tracking timer, and obtains the post-departure status feedback field of the student to be reviewed from the cloud-side scheduling module according to the trigger time of the tracking timer. It performs a difference judgment between the post-departure status feedback field and the expected status field in the correction operation sequence, thereby continuously tracking and quantifying the correction effect and ensuring the effectiveness of the closed loop of education management.

[0005] As a technical solution of the present invention, when parsing the review data packet, the data parsing module performs header marker recognition on the review data packet, extracts the payload type identifier from the data packet encapsulation format to determine the student information segment, reads the encoded field at a fixed offset position from the student information segment and performs preamble verification, sets the encoded field that passes the preamble verification as the identity code, and then sequentially reads the course name field and grade level field from the student information segment and merges them into an academic record field, thereby ensuring the accuracy and completeness of data extraction. When querying historical behavior tags, the data parsing module splits the identity code into a school prefix code and a student serial number, uses the school prefix code as an index to locate the target school record area in the inter-school shared database, performs record matching in the target school record area using the student serial number as the key value to obtain the timeline of behavioral events in the matching records, extracts the event type marker and event severity value from the behavioral event timeline and combines them into historical behavior tags, efficiently utilizing distributed inter-school data and improving query efficiency.

[0006] As a preferred embodiment of the present invention, when the cloud-side scheduling module obtains the basic information queue of students to be reviewed, it sends a queue retrieval request to the data parsing module. The queue retrieval request includes the node number of the cloud-side scheduling module and the current load rate threshold. It receives the list of student identifiers to be reviewed returned by the data parsing module and sequentially reads the enrollment year field and current grade field corresponding to each student identifier according to the order in the list. These fields are then merged into a single student basic information entry in the basic information queue. A load-aware queue retrieval mechanism ensures the stability of data acquisition. When the cloud-side scheduling module retrieves past evaluation interval records from the local cache, it extracts the current grade field from a single student basic information entry as the education stage field and extracts the prefix region code of the school address field as the region code. Using the combination of the education stage field and the region code as the cache key, it searches for a matching index entry in the locally cached evaluation record index table. Based on the record pointer stored in the index entry, it reads the corresponding past evaluation interval record. This combination key-based fast lookup achieves low-latency retrieval from the local cache.

[0007] As a further technical solution of the present invention, when the time period discretization module performs time period discretization processing on the recalled previous evaluation interval records, it reads the evaluation start time and evaluation end time from the previous evaluation interval records, and divides the continuous time axis between the evaluation start time and the evaluation end time into multiple unit time periods according to a preset fixed time step, with the evaluation start time as the origin. For each unit time period, it reads the attendance record field and classroom participation field within that unit time period, and marks the unit time period where both the attendance record field and classroom participation field are not empty as a discrete evaluation time period, ensuring the integrity of the data included in the analysis. When generating the evaluation index quantification table, the time-period discrete module counts the number of late arrivals and absences for each discrete evaluation period, adds them together to obtain the behavior count, marks discrete evaluation periods with behavior counts exceeding a preset discrete time-period threshold as high-concern periods, arranges all high-concern periods in chronological order, writes the time interval between adjacent high-concern periods into the row record of the evaluation index quantification table, and associates the time interval field in the row record with the behavior count of the corresponding high-concern period for storage. In this way, the original behavior data is compactly quantified into a comparable index table, which facilitates subsequent strategy matching.

[0008] As another preferred technical solution of the present invention, when the strategy matching module matches each evaluation indicator in the evaluation indicator quantification table with the preset correction target field, it sequentially reads the name string and numerical range identifier of the current evaluation indicator from the evaluation indicator quantification table, extracts the matching keyword and matching numerical range for each target entry in the correction target field, calculates the character overlap rate between the name string of the current evaluation indicator and the matching keyword, and determines whether the numerical range identifier of the current evaluation indicator falls within the matching numerical range. Target entries with a character overlap rate exceeding the overlap rate threshold and a numerical range identifier falling within the matching numerical range are set as successfully matched correction target fields. The accuracy of strategy matching is improved through dual matching of characters and values. When extracting the correction operation sequence, the strategy matching module combines the target entry numbers of the successfully matched correction target fields into a strategy query vector. Using the strategy query vector as an index, it retrieves the operation step chain corresponding to the target entry number combination in the strategy mapping table of the cloud strategy library. The step instruction code is read in the order of the step sequence number of each operation step in the operation step chain. After sorting all the read step instruction codes, a correction operation sequence is formed, realizing the automatic generation of operation sequences from multi-target combinations.

[0009] As another technical solution of the present invention, when performing difference determination, the tracking and judgment module reads the feedback record timestamp and feedback content code from the post-departure status feedback field, and filters out the post-departure status feedback fields whose feedback record timestamp is later than the end execution timestamp in the correction operation sequence. The feedback content code in the filtered post-departure status feedback fields is compared character by character with the expected status field in the correction operation sequence to count the number of inconsistent characters. When the number of inconsistent characters exceeds the preset difference counting threshold, a difference flag is generated and written into the difference record table of the tracking and judgment module. By comparing the post-departure status in detail, individuals who deviate from the expectations can be identified in a timely manner, thereby improving the timeliness and accuracy of educational management intervention.

[0010] The technical effects and advantages provided by the present invention in the above technical solution are as follows: The recalled previous assessment records are discretized by time period. Using the assessment start time as the origin, the continuous time axis is divided into multiple unit time periods with a fixed time step. Attendance and classroom participation fields for each time period are read, and unit time periods where both are not empty are selected as discrete assessment time periods. For each discrete assessment time period, the counts of late arrivals and absences are calculated and summed to obtain the behavior count. Time periods with behavior counts exceeding a preset threshold are marked as high-concern time periods. All high-concern time periods are arranged chronologically, and the time interval between adjacent high-concern time periods is associated with the behavior count of those high-concern time periods and stored to form a quantitative assessment indicator table. Unlike traditional methods that only output a single comprehensive indicator for the entire assessment cycle, this discretization process outputs the temporal clustering characteristics and interval patterns of behavioral events. This allows the cloud-based strategy matching stage to accurately retrieve correction operation sequences highly adapted to behavioral fluctuation patterns from the strategy library based on the density and interval length of high-concern time periods. This avoids the generalization of correction measures caused by coarse-grained assessment and improves the timing and relevance of interventions. The rows in the evaluation index quantification table also reflect the behavioral relapse interval, thus allowing the execution rhythm of the corrective action sequence to be adapted to the actual behavioral relapse cycle. The tracking and judgment module sets a tracking timer based on the end execution timestamp in the corrective action sequence. When the timer is triggered, it acquires the post-school status feedback field of the student under review, filters out feedback content codes whose timestamps are later than the end timestamp, and performs a character-by-character comparison with the expected status field in the corrective action sequence. The number of inconsistent characters is accumulated, and when the number of inconsistent characters exceeds a preset difference counting threshold, a difference flag is generated and written to the difference record table. This tracking and comparison mechanism compensates for the deficiency of conventional education management systems in acquiring and quantifying off-campus behavioral performance. It enables the automatic capture and deviation measurement of students' behavioral changes after leaving the school environment following corrective intervention. Once a deviation from the expected state is detected, a flag is triggered, and the difference flag is recorded in the difference record table. This facilitates the identification of cases of corrective failure and timely triggering of supplementary interventions, thus forming a coherent corrective verification loop inside and outside the school. This breaks the limitations of relying solely on manual observation within the school, enhancing the continuity and responsiveness of education management. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0012] Figure 1 This is a schematic diagram of the structure of a comprehensive education management system based on edge-cloud interconnection; Figure 2This is a flowchart of the cloud-side scheduling module acquiring basic information queues and recalling records from previous evaluation intervals. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] See Figure 1 This invention provides a comprehensive education management system based on edge-cloud interconnection, including a data parsing module, a cloud-side scheduling module, a time-segmentation module, a strategy matching module, and a tracking and judgment module. The data parsing module receives review data packets submitted by the client, extracts the student's identification code and academic record field by parsing the data structure of the review data packet, and queries historical behavior tags in the inter-school shared database based on the identification code. The cloud-side scheduling module obtains a queue of basic information for students under review, and retrieves past assessment interval records associated with the educational stage field and region code of each student from the local cache. The time-segmentation module performs time-segmentation discretization processing on the retrieved past assessment interval records to obtain multiple discrete assessment periods, and generates a quantitative table of assessment indicators for the students under review based on the behavior count within each discrete assessment period. The strategy matching module matches each assessment indicator in the quantitative table of assessment indicators with a preset remedial target field, and extracts the corresponding remedial operation sequence from the cloud strategy library based on the successfully matched remedial target field. The tracking and judgment module extracts the end execution timestamp in the correction operation sequence, sets a tracking timer, and obtains the post-departure status feedback field of the student to be reviewed from the cloud-side scheduling module according to the trigger time of the tracking timer. The module then performs a difference judgment between the post-departure status feedback field and the expected status field in the correction operation sequence.

[0015] Example 1: In practice, after receiving the review data packet submitted by the client, the data parsing module performs header marking and identification on the review data packet. The review data packet has a predefined data packet encapsulation format, which includes a fixed-length header area containing a payload type identifier field. The data parsing module reads the binary value of the payload type identifier field in the header area and compares it with a preset payload type mapping table to determine the student information segment of the review data packet. The payload type mapping table stores the correspondence between different binary values ​​and student information segment formats. When the binary value of the payload type identifier is a first preset value, the student information segment adopts the first segment structure; when the binary value of the payload type identifier is a second preset value, the student information segment adopts the second segment structure. Based on the determined student information segment structure, the data parsing module extracts all bytes of the student information segment from the review data packet.

[0016] When retrieving the encoded field at a fixed offset from the student information segment, the data parsing module locates the storage position of the encoded field by adding a preset fixed offset to the starting address of the student information segment. The encoded field consists of several bytes. The data parsing module reads all bytes of the encoded field and performs a preamble verification. The preamble verification process is as follows: the data parsing module pre-stores a valid preamble sequence and compares the first few bytes of the encoded field byte-by-byte with the valid preamble sequence. If the comparison matches, the data parsing module determines that the preamble verification has passed and sets the remaining part of the encoded field after removing the valid preamble sequence as the identification code. If the comparison does not match, the data parsing module discards the review data packet and generates a parsing failure notification.

[0017] Following the identification code, the data parsing module sequentially reads the course name field and grade level field from the student information segment. Immediately following the identification code field in the student information segment is the course name length field. The data parsing module reads the value of the course name length field to determine the byte length of the course name field, and then reads the content of the course name field from the student information segment according to that byte length. Following the course name field is the grade level field, which uses fixed-length encoding. The data parsing module reads a fixed-length byte from the end of the course name field to form the grade level field. The data parsing module concatenates the content of the course name field and the content of the grade level field, separated by a delimiter, to form the academic record field.

[0018] In some implementations, the data parsing module queries the inter-school shared database for historical behavior tags based on the identification code. The identification code is a composite code containing a school prefix and a student serial number, where the school prefix is ​​located in the high-order bits and the student serial number is located in the low-order bits. The data parsing module splits the identification code into the school prefix and student serial number as follows: a predefined bit length for the school prefix is ​​used, and a bit string of the predefined bit length is extracted from the highest bit of the identification code to form the school prefix; the remaining bits in the identification code form the student serial number.

[0019] The data parsing module uses the school prefix code as an index to locate the target school's record area in the inter-school shared database. The inter-school shared database is a distributed database, partitioned according to the school prefix code, with each school prefix code's corresponding data stored in an independent target school record area. The data parsing module sends a query request containing the school prefix code to the inter-school shared database, which then returns the storage address of the target school's record area based on the school prefix code.

[0020] Within the target school's record area, the data parsing module performs record matching using the student's serial number as the key. The target school's record area employs a key-value storage structure, where the key is the complete student's serial number, and the value is the corresponding student's behavioral event data. The data parsing module constructs a key-value query command containing the student's serial number, searching for records in the target school's record area whose key is exactly the same as the student's serial number. When a matching record is found, the inter-school shared database returns the timeline of the behavioral events from the matching record.

[0021] The behavioral event timeline is a sequence of event entries arranged chronologically. Each event entry contains an event type tag field and an event severity value field. The data parsing module extracts the event type tag and event severity value for each event entry from the behavioral event timeline. The event type tag is an enumerated value used to identify the category of the behavioral event; the event severity value is a numerical value used to represent the severity level of the behavioral event. The data parsing module combines the extracted event type tag and event severity value by converting the enumerated value of the event type tag into a string and concatenating it with the numerical string of the severity value using an underscore. The combined result forms a historical behavioral label. If the behavioral event timeline contains multiple event entries, the data parsing module generates the same number of historical behavioral labels as the number of event entries. Multiple historical behavioral labels constitute a historical behavioral label set.

[0022] Example 2: In specific implementation, please refer to Figure 2When the cloud-side scheduling module obtains the queue of basic information of students to be reviewed, it sends a queue retrieval request to the data parsing module. The queue retrieval request is a data packet encapsulated according to the communication protocol agreed upon between the cloud-side scheduling module and the data parsing module. The communication protocol adopts a TCP-based binary protocol format. The cloud-side scheduling module reads the values ​​of its own node number register and current load rate statistics register, and fills the node number and current load rate threshold into the corresponding fields of the queue retrieval request packet. The node number is a unique identifier for the cloud-side scheduling module in the cluster, uniformly assigned by the cluster management node configured during deployment; the current load rate threshold is a preset value read by the cloud-side scheduling module from the configuration parameter file. The preset value is a floating-point number between 0.1 and 0.9. The configuration parameter file is set by the operations and maintenance personnel during the system initialization phase according to the hardware processing capabilities of the node where the cloud-side scheduling module is located.

[0023] The cloud-side scheduling module sends queue retrieval requests to the node where the data parsing module resides via the internal message bus. Upon receiving the queue retrieval request, the data parsing module extracts the node number and the current load rate threshold from the message. Internally, the data parsing module maintains a complete list of student identifiers awaiting review. It queries the load balancing mapping table based on the node number, which stores the correspondence between node numbers and the index of the last assigned identifier. Starting from the position after the last assigned identifier index, the data parsing module sequentially selects student identifiers awaiting review, with the number selected determined by the current load rate threshold. The number of identifiers selected... The calculation method is as follows: in, This represents the current load rate threshold extracted from requests pulled from the queue. This represents a constant indicating the maximum number of items that can be retrieved in a single operation, preset by the data parsing module. This indicates the floor function. The value range is an integer from 50 to 500, and is set in the configuration file of the data parsing module. When the calculated value... When the value is less than 1, one student ID to be reviewed is selected by default.

[0024] The data parsing module selects the student identifiers to be reviewed and compiles them into a list of student identifiers to be reviewed, then updates the last assigned identifier index corresponding to the node number in the load balancing mapping table. The data parsing module then returns the list of student identifiers to be reviewed to the cloud-side scheduling module. Upon receiving the list of student identifiers to be reviewed from the data parsing module, the cloud-side scheduling module temporarily stores the list in a memory buffer.

[0025] The cloud-side scheduling module processes each student identifier to be reviewed sequentially according to the order in the list. This order is determined by the order in which the data parsing module returns the list. The cloud-side scheduling module maintains a processing pointer, initially pointing to the first element of the list. It reads the student identifier pointed to by the processing pointer and sends an information query request containing that identifier to the student information database. The student information database stores the mapping between student identifiers and the fields of enrollment year and current grade. The cloud-side scheduling module receives the enrollment year and current grade fields returned by the student information database and merges them. The merging method involves concatenating the strings of the enrollment year and current grade fields with a vertical bar, forming a single student basic information entry. The cloud-side scheduling module appends this single student basic information entry to the end of the basic information queue and moves the processing pointer one position to the right. This process is repeated until the processing pointer exceeds the end of the list of student identifiers to be reviewed, at which point the cloud-side scheduling module completes the acquisition of the basic information queue.

[0026] In some implementations, after the cloud-side scheduling module completes the acquisition of the basic information queue, it retrieves past assessment interval records from the local cache according to the education stage field and the region code of each student to be reviewed in the basic information queue. The cloud-side scheduling module traverses individual student basic information entries in the basic information queue. For the currently traversed individual student basic information entry, the cloud-side scheduling module extracts the current grade field as the education stage field. The part after the vertical bar character in the individual student basic information entry is the current grade field, and the cloud-side scheduling module extracts the string of the current grade field by locating the position of the vertical bar character. The cloud-side scheduling module extracts the prefix region code of the school address field as the region code from the individual student basic information entry. In addition to the enrollment year field and the current grade field, the individual student basic information entry also contains a school address field, which is a complete administrative division address code. The cloud-side scheduling module extracts the first six characters of the school address field as the prefix region code, which corresponds to the prefecture-level administrative division code.

[0027] The cloud-side scheduling module uses the combination of the education stage field and the region code as the cache key. The combination method involves concatenating the string of the education stage field, the underscore character, and the string of the region code sequentially to obtain the complete cache key string. The cloud-side scheduling module uses the cache key string to search for a matching index entry in the local cached evaluation record index table. The local cached evaluation record index table is a hash table structure, where the key is the cache key string and the value is the index entry. Each index entry stores a record pointer and a record length. The record pointer is the starting memory address of the previous evaluation interval record in the local cache record storage area, and the record length is the number of bytes occupied by the previous evaluation interval record.

[0028] The cloud-side scheduling module calculates the hash value of the cached key string, traverses and compares it in the hash bucket corresponding to the evaluation record index table, and locates the index entry whose key value is exactly the same as the cached key string. The cloud-side scheduling module reads the record pointer and record length from the index entry. Based on the record pointer, the cloud-side scheduling module locates the physical address in the local cache record storage area, reads consecutive byte blocks from the record storage area according to the record length, and deserializes the read consecutive byte blocks into a data structure of past evaluation interval records. The data structure of past evaluation interval records includes an evaluation start time field, an evaluation end time field, and a list of attendance participation records for a unit time period field. The cloud-side scheduling module returns the deserialized past evaluation interval records to the caller.

[0029] Example 3: In practice, the time-period discrete module receives recalled past evaluation interval records from the cloud-side scheduling module. The data structure of these records includes an evaluation start time field, an evaluation end time field, and a list of attendance participation records for each time period. The time-period discrete module parses these records, reading the values ​​of the evaluation start time and evaluation end time fields. Both the evaluation start time and evaluation end time fields are stored in UNIX timestamp format, accurate to the second.

[0030] The time-segment discrete module uses the evaluation start time as the origin and divides the continuous time axis between the evaluation start time and the evaluation end time into multiple unit time segments according to a preset fixed time step. The preset fixed time step is a configuration item in the system configuration parameter table, named "Time-Segment Discrete Step Parameter." The Time-Segment Discrete Step Parameter is in seconds and is set by the system administrator during the initialization phase according to the teaching management granularity requirements. The Time-Segment Discrete Step Parameter takes a positive integer value. The time-segment discrete module reads the value of the Time-Segment Discrete Step Parameter from the configuration parameter table and denotes it as... The process of dividing a time segment into units is as follows: Set the current starting time for the segmentation. Initial command Equals the evaluation start time; calculates the current segmentation end time. ;like If the time is less than or equal to the end time of the evaluation, a unit time period is generated, and the start and end times of the unit time period are... to And assign a unit time period number to that unit time period; Continue calculating the next time unit; if If the time exceeds the end time of the evaluation, then the start and end times of the last unit time period will be set to... The evaluation ends at the designated time, and the segmentation ends there. All generated unit time periods constitute a set of unit time periods.

[0031] For each segmented time period, the time period discrete module reads the attendance record field and classroom participation field within that time period. The list of attendance and participation records for each time period in the previous evaluation interval records is a linear list consisting of multiple record items. Each record item corresponds to an observation at a specific time point and includes a timestamp, an attendance record field, and a classroom participation field. The timestamp identifies the time of this observation. The attendance record field is an enumeration type, with values ​​including a first value representing attendance, a second value representing lateness, and a third value representing absence. The classroom participation field is a numeric type, with values ​​ranging from 1 to 5, where 1 represents the lowest participation and 5 represents the highest participation.

[0032] The time-segment discrete module iterates through the list of attendance and participation records for each time segment, filtering out all records whose timestamps fall within the start and end time range of the current time segment. From the filtered records, the module reads the enumerated value of the attendance record field and the value of the classroom participation field for each record. The module then determines whether both the attendance record field and the classroom participation field are non-empty. The criteria for non-emptiness are: the value of the attendance record field is not a predefined empty enumerated value, and the value of the classroom participation field is within the valid range of 1 to 5. If at least one of the filtered records satisfies the condition that both the attendance record field and the classroom participation field are non-empty, the module marks the time segment as a discrete assessment period and assigns a discrete assessment period identifier. If none of the filtered records meet the conditions, the time segment is not marked.

[0033] In some implementations, after marking the discrete assessment periods, the time-discretion module counts the behaviors for each discrete assessment period. The time-discretion module obtains all valid records within the discrete assessment period corresponding to the current discrete assessment period identifier, and reads the attendance record field of each valid record. The time-discretion module maintains two counters: a late arrival event count and an absence event count, both initially set to 0. For each valid record, if the value of the attendance record field represents the second value indicating lateness, the late arrival event count is incremented by 1; if the value of the attendance record field represents the third value indicating absence, the absence event count is incremented by 1. The time-discretion module calculates the behavior count for the discrete assessment period as the sum of the late arrival event count and the absence event count.

[0034] The time-segment discrete module reads the preset discrete time-segment threshold from the system configuration parameter table. The preset discrete time-segment threshold is a critical value used to determine whether a discrete assessment period requires high attention; it is determined by the teaching management strategy and written into the configuration parameter table, and is a positive integer. The time-segment discrete module compares the behavior count for each discrete assessment period with the preset discrete time-segment threshold. When the behavior count for a discrete assessment period is greater than the preset discrete time-segment threshold, the time-segment discrete module marks that discrete assessment period as a high-attention period and generates a high-attention period record. The high-attention period record includes the start time, end time, and behavior count value of the high-attention period. The start time is the start time of the corresponding discrete assessment period, the end time is the end time, and the behavior count is the behavior count of the corresponding discrete assessment period.

[0035] The time-period discretization module sorts all high-concern time-period records in ascending order of their start times, resulting in a high-concern time-period sequence. This module creates a quantitative evaluation index table, a data table containing multiple rows, each describing a time interval and its associated behavior count. The time-period discretization module iterates through the high-concern time-period sequence, processing the first high-concern time-period record in the sequence... Records of high-attention periods, among which For an integer index starting from 1, perform the following operation: If If the current high-attention period is recorded as the first high-attention period in the sequence, then the time interval between adjacent high-attention periods is recorded as 0; if Then read the first The high-attention period records the end time of the high-attention period, using the first high-attention period as the record. The high-attention period record starts at the time of the high-attention period minus the first high-attention period. The high-attention period records the end time of each high-attention period, and the time interval value is obtained in seconds.

[0036] The time-discrete module adds a new record to the evaluation index quantification table, writes the calculated time interval value into the time interval field of this record, and writes the row count field into the first row. The high-attention period behavior count is recorded in each high-attention period record. The final evaluation index quantification table contains a number of rows equal to the number of high-attention period records in the high-attention period sequence. Each row stores the time interval field in association with the corresponding high-attention period behavior count.

[0037] Example 4: In practical implementation, the strategy matching module obtains the evaluation index quantification table from the time-discretion module. The evaluation index quantification table is a data table containing multiple rows of records. Each row contains a name string field and a value interval identifier field. The name string field stores the name of the evaluation index, and the value interval identifier field stores the interval identifier of the quantified value of the evaluation index. The strategy matching module maintains a row pointer in the evaluation index quantification table, initially pointing to the first row. The strategy matching module reads the current row pointed to by the row pointer, extracts the value of the name string field from the current row as the name string of the current evaluation index, and extracts the value of the value interval identifier field from the current row as the value interval identifier of the current evaluation index. The value of the name string field is a string constant composed of letters and underscores, including the strings "high_attention_interval" and "behavior_count_level". The value of the value interval identifier field is represented by an integer range enclosed in square brackets, containing a lower bound integer value and an upper bound integer value, separated by a comma.

[0038] The strategy matching module loads a pre-defined set of treatment target fields from the system configuration store. This set contains multiple target entries, each consisting of a matching keyword field, a matching value range field, and a target entry number field. The matching keyword field stores the string used to match the evaluation indicator name string, including the strings "high_attention_interval" and "behavior_count_level". The matching value range field stores an integer range enclosed in square brackets, containing both the lower and upper bounds of the range. The target entry number field stores a four-digit integer value that uniquely identifies the target entry, ranging from 0001 to 9999.

[0039] The strategy matching module iterates and matches each target entry in the set of treatment target fields. For the currently encountered target entry, the strategy matching module extracts the matching keywords and matching value range of the target entry. The strategy matching module calculates the character overlap rate between the name string of the current evaluation indicator and the matching keywords of the target entry. The character overlap rate is calculated as follows: align the name string of the current evaluation indicator with the matching keywords of the target entry, and compare the characters at the same positions of the two strings one by one to see if they are the same; count the number of positions where the characters are the same in the two strings at the same positions as the common character count; obtain the total character length of the name string of the current evaluation indicator and the total character length of the matching keywords of the target entry; divide twice the number of common characters by the sum of the total character length of the name string of the current evaluation indicator and the total character length of the matching keywords of the target entry to obtain the character overlap rate.

[0040] The character overlap rate is expressed by the formula: in, This represents the calculated character overlap rate. A floating-point number whose value is between 0 and 1; This indicates the number of positions where the name string of the current evaluation indicator and the matching keyword of the target entry have the same character position. It is a non-negative integer; This represents the total length of the string representing the name of the current evaluation indicator. It is a positive integer; This indicates the total length of the matching keywords in the target entry. It is a positive integer.

[0041] The strategy matching module reads the overlap rate threshold from the system configuration parameter table. The overlap rate threshold is a critical proportion used to determine whether a name string match is successful; the threshold value is 0.75. The basis for setting the overlap rate threshold to 0.75 is that when the number of common characters accounts for at least 75% of the sum of the total lengths of both strings, there is a high degree of similarity between the name strings, which can eliminate false matches caused by spelling differences or different suffixes. The strategy matching module will then calculate the... Compare with the overlap rate threshold.

[0042] The strategy matching module determines whether the numerical range identifier of the current evaluation indicator falls within the matching numerical range of the target item. The determination method is: extract the lower bound integer value of the range identifier from the numerical range identifier of the current evaluation indicator. and the integer value of the upper bound of the interval Extract the lower bound integer value from the matching numerical range of the target entry. and the upper bound of the range of integer values ;when Greater than or equal to and Less than or equal to If the value range indicator falls within the range of the matching value, then the value range indicator is determined to fall outside the range of the matching value; otherwise, the value range indicator is determined not to fall outside the range of the matching value.

[0043] The strategy matching module sets target entries that simultaneously meet both conditions—character overlap rate exceeding the overlap rate threshold and numerical range identifier falling within the matching numerical range—as successfully matched treatment target fields. If either condition is not met, the current target entry does not match the current evaluation index, and the strategy matching module continues to traverse the next target entry. After completing the traversal of all target entries in the treatment target field set, the strategy matching module moves the row pointer to the next row and continues to perform the above matching process on the next row of the evaluation index quantification table until all rows of the evaluation index quantification table have been processed.

[0044] In some implementations, the strategy matching module collects the target entry numbers of all successfully matched treatment target fields and combines these target entry numbers into a strategy query vector. The strategy query vector is combined as follows: the target entry number of each successfully matched treatment target field is converted into a string of length 4. If multiple successfully matched treatment target fields exist, these multiple 4-length strings are concatenated end-to-end in ascending order of target entry number value to obtain the strategy query vector. If only one successfully matched treatment target field exists, the strategy query vector is the 4-length string corresponding to that treatment target field.

[0045] The policy matching module uses the policy query vector as an index to retrieve the corresponding operation step chain in the policy mapping table of the cloud policy library. The cloud policy library is a key-value database deployed on a cloud server, and the policy mapping table is a hash mapping table within it. The key of the policy mapping table is the policy query vector, and the value is the storage identifier of the operation step chain. The policy matching module sends a query request containing the policy query vector to the cloud policy library, which then searches the policy mapping table for an entry whose key exactly matches the policy query vector. When a matching entry is found, the cloud policy library returns the storage identifier of the operation step chain, and the policy matching module reads the complete data of the operation step chain based on this identifier.

[0046] The operation step chain is a sequence of step instruction codes arranged in ascending order of step number. Each step instruction code consists of a step number field and an instruction content field. The step number field is a positive integer starting from 1 and incrementing, while the instruction content field is a string encoding representing the specific orthodontic operation. The strategy matching module reads the corresponding instruction content fields sequentially according to the ascending order of the step number field of each step instruction code in the operation step chain. It then concatenates the string encodings of all read instruction content fields in ascending order of step number, inserting a semicolon as a separator between every two instruction content fields. The resulting concatenated string constitutes the orthodontic operation sequence.

[0047] Example 5: In practice, the tracking and determination module obtains the treatment operation sequence from the strategy matching module. The treatment operation sequence is a string composed of multiple operation step instruction codes, separated by semicolons. The last operation step instruction code in the sequence contains an end-execution timestamp field, formatted as "end_timestamp=UNIX timestamp value". The tracking and determination module parses the treatment operation sequence string, locates the content after the last semicolon, searches for the substring "end_timestamp=" within that content, and extracts the integer value after the equals sign as the end-execution timestamp in the treatment operation sequence.

[0048] The tracing and determination module sets the tracing timer. This tracing timer is a software timer, created by the tracing and determination module using the operating system's timer interface. The tracing and determination module sets the trigger timer's value to the end execution timestamp plus a preset tracing offset duration. The preset tracing offset duration is a configuration parameter stored in the system configuration parameter table, named "Tracing Offset Duration Parameter," in seconds, and its value is determined by educational management policies. The tracing and determination module reads the value of the tracing offset duration parameter from the system configuration parameter table, adds the end execution timestamp to the tracing offset duration parameter value, and obtains the absolute UNIX timestamp value of the trigger time. The tracing and determination module writes the trigger time into the tracing timer's expiration time register and starts the tracing timer's countdown.

[0049] The tracking timer generates an interrupt signal when the system clock reaches its trigger point, and the interrupt handling routine of the tracking determination module receives the interrupt signal. At the trigger point of the tracking timer, the tracking determination module sends a request to the cloud-side scheduling module to obtain the post-departure status feedback field. The request message contains the identification code of the student to be reviewed, which is pre-obtained by the tracking determination module from the data parsing module and stored in the tracking task context.

[0050] After receiving the request, the cloud-based scheduling module queries the student departure status database and extracts the departure status feedback field associated with the student's identification code. The departure status feedback field is a variable-length string, using a key-value pair encoding format, with entries separated by semicolons, and the encoding format is "key=value". The cloud-based scheduling module returns the departure status feedback field string to the tracking and judgment module.

[0051] The tracking and determination module receives the post-departure status feedback field returned by the cloud-side scheduling module and parses the string. The module locates the first entry in the string with the key "timestamp," extracts the substring after the equals sign, and converts it to a UNIX timestamp to obtain the feedback record timestamp. The module then locates the entry in the string with the key "content_code," extracts the substring after the equals sign as the feedback content code. The feedback content code is a string composed of letters and numbers, representing the specific category code of the post-departure status.

[0052] The tracking and judgment module filters the feedback record timestamps based on the final execution timestamp in the correction operation sequence. The comparison method is as follows: if the feedback record timestamp is greater than the final execution timestamp, the post-departure status feedback field is determined to be a new status record generated after the final execution timestamp and is retained; if the feedback record timestamp is less than or equal to the final execution timestamp, the post-departure status feedback field is discarded.

[0053] For the selected post-departure status feedback fields, the tracking and judgment module compares the encoded feedback content character by character with the expected status field in the corrective operation sequence. The expected status field is stored in the corrective operation sequence as a string of characters prefixed with "expected_state=". The tracking and judgment module first searches for the substring "expected_state=" in the corrective operation sequence string and extracts the string after the equals sign as the expected status field. The tracking and judgment module compares the characters at the same index position of the encoded feedback content string and the expected status field string one by one, maintaining an inconsistent character counter, initially set to 0. For cases where the string lengths are misaligned, the longer portion is entirely counted as an inconsistent character. During the traversal, whenever the characters at the same index position of two strings are not equal, the inconsistent character counter is incremented by 1. After the count is completed, the number of inconsistent characters is obtained.

[0054] The tracking and judgment module reads the preset difference count threshold from the system configuration parameter table. The preset difference count threshold is a critical integer value used to determine whether a state difference is significant; it is defined by the educational assessment standards and written into the configuration parameter table. The tracking and judgment module compares the count of inconsistent characters with the preset difference count threshold. When the number of inconsistent characters exceeds the preset difference count threshold, the tracking and judgment module generates a difference flag, which is a Boolean true value. When the number of inconsistent characters is less than or equal to the preset difference count threshold, the tracking and judgment module does not generate a difference flag, or generates a Boolean false value.

[0055] The tracking and judgment module writes the true or false value of the difference flag bit into the difference record table. The difference record table is a local data table maintained by the tracking and judgment module, containing an identification code column, an end-execution timestamp column, a difference flag bit column, and a record-writing timestamp column. When the tracking and judgment module adds a new record to the difference record table, it writes the student's identification code into the identification code column, the end-execution timestamp into the end-execution timestamp column, converts the Boolean value of the difference flag bit to an integer and writes it into the difference flag bit column, and writes the current system time's UNIX timestamp into the timestamp column. The difference record table is used for querying and statistical analysis of subsequent educational intervention measures.

[0056] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A comprehensive education management system based on edge-cloud interconnection, characterized in that: include: The data parsing module receives the review data packet submitted by the client, extracts the student's identification code and academic record field by parsing the data structure of the review data packet, and queries the historical behavior tags in the inter-school shared database based on the identification code; The cloud-side scheduling module obtains the basic information queue of students to be reviewed, and retrieves the previous assessment interval records associated with the education stage field and the region code of each student to be reviewed from the local cache according to the education stage field and the region code of each student to be reviewed in the basic information queue. The time-period discretization module performs time-period discretization processing on the recalled previous evaluation interval records to obtain multiple discrete evaluation periods, and generates a quantitative table of evaluation indicators for the students to be reviewed based on the behavior count within each discrete evaluation period. The strategy matching module matches each evaluation indicator in the evaluation indicator quantification table with the preset treatment target field, and extracts the corresponding treatment operation sequence from the cloud strategy library based on the successfully matched treatment target field. The tracking and judgment module extracts the end execution timestamp in the correction operation sequence, sets a tracking timer, and obtains the post-departure status feedback field of the student to be reviewed from the cloud-side scheduling module according to the trigger time of the tracking timer. It then performs a difference judgment between the post-departure status feedback field and the expected status field in the correction operation sequence.

2. The integrated education management system based on edge-cloud interconnection according to claim 1, characterized in that, The data parsing module extracts the student's identification code and academic record fields by parsing the data structure of the review data packet, including: The header marker of the review data packet is identified, the payload type identifier in the data packet encapsulation format is extracted, and the student information segment of the review data packet is determined based on the payload type identifier. Read the coded field at a fixed offset position from the student information segment, perform preamble verification on the coded field, and set the coded field that passes the preamble verification as the identity code; After the identification code, the course name field and grade level field are read sequentially from the student information section, and then the course name field and grade level field are merged into the academic record field.

3. The integrated education management system based on edge-cloud interconnection according to claim 2, characterized in that, The data parsing module queries historical behavior tags in the inter-school shared database based on the identity verification code, including: The identification code is split into a school prefix code and a student serial number. The school prefix code is used as an index to locate the target school record area in the inter-school shared database. Within the target school's record area, records are matched using the student's serial number as the key to obtain the timeline of behavioral events in the matched records; Extract event type tags and event severity values ​​from the timeline of behavioral events, and combine the event type tags and event severity values ​​into historical behavior labels.

4. The integrated education management system based on edge-cloud interconnection according to claim 1, characterized in that, The cloud-side scheduling module obtains a queue of basic information about students to be reviewed, including: Send a queue retrieval request to the data parsing module. The queue retrieval request contains the node number of the cloud-side scheduling module and the current load rate threshold. The data parsing module receives a list of student identifiers to be reviewed based on the node number and the current load rate threshold. According to the order of the student identifiers to be reviewed, the enrollment year field and current grade field corresponding to each student identifier to be reviewed are read in sequence, and the enrollment year field and current grade field are merged into a single student basic information entry in the basic information queue.

5. The integrated education management system based on edge-cloud interconnection according to claim 4, characterized in that, The cloud-side scheduling module retrieves previous assessment interval records associated with the educational stage field and the region code of each student to be reviewed from the local cache, including: Extract the current grade field from a single student's basic information entry as the education stage field, and extract the prefix area code of the school address field from a single student's basic information entry as the area code; Using the combination of the education stage field and the region code as the cache key, search for a matching index entry in the locally cached evaluation record index table; Based on the record pointers stored in the index entries, read the previous evaluation interval records corresponding to the composite keys from the record storage area of ​​the local cache.

6. The integrated education management system based on edge-cloud interconnection according to claim 1, characterized in that, The time-period discretization module performs time-period discretization processing on the recalled previous evaluation interval records to obtain multiple discrete evaluation time periods, including: Read the evaluation start time and evaluation end time from the previous evaluation interval records. Using the evaluation start time as the origin, divide the continuous time axis between the evaluation start time and the evaluation end time into multiple unit time periods according to the preset fixed time step. For each time period, read the attendance record field and class participation field within that time period; Mark the time period when both the attendance record field and the class participation field are not empty as a discrete assessment time period.

7. The integrated education management system based on edge-cloud interconnection according to claim 6, characterized in that, The time-period discrete module generates a quantitative table of evaluation indicators for the students to be reviewed based on the behavior counts within each discrete evaluation period, including: For each discrete assessment period, the count of late arrivals and the count of absences within that period are counted, and the count of late arrivals and the count of absences are added together to obtain the behavior count. Discrete evaluation periods whose behavior counts exceed a preset discrete time period threshold are marked as high-concern periods; Arrange all high-attention periods in chronological order, write the time interval between adjacent high-attention periods into the row record of the evaluation index quantification table, and associate the time interval field in the row record with the behavior count of the corresponding high-attention period.

8. The integrated education management system based on edge-cloud interconnection according to claim 1, characterized in that, The strategy matching module matches each assessment indicator in the assessment indicator quantification table with a preset treatment target field, including: Read the name string and numerical range identifier of the current evaluation indicator sequentially from the evaluation indicator quantification table; For each target entry in the treatment target field, extract the matching keywords and matching value range of the target entry; Calculate the character overlap rate between the name string of the current evaluation indicator and the matching keyword of the target entry, and determine whether the numerical range identifier of the current evaluation indicator falls within the matching numerical range of the target entry. Target entries with a character overlap rate exceeding the overlap rate threshold and a numerical range identifier falling within the matching numerical range are set as successfully matched correction target fields.

9. The integrated education management system based on edge-cloud interconnection according to claim 8, characterized in that, The strategy matching module extracts the corresponding treatment operation sequence from the cloud strategy library based on the successfully matched treatment target field, including: Combine the target entry numbers of the successfully matched treatment target fields into a strategy query vector; Using the policy query vector as an index, retrieve the chain of operation steps corresponding to the combination of target entry numbers in the policy mapping table of the cloud policy library; Read the step instruction codes in the order of the step number of each operation step in the operation step chain, and sort all the read step instruction codes to form an orthodontic operation sequence.

10. The integrated education management system based on edge-cloud interconnection according to claim 1, characterized in that, The tracking and determination module performs a difference determination between the post-departure status feedback field and the expected status field in the correction operation sequence, including: Read the feedback record timestamp and feedback content code from the post-departure status feedback field, and filter out the post-departure status feedback fields whose feedback record timestamp is later than the final execution timestamp in the corrective operation sequence; The feedback content codes in the selected post-departure status feedback fields are compared character by character with the expected status fields in the correction operation sequence, and the number of inconsistent characters is counted. When the number of inconsistent characters exceeds the preset difference count threshold, a difference flag is generated and written to the difference record table of the tracking and judgment module.

Images