Space-time big data-based academic supply and demand dynamic sensing method and system

Abstract

The invention relates to the technical field of information retrieval, in particular to an academic supply and demand dynamic sensing method and system based on space-time big data, and the method comprises the following steps: obtaining the path length of a source and a target college, screening migration direction stable labels, recognizing a trend consistent region, evaluating the local transfer capacity, and fusing a high-frequency path region. And extracting behavior track data, performing sequence comparison on the associated nodes, and generating a prediction chain space offset mark set. According to the method, space label units with stable migration characteristics are effectively distinguished by processing enrollment information and population migration trends and constructing a path length filtering and direction section repeating mechanism, and supply and demand migration origin areas are accurately locked in combination with change trends reflected by continuous direction consistency; spatial aggregation judgment is enhanced through a statistical means of a neighborhood capacity proportion and path intersection frequency, the sensitivity and dynamic response capability of a spatial prediction structure are improved, and active sensing and intervention prompting of a supply and demand change trend under multiple space-time dimensions are realized.

Description

Technical Field

[0001] This invention relates to the field of information retrieval technology, and in particular to a dynamic sensing method and system for degree supply and demand based on spatiotemporal big data. Background Technology

[0002] Information retrieval technology involves the collection, analysis, matching, and query processing of massive amounts of structured and unstructured data. Its core aspects include information organization, index building, query parsing, semantic understanding, and relevance ranking. It is widely used in search engines, database systems, knowledge graphs, and intelligent recommendation systems, and is a crucial foundational technology supporting large-scale data analysis and knowledge discovery. Traditional methods for predicting degree demand utilize enrollment data, demographic information, and regional economic and social development data. Through query processing and information matching techniques in information retrieval, combined with time-series trend analysis in data mining, it statistically estimates the supply and demand of degree programs in different regions and time periods. This is accomplished using educational statistical information retrieval and rule-based supply and demand analysis.

[0003] Existing technologies mainly rely on structured queries and rule-based supply and demand statistics, lacking in-depth identification of the dynamic continuity of population migration directions and paths. When faced with ambiguous migration trends or unclear boundary transitions between regions, problems such as generalized label recognition or discontinuous predicted distribution are likely to occur. Especially under conditions with a large time span or frequent interactions between regions, the phenomenon of delayed early warning or imbalanced regulation is likely to occur in the context of complex migration. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a dynamic sensing method and system for degree supply and demand based on spatiotemporal big data.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a dynamic sensing method for degree supply and demand based on spatiotemporal big data, comprising the following steps: S1: Obtain the student's origin location, target university coordinates and enrollment year number, calculate the spatial path length between the origin and target university, filter the label units with path lengths within an acceptable range, and generate a migration-stable label list; S2: Call the migration stability label list to obtain the direction of change in the number of students admitted to adjacent units, mark the units that show the same direction continuously, take the unit number with the same trend at the boundary as the origin region of supply and demand migration, and generate a set of spatial prediction starting units. S3: Call the spatial prediction starting point unit set, collect the student capacity and enrollment data in the surrounding two layers of grid, calculate the proportion of transferable capacity in the overall capacity within the local area, record the population migration trend in the area, and generate a predicted supply and demand aggregation structure block; S4: Based on the geographic number and time record of the spatial unit in the predicted supply and demand aggregation structure block, extract the corresponding enrollment number, migration direction trajectory and education acceptance quantity information, record the type and start and end time of each data call in chronological order, identify the reading trajectory of continuous data, and generate a predictive variable time behavior cue map.

[0006] As a further embodiment of the present invention, the migration stability label list includes label number, number of repeating direction segments, and direction consistency identifier; the spatial prediction starting point unit set includes trend consistency boundary unit number, direction continuity marker, and initial supply and demand characteristic value; the prediction supply and demand aggregation structure block includes path intersection frequency, transfer capacity ratio, and spatial fusion unit number; and the prediction variable time behavior cue graph includes variable call type, variable start and end time, and data trajectory sequence.

[0007] As a further aspect of the present invention, the step of obtaining the migration stability tag list specifically includes: S111: Obtain the student's origin location, target school coordinates and enrollment year number, calculate the spatial path length between the student's origin and the corresponding target school, determine whether it is less than the path length filtering threshold, and extract the tag unit number that meets the condition after filtering to obtain the path filtering tag number set; S112: Call the path to filter the tag number set, obtain the corresponding source coordinates in three consecutive enrollment years, compare the latitude and longitude change directions between the source coordinates in consecutive years, count the consistency of the migration direction of the tag unit source within three years, and generate a list of tags with consistent migration direction. S113: Call the migration direction consistent label list, count whether the source migration direction segments in consecutive years are consistent for the label unit number, record the label number whose consecutive direction consistency exceeds the set direction repetition threshold as stable unit number, calculate the direction repetition trend degree of the label unit, and obtain the migration stable label list based on the case where the direction repetition trend degree exceeds the direction repetition threshold.

[0008] As a further aspect of the present invention, the step of obtaining the spatial prediction starting point unit set specifically includes: S211: Call the enrollment data of the units in the migration stability label list, select adjacent units in sequence to construct a comparison combination, perform difference calculation by combining the enrollment values ​​of each pair of adjacent units, and determine the direction of the change trend of the enrollment within the combination by comparing the difference direction between the enrollment value of the latter unit and the former unit, and generate the change trend direction value of the enrollment. S212: Based on the trend direction value of the number of admitted students, construct a continuous trend judgment sequence, call the unit trend direction, identify adjacent unit combinations that continuously appear in the same trend direction, number and mark the combinations that meet the condition of consistent continuous trend direction, and generate a continuous trend number unit sequence. S213: Based on the continuous trend numbering unit sequence, filter the boundary units corresponding to the numbers at both ends of the sequence, compare whether the trend directions at the boundaries are consistent, and if the directions are consistent, extract all the unit numbers corresponding to the numbers as the origin region of supply and demand migration, and establish a spatial prediction starting point unit set.

[0009] As a further aspect of the present invention, the step of obtaining the predicted supply and demand aggregation structure block specifically includes: S311: Call the spatial prediction starting point unit set, collect the student capacity and enrollment data in the surrounding two-layer grid, calculate the ratio of the total student capacity of the two-layer grid in the local area to the real-time enrollment, and compare the ratio with the ratio of the total capacity of the whole area to the total enrollment to generate the capacity transfer difference rate. S312: Based on the capacity transfer difference rate, record the inflow and outflow directions of the grid population, mark the population paths with the same direction between adjacent grids in sequence, and count the total frequency of path intersections in the spatial range. Filter out areas where the path intersection frequency exceeds the average distribution number of paths to obtain the number of path intersection segments. S313: Based on the number of intersection segments of the path, extract the set of spatial units with intersection frequency exceeding the average value in the spatial coordinate system according to the spatial distribution adjacency relationship, and calculate the supply and demand deviation intensity value by combining the capacity transfer difference rate of the corresponding location with the consistency ratio of the grid population flow direction within the coverage area. Then, group the supply and demand deviation intensity value in adjacent areas according to the spatial adjacency relationship to obtain the predicted supply and demand aggregation structure block.

[0010] As a further aspect of the present invention, the steps for obtaining the predictor variable time behavior cue map are specifically as follows: S411: Based on the geographic number and time record of the spatial unit in the predicted supply and demand aggregation structure block, extract the corresponding enrollment number, migration direction trajectory and education acceptance number information, record the data call type and start and end time in time order, take the time record as the main sequence, sort the time interval of the call node and match the geographic number to generate a geographic time node matching table. S412: Call the information on the number of students enrolled and the number of students admitted to education within the time period in the geographical time node matching table, group the data by time period, calculate the direction and magnitude of data change within the group, determine the vector angle between the change direction and the migration direction trajectory, identify continuous time periods in which the angle between the two changes is continuously close, and obtain the convergent trajectory matching segment. S413: Based on the time period and geographic number interval in the convergent trajectory matching segment, extract the continuously read trajectory segment, perform joint judgment on the associated enrollment number, education acceptance number and corresponding time length information, calculate the time behavior clue aggregation value, use the geographic number as an index, evaluate the mapping relationship between the value and the trajectory sequence, and obtain the predictive variable time behavior clue map.

[0011] As a further aspect of the present invention, the method further includes step S5: S5: Call the time behavior clue diagram of the predicted variable, perform a sequence comparison of the correlation and time interval according to the order of the predicted variable call, screen out the node numbers of the sequence break or time conflict, map the numbers to the corresponding spatial units in the predicted supply and demand aggregation structure block and mark them, and generate a prediction chain spatial offset mark set. The predicted chain spatial offset tag set includes offset node number, conflict time interval, and spatial location label.

[0012] As a further aspect of the present invention, the step of obtaining the predicted chain spatial offset marker set specifically includes: S511: Call the predictor node number, timestamp information and call order number in the predictor time behavior clue graph, compare the order number and timestamp between adjacent nodes one by one, extract the node numbers with discontinuous order numbers or reversed order of timestamps, and generate an abnormal node number set. S512: Based on the node numbers in the abnormal node number set, locate and predict the spatial unit number corresponding to the number in the supply and demand aggregation structure block, and match it through the spatial unit number list to identify the one-to-one correspondence between the node number and the spatial unit number, and generate a spatial number offset comparison set. S513: Based on the spatial unit number in the spatial number offset reference set, call the annotation field of the spatial marker, mark the spatial unit position with an identifier, and determine the identifier value according to the fracture type or time conflict type of the node to generate the predicted chain spatial offset marker set.

[0013] The dynamic perception system for degree supply and demand based on spatiotemporal big data is used to execute the aforementioned dynamic perception method for degree supply and demand based on spatiotemporal big data. The system includes: The migration positioning module obtains the student's origin location, target university coordinates, and enrollment year number, calculates the spatial straight-line path length between the origin and the target university, and generates a migration stability label list. The trend identification module calls the tag unit number in the migration stability tag list to obtain the difference in the number of students admitted to adjacent units, and determines the direction of change by combining the positive or negative value of the difference in the number of students admitted, and generates a set of spatial prediction starting units. The capacity estimation module calls the coordinates of the tag unit corresponding to the centralized number of the spatial prediction starting unit, collects the student carrying capacity and enrollment capacity of the two layers of grids around the unit, records the population movement direction of the unit, and generates a predicted supply and demand aggregation structure block. The behavioral clue module calls the spatial coordinate number and time record corresponding to the predicted supply and demand aggregation structure block, extracts the enrollment number, population migration direction trajectory and educational capacity of the unit, marks the calling order of each data record in time order, and generates a time behavioral clue map of the predicted variables. The offset marker module calls the predictor variable time behavior cue map, compares the calling order number and time interval value of each set of numbers, and compares whether there is a reversal or jump in the calling order and time sequence between adjacent numbers, and generates a prediction chain spatial offset marker set.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by processing enrollment information and population migration trends, a path length filtering and directional segment repetition mechanism is constructed to effectively distinguish spatial label units with stable migration characteristics. Combined with the changing trends reflected by the consistency of continuous directions, the origin regions of supply and demand migration are accurately located. The spatial aggregation judgment is strengthened by statistical means of neighborhood capacity ratio and path intersection frequency, improving the expression accuracy of dynamic connections between regions. By using variable call trajectories in time sequence, the continuous call relationship of multi-dimensional information is identified. On this basis, a spatial offset marker set is constructed to identify the break or conflict position of supply and demand relationship in the time evolution, improve the sensitivity and dynamic response capability of spatial prediction structure, and realize the proactive perception and intervention prompts of supply and demand change trends in multiple spatiotemporal dimensions. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the workflow of the present invention; Figure 2 This is a flowchart illustrating the process of obtaining the migration stability tag list in this invention. Figure 3 This is a flowchart illustrating the process of obtaining the spatial prediction starting point unit set in this invention. Figure 4 This is a flowchart illustrating the process of obtaining the predicted supply and demand aggregation structure block in this invention. Figure 5 This is a flowchart of the process for obtaining the time behavior cue map of predictive variables in this invention; Figure 6 This is a flowchart illustrating the process of obtaining the predicted chain spatial offset marker set in this invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0017] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0018] Please see Figure 1 This invention provides a technical solution: a dynamic sensing method for the supply and demand of academic degrees based on spatiotemporal big data, comprising the following steps: S1: Obtain the student's origin location, target school coordinates, and enrollment year number; calculate the spatial path length between the origin and target school; filter tag units whose path length is within an acceptable range; combine the origin migration direction information of the tag units in three consecutive years; count the number of repeated direction segments; and generate a migration stable tag list. S2: Call the migration stability label list to obtain the direction of change in the number of students admitted in adjacent units, mark the units that show the same direction continuously, take the unit number with the same trend at the boundary as the origin region of supply and demand migration, and generate a set of spatial prediction starting units. S3: Call the spatial prediction starting unit set, collect the student capacity and enrollment data in the surrounding two layers of grid, calculate the proportion of transferable capacity in the local area to the overall capacity, record the population migration trend in the area, count the number of times the population path intersection location appears in the spatial range, perform spatial fusion processing on areas where the frequency of occurrence exceeds the average distribution of paths, and generate a predicted supply and demand aggregation structure block. S4: Based on the geographic number and time record of the spatial unit in the predicted supply and demand aggregation structure block, extract the corresponding enrollment number, migration direction trajectory and education acceptance number information, record the type and start and end time of each data call in chronological order, identify the reading trajectory of continuous data, and generate a time behavior cue map of the predicted variables. S5: Call the time behavior clue diagram of the predictor variables, perform a sequence comparison of the correlation and time interval according to the order of the predictor variable calls, screen out the node numbers of the sequence break or time conflict, map the numbers to the corresponding spatial units in the predicted supply and demand aggregation structure block and mark them, and generate a prediction chain spatial offset mark set. The migration stability label list includes label number, number of repeating direction segments, and direction consistency identifier; the spatial prediction starting point unit set includes trend consistency boundary unit number, direction continuity marker, and initial supply and demand characteristic value; the prediction supply and demand aggregation structure block includes path intersection frequency, transfer capacity ratio, and spatial fusion unit number; the prediction variable time behavior cue map includes variable call type, variable start and end time, and data trajectory sequence; and the prediction chain spatial offset marker set includes offset node number, conflict time interval, and spatial location label.

[0019] Please see Figure 2 The specific steps for obtaining the migration stable tag list are as follows: S111: Obtain the student's origin location, target school coordinates and enrollment year number, calculate the spatial path length between the student's origin and the corresponding target school, determine whether it is less than the path length filtering threshold, and extract the tag unit number that meets the condition after filtering to obtain the path filtering tag number set; The enrollment year numbers are sequentially assigned, with 2020, 2021, and 2022 designated as Y1, Y2, and Y3 respectively. These are used as association fields to extract the geographical coordinates of target universities, assigning university number A001 with coordinates of longitude 116.39 and latitude 39.91. Batch parsing of student origin data extracts the administrative division identifiers and corresponding latitude and longitude coordinates of each origin, labeling them from B001 to B500, totaling 500 origin data points. Each origin's coordinates are paired with its enrollment year number. A formula for calculating the distance between two points is then established by calling a geographical path length calculation expression. Batch calculations were performed to determine the spatial path length between students' origin and target universities. A path length threshold of 250 kilometers was set. Validation using sample data revealed that 250 kilometers represents a moderate migration radius. See the following path length example for further details: Table 1: Example Table of Path Length Calculation

[0020] As shown in Table 1, the path lengths corresponding to B001 and B002 are both below the screening threshold of 250 kilometers, so they are retained for the next step of analysis. B003 is removed because its path length is greater than 250 kilometers. The source location tag unit numbers after screening are combined into a set to generate a path screening tag number set.

[0021] S112: Call the path filter tag number set, obtain the corresponding source coordinates in three consecutive enrollment years, compare the latitude and longitude change direction between the source coordinates in consecutive years, count the consistency of the migration direction change of the tag unit source in three years, and generate a list of tags with consistent migration direction. Extract the source location coordinates of the corresponding tags for three consecutive enrollment years (Y1 to Y3). Assuming that all three years of data exist for each tag unit as a premise for analysis, obtain the latitude and longitude of tag number S101 in Y1, Y2, and Y3 as (116.20, 39.80), (116.50, 40.10), and (116.80, 40.40), respectively. Calculate the direction vectors between adjacent years sequentially, and calculate the direction code based on the latitude and longitude differences between two years. Set the direction code as follows: Northeast (NE) = +1, Southwest (SW) = -1, and the remaining directions as: North (N) = +2, South (S) = -2, East (E) = +3, and West (W) = -3. The direction code calculation rules are as follows: ; Substituting the latitude and longitude of S101, Y1→Y2 is calculated by increasing both longitude and latitude by 0.3, and the direction of belonging is increased by 1. Y2→Y3 is also increased by 1. Since the two directions have the same code, the consistency of the direction change is recorded as 1. If the direction changes between two years, it is recorded as 0. The determination is based on whether the number of consistency counts is greater than 1. If the number of consistency counts is 2, the tag number is retained; if it is 0, it is removed. This determination rule is based on the logic of continuous path change. Therefore, the S101 tag meets the conditions and is included in the migration direction consistency tag list.

[0022] S113: Retrieve the list of tags with consistent migration directions, and statistically analyze whether the migration direction segments of the origin locations are consistent across consecutive years for each tag unit number. Record the tag numbers whose consecutive direction consistency exceeds a set direction repetition threshold as stable unit numbers, using the following formula: ; Calculate the directional repetition trend of the label unit, and obtain a list of migration-stable labels if the directional repetition trend exceeds the directional repetition threshold. in, This represents the degree of repetition in a particular direction. Representative year The source location direction code, Representative tag unit Spatial path length value, Representative tag unit The difference in latitude, Representative tag unit Number of directional changes within three years This represents the number of tag units; The formula calculation logic considers the impact of changes in direction encoding, path length, and latitude difference on the direction trend within each time period, and normalizes the results using the direction change frequency to ensure both sensitivity and stability of the calculation results. This indicates the change in directional coding between consecutive years; a larger value indicates a more drastic change in direction. Indicates the tag unit number Year to the The spatial span between years includes not only path length It also includes latitudinal migration. The sum of the squares of the two reflects the magnitude of the spatial directional shift; furthermore, the denominator The sign indicates the number of directional changes plus one, introduced as a normalization term to prevent the trend value from becoming too large when directional changes are frequent, while avoiding the calculation problems of dividing by zero; the absolute value sign ensures the non-negativity of the trend value, adapting to different directional changes, by applying the sign to all... By summing up the changes in segment direction, the directional repetition trend of the label unit can be obtained; The directional repetition trend is used to quantify the coupling strength between the change in the migration direction of the origin of the tag unit in consecutive years and the spatial characteristics of the path. It reflects the consistency of the directional change by calculating the comprehensive ratio of the directional coding difference, the path length and the latitude change. The consistency count of the source direction segment is counted in three consecutive years. If the direction code of S101 is +1 in Y1→Y2→Y3, then there are 2 direction repetition segments. The direction repetition threshold is set to 1 segment, that is, the tag with the same direction segment more than once is regarded as a direction stable unit. Let the direction code of S101 be [+1, +1, +1], the path length values ​​be [145.6, 148.3] kilometers, the latitude difference be [0.3, 0.3], and the number of direction changes be 0. Then, substituting these values ​​into the calculation: Calculation for the first segment: ; Calculation in section 2: ; Substitute into the formula to calculate: ; Let the direction code of S102 be [+1, -1, +1], the path length be [245.3, 240.2], the latitude difference be [0.5, 0.6], the number of direction changes be 2, and the direction code changes be (-2, +2) respectively; Calculation for the first segment: ; Calculation in section 2: ; Substitute into the formula to calculate: ; Based on the above calculations, a higher trend value indicates a greater change in direction and instability in direction. If the trend value threshold is set to 100, then S101 is determined to be stable migration and S102 is determined to be unstable migration, thus generating a list of stable migration labels.

[0023] Please see Figure 3 The specific steps for obtaining the spatial prediction starting unit set are as follows: S211: Call the enrollment data of the units in the migration stability label list, select adjacent units in turn to build a comparison combination, perform difference calculation by combining the enrollment values ​​of each pair of adjacent units, and determine the direction of the change trend of the enrollment within the combination by comparing the difference direction between the enrollment value of the latter unit and the former unit, and generate the change trend direction value of the enrollment. Extract complete unit data tables from the information platform. The content should include the unique identifier of each unit and the number of students admitted in the corresponding year. Set the unit numbers as 1, 2, 3, etc., and record the number of students admitted in a specific year, such as 120 students admitted in 2021 and 140 students admitted in 2022. After extraction, sort them in ascending order by number. When constructing adjacent unit combinations, a sliding window method can be used to pair them one by one. Form the first group with units numbered 1 and 2, and the second group with units numbered 2 and 3. Extract the number of students admitted in each pair, and subtract the number of students admitted in the previous unit from the number of students admitted in the next unit to obtain the difference in the number of students admitted. This can determine the direction of the trend of the combination and generate the direction value of the trend of the number of students admitted.

[0024] S212: Based on the trend direction value of the number of students admitted, construct a continuous trend judgment sequence, call the unit trend direction, identify the combination of adjacent units that continuously appear in the same trend direction, number and mark the combination that meets the condition of consistent continuous trend direction, and generate a continuous trend numbered unit sequence. The trend direction value of each pair of adjacent units is called. The trend direction value consists of three states: "increase", "decrease" and "no change". The direction values ​​are arranged in order of combination number to form an initial trend direction sequence. The sequence is identified by traversal to identify whether there are consecutive segments with the same trend direction value. If there are multiple consecutive combinations with the trend direction of "increase", it can be determined that the combination constitutes a continuous upward trend segment. If there are two consecutive combinations with "decrease" trends, it can be determined as a group of continuous downward trend segments. Each group of combinations with the same trend direction is identified and given a unique number mark to establish a trend segment number list. At the same time, the unit numbers covered by the trend direction value are also recorded. The trend number T01 is set to cover unit numbers 2, 3, and 4, indicating that it is formed by the combination (2,3) and (3,4) and the trend segment direction is "increase". By traversing the entire trend direction sequence, the combinations that meet the continuous consistency condition can be divided and numbered to generate a continuous trend number unit sequence.

[0025] S213: Based on the continuous trend numbering unit sequence, filter the boundary units corresponding to the numbers at both ends of the sequence, compare whether the trend directions at the boundaries are consistent, and if the directions are consistent, extract all the unit numbers corresponding to the numbers as the origin region of supply and demand migration, and establish a spatial prediction starting point unit set. This requires checking the beginning and end of the unit range contained in each continuous trend segment, extracting the first and last unit numbers covered by the trend segment, and comparing the trend direction values. If the trend directions corresponding to these two boundary numbers are consistent, they are considered to be consistent. Under this condition, the unit numbers contained in the trend segment are listed as the origin region of supply and demand migration and stored in the starting point unit set list. Let the trend segment T02 contain unit numbers 5 to 8. If the direction of units 5 to 6 is "increase", 6 to 7 is "increase", and 7 to 8 is "increase", then the direction of the first and last numbers 5 and 8 is consistent. Units 5, 6, 7, and 8 can be included as a member segment in the prediction starting point set. If the trend starting point of a segment is "increase" and the ending is "decrease", then the segment does not meet the boundary consistency requirement and should be removed. Establish the spatial prediction starting point unit set.

[0026] Please see Figure 4 The specific steps for obtaining the predicted supply and demand aggregation structure block are as follows: S311: Call the spatial prediction starting point unit set, collect the student capacity and enrollment data in the two surrounding grid layers, calculate the ratio of the total student capacity of the two grid layers in the local area to the real-time enrollment, and compare the ratio with the ratio of the total capacity of the whole area to the total enrollment to generate the capacity transfer difference rate. Urban space was grid-coded using a standard hexagonal grid division method. Twenty-one grids, numbered G3100 to G3120, were selected as the prediction starting point unit set. Data was collected from the surrounding two layers of grids—the first and second rings of adjacent grids around each unit—resulting in 231 grid units. Two data items were recorded in each unit: student capacity and enrollment data. Student capacity represents the total number of school places within the grid, and enrollment data represents the number of students admitted to the school in that year. Taking G3101 as an example, its surrounding two layers of grids include 10 grid units numbered G3102 to G3110. The area's capacity was set at 680 students, and the actual enrollment was 712 students. The local capacity ratio for this area was then calculated as follows: For the entire area, the total capacity is 16,320 people, and the total enrollment is 15,600 people, resulting in an overall capacity ratio of [missing information]. The local capacity ratio is compared with the overall capacity ratio, and the relative difference ratio is calculated. The above operations are carried out on 21 prediction starting units to generate capacity transfer difference rates.

[0027] S312: Based on the capacity transfer difference rate, record the inflow and outflow directions of the grid population, mark the population paths with the same direction between adjacent grids in sequence, and count the total frequency of path intersections in the spatial range. Filter out areas where the path intersection frequency exceeds the average distribution number of paths to obtain the number of path intersection segments. To record the inflow and outflow directions of the population in each grid, it is necessary to statistically analyze the number of registered residents moving in and out of each unit grid over the past 12 months. This data is extracted using a population flow ledger device. Taking grid G3101 as an example, the number of residents moving in during the year was 152, and the number moving out was 138. The net migration direction is inflow, with northward migration marked as positive. Migration paths with consistent directions between adjacent grids are marked as continuous population paths to identify their continuity. The consistency of path directions is determined by overlaying path direction vectors. For any intersection node, the number of times it is traversed by different paths is counted. If node N12 is traversed by 7 paths, its intersection frequency is 7. An average distribution number of paths is set as the intersection threshold. Based on a sample of 287 paths and 1161 nodes, the average distribution number is... If the frequency of a node exceeds the threshold by more than 3 times, i.e., more than 0.741 times, it is counted as a valid intersection node. After rounding, the threshold is set to 1. Nodes with a path intersection frequency greater than 1 are determined to be intersection areas. A total of 83 unit areas that meet the conditions are identified within the spatial range, and the number of path intersection segments is obtained.

[0028] S313: Based on the number of path intersection segments, extract the set of spatial units with intersection frequencies exceeding the average value within the spatial coordinate system according to spatial distribution adjacency relationships. Calculate this using the formula: ; Calculate the supply-demand deviation strength value, and group the supply-demand deviation strength value in adjacent regions according to the spatial adjacency relationship to obtain the predicted supply-demand aggregation structure block; in, This represents the strength of the supply-demand deviation. Representing the Cell capacity transfer difference rate, Representing the The number of path intersection segments in a unit. Representing the The ratio of consistency in population flow direction within a unit. Representing the The number of adjacencies between spatial units covered by a single unit. Represents the total number of fusion units within the region; Formula calculation logic: Based on the capacity transfer difference rate of each unit within the fusion region. Number of paths intersecting Consistency ratio with population flow direction The square root of the product and the difference are used to quantify the degree of matching between the supply and demand sides in structural dynamics; the adjacency number in the denominator... Adding 1 as a complexity adjustment term is used to dilute the impact of local extreme deviations. The degree of supply and demand matching deviation of the calculation unit is calculated and averaged to obtain the overall supply and demand deviation intensity value of the fusion region. The supply-demand deviation intensity value is used to measure the degree of structural matching difference between the capacity of educational resources and the population flow trend within a spatial integration unit. The larger the value, the greater the structural mismatch between the supply capacity and demand pressure in the region. Spatial fusion regions are extracted, and the set of cells covered by high-frequency intersection areas is identified. The corresponding capacity transfer difference rate and population flow direction consistency ratio are extracted. This ratio is represented by the proportion of the dominant migration direction within the region. Taking cell G3105 as an example, its internal 12-month migration direction statistics show 63% northward, 22% southward, and 15% in the remaining directions. The dominant direction consistency ratio is taken as the maximum value of 0.63. The number of adjacencies between the covered grids of the fusion cells is also counted. The capacity transfer difference rate is 8. The number of path intersection segments is 0.127. If the value is 5, then the directional consistency ratio is... It is 0.63; Table 2: Key Parameters of the Integration Area

[0029] As shown in Table 2, taking region A1 as an example, substituting the above parameters, the following calculations are performed: Calculate the square root part: ; Calculation of numerator: ; The denominator term is ; Substitute into the formula to calculate: ; Calculations were performed separately for each of the 83 regions, and the averages were summed. ; The results indicate that the current spatial integration area exhibits a certain deviation from the clustering trend, requiring adjustments in actual supply and demand allocation. Regions with values ​​significantly higher than the average are integrated, and structural aggregation is performed based on adjacency relationships to obtain predicted supply and demand aggregated structural blocks.

[0030] Please see Figure 5 The specific steps for obtaining the predictor variable time behavior cue map are as follows: S411: Based on the geographic number and time record of the spatial unit in the predicted supply and demand aggregation structure block, extract the corresponding enrollment number, migration direction trajectory and education acceptance quantity information, record the data call type and start and end time in time order, take the time record as the main sequence, sort the time interval of the call node and match the geographic number to generate a geographic time node matching table. This involves extracting spatial IDs and corresponding time records from the education resource scheduling platform of 16 cities in a province in 2023. In actual deployment, the IDs are formatted as "City ID + District ID + Time Code," with "05030820230105" representing "City ID 05, District ID 03, Spatial Block 08, January 5, 2023." Time records are accurate to the day. After extracting the time records, the enrollment numbers under the corresponding IDs need to be extracted. This type of data is processed by the district-level education centers according to... Weekly updated statistics show that number 05030820230105 corresponds to an enrollment of 124 students. Migration trajectory data is derived from records of inter-district transfers of registered residents within the district. Path vector data is extracted using time series data from the receiving and sending district numbers to form migration trajectory records. A 7-day continuous data set from number 050308 to number 050408 constitutes the migration direction path vector. The trajectory direction angle is calculated from the angle formed by the geographical coordinates (latitude and longitude) of the two locations. The angle calculation formula uses... , where vector These are the displacement vectors for the two paths, respectively. If the number of students admitted to education program number 05040820230106 is 118, then a data table can be constructed as follows: Table 3: Example of Educational Data Retrieval

[0031] Referring to Table 3, a correlation table is constructed by combining the data index numbers within the time period. At the same time, records with consecutive numbers and time differences not exceeding 1 day are identified and marked as belonging to the same geographic unit of continuous data stream, thus obtaining a geographic time node matching table.

[0032] S412: Call the information on the number of students enrolled and the number of students admitted to education within a time period in the geographic time node matching table, group the data by time period, calculate the direction and magnitude of data change within the group, determine the vector angle between the change direction and the migration direction trajectory, identify continuous time periods in which the angle between the two changes is continuously close, and obtain the convergent trajectory matching segment. Dividing the data into time periods, with each period lasting 7 days, we extracted enrollment and educational admission data to form two sequences. For example, in the first period, the enrollment sequence for each day is [124, 126, 125, 123, 121, 120, 119], and the educational admission sequence is [117, 118, 117, 116, 115, 114, 113]. We then calculated the difference vector between the two sequences. and The direction of change of each pair of difference vectors is used as the benchmark for migration change. The angle between the vectors is then used to determine whether it remains stable. For example, if the change in enrollment from day 1 to day 2 is 2 and the change in acceptance is 1, then the angle is... This indicates that the direction of change is consistent; similarly, the trend of change in each day within the entire cycle is judged. The angle change is set to less than 30° as the criterion for determining the convergence segment. If the angle is less than 30° for five consecutive days within the cycle, it is considered to form a convergence trajectory segment. In the above sequence, the angles of the 1st to the 5th days are 0°, 12°, 25°, 16° and 27° respectively, which meet the set conditions. Therefore, this segment is identified as a convergence trajectory matching segment.

[0033] S413: Based on the time period and geographic number interval in the convergent trajectory matching segment, extract the continuously read trajectory segment, and perform a joint judgment on the associated enrollment number, educational acceptance number, and corresponding time length information, using the formula: ; Calculate the aggregated value of time behavior clues, use geographic IDs as indexes, evaluate the mapping relationship between numerical values ​​and trajectory sequences, and obtain a time behavior clue map of predictor variables; in, The aggregated value of time-based behavioral clues, Indicates the first The number of students admitted to each district Indicates the first The number of students admitted to education in each district, Indicates the first The migration direction and angle of each segment Indicates the first The magnitude of trajectory change in each segment, Indicates the first The time length of each segment Number of segments; The formula's calculation logic is as follows: The core idea is to normalize and amplify the absolute difference between the number of students enrolled and the number of students admitted to education in each trajectory segment, while also considering the direction and angle of the migration path. With the amplitude of trajectory change The combined influence of these two directional parameters is determined by combining the square root of the sum of squares to reflect the overall intensity of change within the segment, and then further analyzed by time length. To adjust for the interference of short-term drastic fluctuations on the overall result, in the overall formula structure, the numerator is the amplification factor for the spatial vector change, and the denominator is the time decay term. This is achieved by adjusting the... The normalized aggregate performance value of the entire convergent trajectory segment is obtained by summing the sums of the segments. The time behavior clue aggregation value is used to quantify the difference between the number of students enrolled and the number of students receiving education in multiple continuous trajectory segments. Combined with the direction, angle and change of migration trajectory, the value is aggregated after normalization to represent the strength of the consistency of supply and demand dynamics in a specific geographical segment over time. The larger the value, the more coordinated the population flow and changes in educational resources in the corresponding time period have. Taking the first segment as an example, if the enrollment is 124 students, the number of students admitted to education is 117, the trajectory angle is 25.3°, the trajectory amplitude is 10.2, and the time length is 1 day; Difference: ; Square root part: ; Denominator: ; The result for the first paragraph is: ; If a total of 3 trajectory segments are extracted, the parameters are as follows: Table 4: Trajectory Segment Data Table

[0034] Section 2: , , , ; Section 3: , , , ; Substitute into the formula to calculate: ; The results indicate that the aggregation intensity of the predictor variable time behavior cue map mapped by the current convergent trajectory matching segment in the region numbered 050308-050408 is 210.09, which can be used as a numerical basis for subsequent regional aggregation feature analysis.

[0035] Please see Figure 6 The specific steps for obtaining the predicted chain spatial offset marker set are as follows: S511: Call the predictor node number, timestamp information and call order number in the predictor time behavior clue graph, compare the order number and timestamp between adjacent nodes one by one, extract the node numbers with discontinuous order numbers or reversed order of timestamps, and generate an abnormal node number set. The node information is summarized in numerical order and organized into a list with a clear calling order and time order. The sequential numbers of adjacent node pairs are compared one by one to observe whether they show a natural increase. If there is a jump in the sequential number, the previous node is set to 5 and the next node to 7, or the previous node is 7 and the next node is 6. The discontinuity is considered abnormal. The corresponding timestamps are then compared to determine whether there is a reversal. If the timestamp of a node is later than that of the next node, the node pair is recorded as abnormal. Nodes that meet any of the above abnormal conditions are included in the abnormal node number set.

[0036] S512: Based on the node numbers in the abnormal node number set, locate and predict the spatial unit numbers corresponding to the numbers in the supply and demand aggregation structure block, and identify the one-to-one correspondence between node numbers and spatial unit numbers by matching through the spatial unit number list, and generate a spatial number offset comparison set. It is necessary to compare the records in the predicted supply and demand aggregation structure block and query the spatial unit number corresponding to each abnormal node. Each structure block data contains a unique spatial number and node number. By traversing and matching, abnormal nodes can be mapped to spatial units one by one. For example, if an abnormal node number is N12, it can be found in the structure block list in the spatial unit numbered S05, thus confirming the correspondence between the two. During execution, the matching process is completed by establishing a mapping table. If the matching fails, it is recorded as an invalid node. This process requires an integrity check of the structure block data to ensure that each node number can find a unique matching spatial number and generate a spatial number offset reference set.

[0037] S513: Based on the spatial unit number in the spatial number offset reference set, call the annotation field of the spatial marker, mark the spatial unit position with an identifier, and determine the identifier value according to the fracture type or time conflict type of the node to generate the predicted chain spatial offset marker set. Based on the listed spatial unit numbers, a set of annotation information fields containing spatial location and status fields is extracted from the structural data. Each spatial number is assigned a new identifier. The anomaly type of the node is determined, and the judgment is set to determine whether it is a time conflict or a break. Then, a fixed label is selected according to the classification rules. The spatial unit numbers to be processed are marked and updated one by one and saved to a new spatial offset label set. Special attention should be paid to the uniqueness of the spatial numbers during the labeling process. There should be no duplicate labeling or omissions, so as to facilitate subsequent calls and visualization. A spatial offset label set of the prediction chain is generated.

[0038] The dynamic sensing system for degree supply and demand based on spatiotemporal big data is used to execute the aforementioned dynamic sensing method for degree supply and demand based on spatiotemporal big data. The system includes: The migration positioning module obtains the student's origin location, target university coordinates, and enrollment year number, calculates the spatial straight-line path length between the origin and the target university, and generates a migration stability label list. The trend identification module calls the tag unit number in the migration stability tag list to obtain the difference in the number of students admitted to adjacent units. It then determines the direction of change by combining the positive or negative value of the difference in the number of students admitted to generate a set of spatial prediction starting units. The capacity estimation module calls the coordinates of the tag unit corresponding to the centralized number of the starting unit of the spatial prediction, collects the number of students and enrollments in the two layers of grids around the unit, records the population movement direction of the unit, and generates a predicted supply and demand aggregation structure block. The behavioral cue module calls the spatial coordinate number and time record corresponding to the predicted supply and demand aggregation structure block, extracts the enrollment number, population migration direction trajectory and educational capacity of the unit, marks the calling order of each data record in time order, and generates a time behavioral cue map of the predicted variables. The offset marker module calls the predictor variable time behavior cue map, and compares the calling order number and time interval value of each set of numbers to see if there is a reversal or jump in the calling order and time sequence between adjacent numbers, generating a prediction chain spatial offset marker set.

[0039] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A dynamic sensing method for degree supply and demand based on spatiotemporal big data, characterized in that, Includes the following steps: S1: Obtain the student's origin location, target university coordinates and enrollment year number, calculate the spatial path length between the origin and target university, filter the label units with path lengths within an acceptable range, and generate a migration-stable label list; S2: Call the migration stability label list to obtain the direction of change in the number of students admitted to adjacent units, mark the units that show the same direction continuously, take the unit number with the same trend at the boundary as the origin region of supply and demand migration, and generate a set of spatial prediction starting units. S3: Call the spatial prediction starting point unit set, collect the student capacity and enrollment data in the surrounding two layers of grid, calculate the proportion of transferable capacity in the overall capacity within the local area, record the population migration trend in the area, and generate a predicted supply and demand aggregation structure block; S4: Based on the geographic number and time record of the spatial unit in the predicted supply and demand aggregation structure block, extract the corresponding enrollment number, migration direction trajectory and education acceptance quantity information, record the type and start and end time of each data call in chronological order, identify the reading trajectory of continuous data, and generate a predictive variable time behavior cue map.

2. The dynamic sensing method for degree supply and demand based on spatiotemporal big data according to claim 1, characterized in that, The migration stability label list includes label number, number of repeating direction segments, and direction consistency identifier; the spatial prediction starting point unit set includes trend consistency boundary unit number, direction continuity marker, and initial supply and demand characteristic value; the prediction supply and demand aggregation structure block includes path intersection frequency, transfer capacity ratio, and spatial fusion unit number; and the prediction variable time behavior cue graph includes variable call type, variable start and end time, and data trajectory sequence.

3. The dynamic sensing method for degree supply and demand based on spatiotemporal big data according to claim 1, characterized in that, The specific steps for obtaining the migration stability label list are as follows: S111: Obtain the student's origin location, target school coordinates and enrollment year number, calculate the spatial path length between the student's origin and the corresponding target school, determine whether it is less than the path length filtering threshold, and extract the tag unit number that meets the condition after filtering to obtain the path filtering tag number set; S112: Call the path to filter the tag number set, obtain the corresponding source coordinates in three consecutive enrollment years, compare the latitude and longitude change directions between the source coordinates in consecutive years, count the consistency of the migration direction of the tag unit source within three years, and generate a list of tags with consistent migration direction. S113: Call the migration direction consistent label list, count whether the source migration direction segments in consecutive years are consistent for the label unit number, record the label number whose consecutive direction consistency exceeds the set direction repetition threshold as stable unit number, calculate the direction repetition trend degree of the label unit, and obtain the migration stable label list based on the case where the direction repetition trend degree exceeds the direction repetition threshold.

4. The dynamic sensing method for degree supply and demand based on spatiotemporal big data according to claim 3, characterized in that, The specific steps for obtaining the spatial prediction starting point unit set are as follows: S211: Call the enrollment data of the units in the migration stability label list, select adjacent units in sequence to construct a comparison combination, perform difference calculation by combining the enrollment values ​​of each pair of adjacent units, and determine the direction of the change trend of the enrollment within the combination by comparing the difference direction between the enrollment value of the latter unit and the former unit, and generate the change trend direction value of the enrollment. S212: Based on the trend direction value of the number of admitted students, construct a continuous trend judgment sequence, call the unit trend direction, identify adjacent unit combinations that continuously appear in the same trend direction, number and mark the combinations that meet the condition of consistent continuous trend direction, and generate a continuous trend number unit sequence. S213: Based on the continuous trend numbering unit sequence, filter the boundary units corresponding to the numbers at both ends of the sequence, compare whether the trend directions at the boundaries are consistent, and if the directions are consistent, extract all the unit numbers corresponding to the numbers as the origin region of supply and demand migration, and establish a spatial prediction starting point unit set.

5. The dynamic sensing method for degree supply and demand based on spatiotemporal big data according to claim 4, characterized in that, The specific steps for obtaining the predicted supply and demand aggregation structure block are as follows: S311: Call the spatial prediction starting point unit set, collect the student capacity and enrollment data in the surrounding two-layer grid, calculate the ratio of the total student capacity of the two-layer grid in the local area to the real-time enrollment, and compare the ratio with the ratio of the total capacity of the whole area to the total enrollment to generate the capacity transfer difference rate. S312: Based on the capacity transfer difference rate, record the inflow and outflow directions of the grid population, mark the population paths with the same direction between adjacent grids in sequence, and count the total frequency of path intersections in the spatial range. Filter out areas where the path intersection frequency exceeds the average distribution number of paths to obtain the number of path intersection segments. S313: Based on the number of intersection segments of the path, extract the set of spatial units with intersection frequency exceeding the average value in the spatial coordinate system according to the spatial distribution adjacency relationship, and calculate the supply and demand deviation intensity value by combining the capacity transfer difference rate of the corresponding location with the consistency ratio of the grid population flow direction within the coverage area. Then, group the supply and demand deviation intensity value in adjacent areas according to the spatial adjacency relationship to obtain the predicted supply and demand aggregation structure block.

6. The dynamic sensing method for degree supply and demand based on spatiotemporal big data according to claim 5, characterized in that, The specific steps for obtaining the predictor variable time behavior cue map are as follows: S411: Based on the geographic number and time record of the spatial unit in the predicted supply and demand aggregation structure block, extract the corresponding enrollment number, migration direction trajectory and education acceptance number information, record the data call type and start and end time in time order, take the time record as the main sequence, sort the time interval of the call node and match the geographic number to generate a geographic time node matching table. S412: Call the information on the number of students enrolled and the number of students admitted to education within the time period in the geographical time node matching table, group the data by time period, calculate the direction and magnitude of data change within the group, determine the vector angle between the change direction and the migration direction trajectory, identify continuous time periods in which the angle between the two changes is continuously close, and obtain the convergent trajectory matching segment. S413: Based on the time period and geographic number interval in the convergent trajectory matching segment, extract the continuously read trajectory segment, perform joint judgment on the associated enrollment number, education acceptance number and corresponding time length information, calculate the time behavior clue aggregation value, use the geographic number as an index, evaluate the mapping relationship between the value and the trajectory sequence, and obtain the predictive variable time behavior clue map.

7. The dynamic sensing method for degree supply and demand based on spatiotemporal big data according to claim 1, characterized in that, The method further includes step S5: S5: Call the time behavior clue diagram of the predicted variable, perform a sequence comparison of the correlation and time interval according to the order of the predicted variable call, screen out the node numbers of the sequence break or time conflict, map the numbers to the corresponding spatial units in the predicted supply and demand aggregation structure block and mark them, and generate a prediction chain spatial offset mark set. The predicted chain spatial offset tag set includes offset node number, conflict time interval, and spatial location label.

8. The dynamic sensing method for degree supply and demand based on spatiotemporal big data according to claim 7, characterized in that, The specific steps for obtaining the predicted chain spatial offset marker set are as follows: S511: Call the predictor node number, timestamp information and call order number in the predictor time behavior clue graph, compare the order number and timestamp between adjacent nodes one by one, extract the node numbers with discontinuous order numbers or reversed order of timestamps, and generate an abnormal node number set. S512: Based on the node numbers in the abnormal node number set, locate and predict the spatial unit number corresponding to the number in the supply and demand aggregation structure block, and match it through the spatial unit number list to identify the one-to-one correspondence between the node number and the spatial unit number, and generate a spatial number offset comparison set. S513: Based on the spatial unit number in the spatial number offset reference set, call the annotation field of the spatial marker, mark the spatial unit position with an identifier, and determine the identifier value according to the fracture type or time conflict type of the node to generate the predicted chain spatial offset marker set.

9. A dynamic sensing system for degree supply and demand based on spatiotemporal big data, characterized in that, The system is used to implement the dynamic sensing method for degree supply and demand based on spatiotemporal big data as described in any one of claims 1-8, and the system comprises: The migration positioning module obtains the student's origin location, target university coordinates, and enrollment year number, calculates the spatial straight-line path length between the origin and the target university, and generates a migration stability label list. The trend identification module calls the tag unit number in the migration stability tag list to obtain the difference in the number of students admitted to adjacent units, and determines the direction of change by combining the positive or negative value of the difference in the number of students admitted, and generates a set of spatial prediction starting units. The capacity estimation module calls the coordinates of the tag unit corresponding to the centralized number of the spatial prediction starting unit, collects the student carrying capacity and enrollment capacity of the two layers of grids around the unit, records the population movement direction of the unit, and generates a predicted supply and demand aggregation structure block. The behavioral clue module calls the spatial coordinate number and time record corresponding to the predicted supply and demand aggregation structure block, extracts the enrollment number, population migration direction trajectory and educational capacity of the unit, marks the calling order of each data record in time order, and generates a time behavioral clue map of the predicted variables. The offset marker module calls the predictor variable time behavior cue map, compares the calling order number and time interval value of each set of numbers, and compares whether there is a reversal or jump in the calling order and time sequence between adjacent numbers, and generates a prediction chain spatial offset marker set.

Images