School-level academic pressure prediction method and device, electronic equipment and storage medium

Abstract

The invention discloses a school-level academic pressure prediction method and device, electronic equipment and a storage medium, and belongs to the field of data processing, and the method comprises the steps: obtaining a sex-based age structure of a reference year, a population natural change parameter of each prediction year, and a population density control threshold value of each administrative region; and school education resource attributes and space coordinates, and residential space coordinates. The regional migration rate is determined based on the population structure and the density threshold value, year-by-year school age population of each administrative region is calculated in combination with the natural change parameters, and then the year-by-year number of students to be entered is obtained through distribution according to the residence point population proportion. Inputting the educational resource attributes into a multi-layer perceptron model to generate school attraction features, and calculating an enrollment probability in combination with a general learning distance; and predicting the number of years-by-year students in each school, and generating a school pressure index based on the maximum bearing capacity of the school. By implementing the method, the problem that the supply and demand contradiction of a single school cannot be accurately reflected due to the fact that the school pressure prediction granularity is rough in the prior art can be solved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a method, apparatus, electronic device, and storage medium for predicting academic pressure at the university level. Background Technology

[0002] Accurately predicting the demand for school places in each specific school during the compulsory education stage is crucial for optimizing the allocation of educational resources, scientifically formulating teacher recruitment plans, and rationally delineating school district boundaries. Especially in the context of urbanization with high population mobility and uneven distribution of educational resources, simply knowing the total number of school-age children at the administrative district level is insufficient for refined management. Education departments urgently need to grasp the dynamics of supply and demand at the micro level, that is, the capacity of each school to fill its places in the coming years, in order to proactively expand facilities or adjust enrollment policies and avoid the structural contradiction of "a shortage of places in some areas" while "resources are idle" in neighboring areas.

[0003] However, current methods for predicting school place pressure mainly rely on macro-level demographic statistics at the administrative district or regional level. These methods generally suffer from insufficient granularity, failing to effectively refine the macro-level school-age population size down to specific residential areas, thus making it impossible to accurately determine the specific location of students within a region. Furthermore, existing prediction methods often assume a static distribution of students by school district, ignoring the differences in school attractiveness arising from varying educational resource attributes between schools, and failing to accurately assess the actual constraints of commuting distance on school choices. This lack of consideration for "specific student location" and "educational resource attributes and commuting distance" makes it difficult for current technologies to scientifically calculate the actual probability of students from a particular residential area choosing a specific school, resulting in significant deviations in the predicted numbers of students and school place pressure for individual schools. Summary of the Invention

[0004] This invention provides a method, device, electronic device, and storage medium for predicting degree pressure at the school level, which can solve the problems of coarse granularity in degree pressure prediction in the prior art, inability to accurately reflect the supply and demand contradiction of individual schools, and low prediction accuracy due to ignoring the differences in educational resource attributes and the impact of commuting distance on enrollment choices.

[0005] One embodiment of the present invention provides a method for predicting academic degree pressure at the university level, comprising: The target city's baseline year gender and age structure, natural population change parameters for all predicted years, population density control thresholds for each administrative region, educational resource attribute data for each school, spatial coordinate data for each school, and spatial coordinate data for each settlement are obtained; wherein, the predicted year includes the baseline year. Based on the gender and age structure of the baseline year and the population density control threshold, the regional migration rate for each predicted year is determined; based on the gender and age structure of the baseline year and the natural population change parameters for all predicted years, combined with the regional migration rate, the total number of school-age population in each administrative region is extrapolated year by year. According to the preset population proportion of each settlement within its administrative region, the total number of school-age children is allocated to the corresponding settlements year by year to obtain the number of children waiting to enroll in school for each settlement each year. The educational resource attribute data is input into a preset multilayer perceptron model, so that the multilayer perceptron model generates school attractiveness features based on the educational resource attribute data; Based on the spatial coordinate data of each school and each residential area, the commuting distance from each residential area to each school is calculated; based on the school attractiveness characteristics and the commuting distance, the probability of each residential area choosing each school for enrollment is calculated. The projected number of students enrolled in each school is calculated annually based on the number of students awaiting enrollment in each residential area and the enrollment probability. Based on the projected number of students enrolled and the maximum capacity of the school calculated according to the educational resource attribute data, an annual school-level degree pressure index is generated.

[0006] Furthermore, determining the regional migration rate for each predicted year based on the population density control threshold includes: Based on the gender and age structure of the baseline year, the total population of each administrative region in the baseline year is obtained; Based on the population density control threshold, determine the maximum carrying capacity of each administrative region; The total population to be relocated for each administrative region is determined based on the total population of the base year and the maximum carrying capacity. Based on the total relocation population, the preset forecast duration, and the total population of the base year, calculate the estimated total population of each administrative region in each forecast year. Based on the total relocation population, the preset prediction duration, and the estimated total population for each prediction year, calculate the regional migration rate for each prediction year.

[0007] Furthermore, the step of calculating the estimated total population of each administrative region in each forecast year based on the total relocation population, the preset forecast duration, and the total population of the base year includes: For each administrative region, the ratio of the total relocation population to the preset forecast duration is calculated to obtain the average annual migration scale of the current administrative region. If the total population of the current administrative region in the base year is greater than the maximum carrying capacity of the current administrative region, then the estimated total population of the current administrative region in each forecast year is obtained by subtracting the average annual migration scale from the total population of the current administrative region in the base year. If the total population of the current administrative region in the base year is less than or equal to the maximum carrying capacity of the current administrative region, then the estimated total population of the current administrative region in each forecast year is obtained by adding the average annual migration scale to the total population of the current administrative region in the base year.

[0008] Furthermore, the step of extrapolating the total number of school-age children in each administrative region year by year based on the age structure by gender in the baseline year and the natural population change parameters for all predicted years, combined with the regional migration rate, includes: Repeat the annual forecasting operation for the school-age population until the current forecast year reaches the forecast termination year, and generate the total number of school-age population for each administrative region for each year. The annual forecasting operation for the school-age population includes: Based on the natural population change parameters for the current forecast year, construct the sex-specific survival rate matrix and the sex-specific fertility matrix for the current forecast year; where the initial current forecast year is the base year. Calculate the natural population growth for the current forecast year based on the sex-specific survival rate matrix, the sex-specific fertility matrix, and the sex-specific age structure for the current forecast year. The total population is calculated based on the gender and age structure of the current predicted year. The current total population, the regional migration rate of the current predicted year, and the preset gender migration pattern vector are multiplied to obtain the mechanical migration population of the current predicted year. The sex-based age structure for the next forecast year is obtained by overlaying the natural population growth for the current forecast year with the mechanical migration population for the current forecast year. Extract the school-age population corresponding to the school system from the gender-based age structure of the next predicted year, and use it as the total school-age population of each administrative region in the next predicted year. Update the next predicted year to the new current predicted year, and update the gender-based age structure of the next predicted year to the new gender-based age structure of the current predicted year.

[0009] Furthermore, the training of the multilayer perceptron model is based on a joint prediction model; the joint prediction model includes a multilayer perceptron model and a spatial interaction computation sub-model connected in series. The training process of the joint prediction model includes: Obtain the historical training dataset for the reference year; wherein, the historical training dataset includes several historical data samples, each historical data sample includes the educational resource attribute data of each school in the reference year, the number of students waiting to enroll in each residential area in the reference year, the spatial coordinate data of each residential area in the reference year, and the actual number of students enrolled in each school in the reference year as labels. The historical training dataset is randomly divided into several batches of training samples according to a preset batch size; Each batch of training samples is sequentially input into the joint prediction model for iterative training until a preset number of training rounds is reached, completing the training of the joint prediction model. In each iteration, the joint prediction model uses a multilayer perceptron model to generate predicted attractiveness features for each school in a reference year based on the educational resource attribute data in the current batch of samples. Using a spatial interaction computation sub-model, it generates predicted enrollment numbers for each school in the reference year based on the predicted attractiveness features, combined with spatial coordinate data and the number of students awaiting enrollment in the current batch of samples. A preset loss function is used to calculate the difference between the predicted enrollment numbers for each school in the reference year and the actual enrollment numbers in the current batch of samples, generating a loss function value. Finally, a preset optimizer updates the learnable network parameters in the multilayer perceptron model based on the loss function value.

[0010] Furthermore, the step of calculating the probability of each settlement choosing each school based on the school's attractiveness characteristics and the commuting distance includes: A nonlinear attenuation calculation is performed on the travel distance using a preset distance attenuation factor to obtain a space impedance factor that characterizes the resistance to space movement. The school attractiveness characteristics are coupled with the spatial impedance factor to generate the school attraction intensity of each residential area. The attractiveness of schools to each residential area is normalized to determine the probability of each residential area choosing a school for enrollment.

[0011] Furthermore, the step of generating an annual school-level degree pressure index based on the predicted number of enrolled students and the school's maximum capacity calculated according to the educational resource attribute data includes: Obtain the land area, number of teachers, and number of classes of each school from the educational resource attribute data; Using preset construction standard parameters and preset weighting factors, the land area, the number of teachers, and the number of classes are weighted and calculated to obtain the maximum carrying capacity of each school; The ratio of the predicted number of students enrolled to the maximum capacity is calculated to obtain the annual school-level degree pressure index.

[0012] Based on the above method embodiments, the present invention provides corresponding apparatus embodiments.

[0013] An embodiment of the present invention provides a school-level degree pressure prediction device, comprising: a data acquisition module, a macro-population projection module, a micro-allocation module, a feature extraction module, a spatial interaction module, and a pressure assessment module; The data acquisition module is used to acquire the baseline year gender and age structure of the target city, natural population change parameters for all predicted years, population density control thresholds for each administrative region, educational resource attribute data for each school, spatial coordinate data for each school, and spatial coordinate data for each settlement; wherein, the predicted year includes the baseline year. The macro-population extrapolation module is used to determine the regional migration rate for each predicted year based on the gender and age structure of the baseline year and the population density control threshold; and to extrapolate the total number of school-age children in each administrative region year by year based on the gender and age structure of the baseline year and the natural population change parameters for all predicted years, combined with the regional migration rate. The micro-allocation module is used to allocate the total number of school-age children to the corresponding residential areas year by year according to the preset population proportion of each residential area in its administrative region, so as to obtain the number of children waiting to enroll in school in each residential area year by year. The feature extraction module is used to input the educational resource attribute data into a preset multilayer perceptron model, so that the multilayer perceptron model generates school attractiveness features based on the educational resource attribute data. The spatial interaction module is used to calculate the commuting distance from each residential point to each school based on the spatial coordinate data of each school and the spatial coordinate data of each residential point; and to calculate the probability of each residential point choosing each school for enrollment based on the school attractiveness characteristics and the commuting distance. The stress assessment module is used to calculate the predicted number of students enrolled in each school each year based on the number of students waiting to enroll in each residential area and the enrollment probability, and to generate an annual school-level degree pressure index based on the predicted number of students enrolled and the maximum capacity of the school calculated based on the educational resource attribute data.

[0014] Based on the above method embodiments, the present invention provides corresponding electronic device embodiments.

[0015] One embodiment of the present invention provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements any one of the above-described method embodiments of the school-level degree pressure prediction method.

[0016] Based on the above method embodiments, the present invention provides corresponding storage medium embodiments.

[0017] One embodiment of the present invention provides a storage medium storing a computer program thereon, wherein, when the computer program is running, it controls the device where the storage medium is located to execute any of the above-described method embodiments of the school-level degree pressure prediction method.

[0018] Compared with the prior art, the present invention has the following beneficial effects: This invention provides a method, apparatus, electronic device, and storage medium for predicting school-level enrollment pressure. The method acquires the baseline year's gender and age structure of the target city, natural population change parameters for each predicted year, population density control thresholds for each administrative region, educational resource attributes and spatial coordinate data for each school, and spatial coordinate data for each settlement; the predicted year includes the baseline year. Based on the baseline year's population structure and population density control thresholds, the regional migration rate for each predicted year is calculated; the total number of school-age children in each administrative region is extrapolated annually based on the natural population change parameters. According to the preset population proportion of each settlement within its administrative region, the school-age children are allocated to corresponding settlements annually, yielding the number of students awaiting enrollment each year. Educational resource attributes are input into a preset multilayer perceptron model to generate school attractiveness features; and the commuting distance is calculated based on the spatial coordinates of schools and settlements. The probability of settlements choosing each school is obtained based on the school attractiveness features and commuting distance. Finally, the predicted number of students enrolled in each school annually is calculated based on the number of students awaiting enrollment and the enrollment probability, and an annual school-level enrollment pressure index is generated by combining this with the school's maximum capacity.

[0019] This invention effectively overcomes the technical shortcomings of existing technologies, such as coarse prediction granularity and inability to accurately determine the specific location of students, by allocating the school-age population at the administrative district level to specific residential areas according to a preset proportion each year. At the same time, it generates school attractiveness characteristics based on educational resource attribute data through a multilayer perceptron model, and calculates the enrollment probability by combining it with the commuting distance determined based on spatial coordinates. This solves the prediction bias problem caused by ignoring the differences in school attractiveness and the impact of commuting distance on school selection behavior, thereby achieving accurate assessment of school-level degree pressure based on real supply and demand. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating a method for predicting academic pressure at the university level, provided in one embodiment of the present invention.

[0021] Figure 2 This is a schematic diagram of the process for determining the regional migration rate for each predicted year, provided by an embodiment of the present invention.

[0022] Figure 3 This is a schematic diagram of the structure of a school-level degree pressure prediction device provided in an embodiment of the present invention. Detailed Implementation

[0023] 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, and 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.

[0024] like Figure 1 As shown, to address the problems of coarse granularity in existing degree pressure prediction methods, inability to accurately reflect the supply and demand imbalance of individual schools, and low prediction accuracy due to neglecting the impact of differences in educational resource attributes and commuting distance on enrollment choices, an embodiment of the present invention provides a school-level degree pressure prediction method, which includes at least the following steps: Step S1: Obtain the baseline year gender and age structure of the target city, natural population change parameters for all predicted years, population density control thresholds for each administrative region, educational resource attribute data for each school, spatial coordinate data for each school, and spatial coordinate data for each settlement; wherein, the predicted year includes the baseline year. Specifically, data acquisition is fundamental for subsequent high-precision predictions. For population data, the system extracts the sex-specific age structure of the target city in the simulation's starting year (i.e., the baseline year) through a census database or statistical yearbook interface. Mathematically, this data is defined as a population vector, containing the number of males and females in each age group (typically from 0 years old to the highest simulated age), serving as the initial state for subsequent population dynamics extrapolation. Simultaneously, to support dynamic simulations over long periods, the system needs to acquire all natural population change parameters covering the entire prediction period from the baseline year to the prediction's end year. These parameters specifically include sex-specific age-specific fertility rates, total fertility rates, sex ratio at birth, and sex-specific life expectancy or mortality probabilities. These parameters can be sequence values ​​extrapolated from historical trends or expected target values ​​set according to urban development plans, used to construct dynamically evolving life tables and fertility patterns over time, thereby supporting iterative calculations of the Leslie model in the time dimension.

[0025] Regarding spatial planning and constraint data, the system obtains population density control thresholds for each administrative region. These thresholds are based on the target city's territorial spatial planning (such as the 2021-2035 plan), which sets upper limits on population per unit area for different functional zones (such as the central urban area and the outer urban area). Figure 2As shown, this population density control threshold is a key prerequisite parameter for subsequently estimating the maximum carrying capacity of each administrative region, thereby determining the number of mobile populations affected by population relocation policies, and calculating the regional migration rate. By obtaining the administrative boundaries of each administrative region and the corresponding density control indicators, the system can transform macro-level policy constraints into calculable mathematical boundary conditions.

[0026] Furthermore, regarding the micro-level matching of educational resource supply and demand, the system acquires educational resource attribute data, spatial coordinate data of each school, and spatial coordinate data of each residential area. The educational resource attribute data includes, but is not limited to, hard indicators such as the number of full-time teachers, the number of county-level or higher key teachers, the number of classes, and the school's land area. This data will serve as input features for the subsequent multilayer perceptron model, used to quantify the non-linear attractiveness of schools to students. The spatial coordinate data uses a unified geographic coordinate system (such as latitude and longitude or projected coordinates) to obtain the precise locations of schools and residential areas (such as neighborhoods and villages), providing data support for subsequent calculations of commuting distances based on Euclidean distance or road network distance. By constructing the aforementioned multidimensional dataset containing population attributes, policy constraints, educational resources, and spatial geographic information, this invention can provide comprehensive and accurate data input for subsequent end-to-end calculations from macro to micro levels, from population projection to assessment of school place pressure, ensuring that the prediction results accurately reflect the supply and demand relationship of educational resources in the complex urban environment.

[0027] Step S2: Determine the regional migration rate for each predicted year based on the gender and age structure of the baseline year and the population density control threshold; based on the gender and age structure of the baseline year and the natural population change parameters for all predicted years, and in conjunction with the regional migration rate, extrapolate the total number of school-age children in each administrative region year by year. In a preferred embodiment, determining the regional migration rate for each predicted year based on the population density control threshold includes: Based on the gender and age structure of the baseline year, the total population of each administrative region in the baseline year is obtained; Based on the population density control threshold, determine the maximum carrying capacity of each administrative region; The total population to be relocated for each administrative region is determined based on the total population of the base year and the maximum carrying capacity. Based on the total relocation population, the preset forecast duration, and the total population of the base year, calculate the estimated total population of each administrative region in each forecast year. Based on the total relocation population, the preset prediction duration, and the estimated total population for each prediction year, calculate the regional migration rate for each prediction year.

[0028] In a preferred embodiment, calculating the estimated total population of each administrative region in each forecast year based on the total relocation population, the preset forecast duration, and the total population of the base year includes: For each administrative region, the ratio of the total relocation population to the preset forecast duration is calculated to obtain the average annual migration scale of the current administrative region. If the total population of the current administrative region in the base year is greater than the maximum carrying capacity of the current administrative region, then the estimated total population of the current administrative region in each forecast year is obtained by subtracting the average annual migration scale from the total population of the current administrative region in the base year. If the total population of the current administrative region in the base year is less than or equal to the maximum carrying capacity of the current administrative region, then the estimated total population of the current administrative region in each forecast year is obtained by adding the average annual migration scale to the total population of the current administrative region in the base year.

[0029] In a preferred embodiment, the step of extrapolating the total number of school-age children in each administrative region year by year based on the gender and age structure of the baseline year and the natural population change parameters for all predicted years, combined with the regional migration rate, includes: Repeat the annual forecasting operation for the school-age population until the current forecast year reaches the forecast termination year, and generate the total number of school-age population for each administrative region for each year. The annual forecasting operation for the school-age population includes: Based on the natural population change parameters for the current forecast year, construct the sex-specific survival rate matrix and the sex-specific fertility matrix for the current forecast year; where the initial current forecast year is the base year. Calculate the natural population growth for the current forecast year based on the sex-specific survival rate matrix, the sex-specific fertility matrix, and the sex-specific age structure for the current forecast year. The total population is calculated based on the gender and age structure of the current predicted year. The current total population, the regional migration rate of the current predicted year, and the preset gender migration pattern vector are multiplied to obtain the mechanical migration population of the current predicted year. The sex-based age structure for the next forecast year is obtained by overlaying the natural population growth for the current forecast year with the mechanical migration population for the current forecast year. Extract the school-age population corresponding to the school system from the gender-based age structure of the next predicted year, and use it as the total school-age population of each administrative region in the next predicted year. Update the next predicted year to the new current predicted year, and update the gender-based age structure of the next predicted year to the new gender-based age structure of the current predicted year.

[0030] Specifically, in order to incorporate the role of urban planning in regulating population distribution into long-term forecasting, the system first performs the step of determining the regional migration rate. For example... Figure 2 As shown, the process first aggregates the population of all age groups in each administrative region based on the gender and age structure of the baseline year, obtaining the actual total population of each administrative region in the baseline year. Simultaneously, the system reads the population density control threshold for each administrative region. This threshold is typically derived from the city's land use plan (e.g., the 2021-2035 plan), setting different density upper limits (e.g., people / square kilometer) for the central urban area and the outer urban area. The system combines the buildable land area of ​​each administrative region with the population density control threshold, determining the maximum carrying capacity of each administrative region through a product operation. Subsequently, the absolute value of the difference between the total population of the baseline year and the maximum carrying capacity is calculated to determine the total population to be relocated from each administrative region during the forecast period (the outflow for overloaded areas and the potential inflow capacity for underloaded areas).

[0031] After obtaining the total relocation population, in order to transform the static total constraint into a dynamic annual indicator, the system calculates the estimated total population of each administrative region in each forecast year based on the total relocation population, the preset forecast period (e.g., 10 or 15 years), and the total population of the base year. This calculation follows the logic of "linear gradual change under policy guidance." For each administrative region, the system first calculates the ratio of the current total relocation population to the preset forecast period to obtain the current average annual migration scale of the administrative region. This average annual migration scale represents the average number of people who should have a net outflow or net inflow to the region each year under the guidance of the plan. Next, the system determines the current population status of each administrative region: if the current total population of the administrative region in the base year is greater than the current maximum carrying capacity of the administrative region (usually corresponding to...), then... Figure 2 The high-density central area indicates that the region faces population relocation pressure. The system uses the total population of the current administrative region in the base year as a basis, subtracting the average annual migration scale year by year to generate the estimated total population sequence for each predicted year. Conversely, if the total population of the current administrative region in the base year is less than or equal to the maximum carrying capacity of the current administrative region (usually corresponding to...), the system generates the estimated total population sequence for the administrative region in each predicted year. Figure 2 The system considers the surrounding urban areas (as indicated by the population carrying capacity of the region) to generate a projected total population sequence for each projected year. Based on this, the system uses the total population of the current administrative region in the base year as a foundation, adding the estimated annual migration scale year by year to generate the projected total population sequence for each projected year. Then, using the total relocation population (corresponding to the migration scale at the numerator level), the preset prediction duration, and the projected total population for each projected year (corresponding to the annual population base at the denominator level), the system calculates the regional migration rate for each projected year. This regional migration rate is a dynamically changing parameter sequence over time, ensuring that the annual mechanical migration volume always conforms to the planned average annual migration scale, even with changes in the population base.

[0032] After determining the key regional migration rate parameters, the system enters the population iterative extrapolation stage based on the Leslie matrix. This stage extrapolates the total school-age population for each administrative region year by year, based on the sex-age structure of the baseline year and the natural population change parameters for all predicted years, combined with the previously calculated regional migration rates. This process is achieved by repeatedly executing the year-by-year school-age population prediction operation until the current prediction year reaches the prediction termination year. The year-by-year school-age population prediction operation includes two core parts: natural growth simulation and mechanical growth simulation.

[0033] In the natural growth simulation, the system constructs a sex-specific survival rate matrix and a sex-specific fertility matrix for the current predicted year based on natural population change parameters (including sex-specific age-specific fertility rates, mortality probabilities, and sex ratio at birth). The initial current predicted year is set as the baseline year. Subsequently, based on the sex-specific survival rate matrix, sex-specific fertility matrix, and sex-specific age structure for the current predicted year, the system calculates the natural population growth for the current predicted year. This calculation follows matrix operation rules, where the natural population growth vector equals the product of the survival rate matrix and the population vector plus the product of the fertility matrix and the population vector. In the mechanical growth simulation, to accurately reflect the age and sex characteristics of the migrating population, the system calculates the current total population based on the sex-specific age structure for the current predicted year (this value is a real-time value during the simulation). The system then multiplies this current total population, the corresponding value of the regional migration rate determined in the previous steps for the current predicted year, and a preset sex-specific migration pattern vector to obtain the mechanically migrating population for the current predicted year. Among them, the gender-specific migration pattern vector is used to characterize the distribution ratio of the migrant population at different ages and genders (for example, the migration ratio of young and middle-aged people is usually higher).

[0034] Finally, the system vector-superimposes the natural population growth and mechanical migration population of the current predicted year to obtain the gender-specific age structure for the next predicted year, completing one annual iteration. From the gender-specific age structure of the next predicted year, the system extracts the population size for the corresponding age group based on the local education system (e.g., 6-12 years old is primary school), which is used as the total school-age population for each administrative region in the next predicted year. Subsequently, the next predicted year is updated to the new current predicted year, and the gender-specific age structure of the next predicted year is updated to the new current year's gender-specific age structure, entering the next cycle. The above population projection process can be uniformly expressed by the following formula: in, This is the gender-specific age structure vector for the next predicted year; A survival rate matrix by gender; A sex-specific fertility matrix; This is the gender-based age structure vector for the current predicted year; This represents the simulated total population scalar for the current projected year. This represents the regional migration rate corresponding to the current predicted year; This is a vector representing the gender-specific migration pattern. Through the above steps, this invention can transform urban population density control policies into specific mathematical constraints while considering the natural reproduction patterns of the population. This enables accurate projection of the total school-age population in each administrative region over a long period, providing a reliable total baseline for subsequent micro-level assessments of school place pressure.

[0035] Step S3: According to the preset population ratio of each settlement within its administrative region, the total number of school-age children is allocated to the corresponding settlements year by year to obtain the number of children waiting to enroll in school for each settlement year by year. Specifically, to achieve refined spatial allocation, the system first identifies and analyzes the population structure of settlements within each administrative region based on population data from a baseline year (such as census data or grid management data). Settlements, as the basic spatial units in this invention that bear residential functions and generate educational needs, can correspond to specific residential communities, natural villages, or community committees. The system calculates the ratio of the population of each settlement to the total population of its respective administrative region in the baseline year, and defines this ratio as the preset population proportion of each settlement within its administrative region. This proportion characterizes the spatial distribution structure of the population within the administrative region in the baseline year and serves as a weighting coefficient in the prediction model for mapping macro-level totals to micro-level units.

[0036] Based on this, for each forecast year within the forecast period, the system uses the preset population proportion to decompose and calculate the total school-age population of each administrative region obtained in step S2. Specifically, it multiplies the total school-age population of a given administrative region in a specific forecast year by the preset population proportion of a certain settlement within that administrative region, thus obtaining the number of students awaiting enrollment in that settlement in that forecast year. This calculation process assumes that although the total population of each administrative region changes dynamically due to natural growth and mechanical migration within the forecast period, the relative population distribution structure among settlements within the administrative region remains stable or inherits from the baseline year. The mathematical expression of the above allocation process is as follows: In the formula, Indicates the first The pre-defined population percentage of each settlement within its administrative region; Indicates the base year Time Population of each settlement; Indicates the base year Time The total population of the administrative region to which each settlement belongs; Indicates the first Residential areas in the predicted year The number of students awaiting enrollment; Indicates the first The administrative district to which each settlement belongs in the predicted year The total number of school-age children. By implementing this spatial allocation step, the present invention can accurately map the total number of school-age children in a region affected by macro-population policies and natural changes to specific geographic spatial coordinates, thereby clarifying the micro-spatial distribution characteristics of educational demand within the city and solving the technical problem in the prior art that the prediction granularity is only at the administrative district level and cannot accurately identify the differences in enrollment pressure in specific communities or neighborhoods.

[0037] Step S4: Input the educational resource attribute data into a preset multilayer perceptron model so that the multilayer perceptron model generates school attractiveness features based on the educational resource attribute data; In a preferred embodiment, the training of the multilayer perceptron model is based on a joint prediction model; the joint prediction model includes a multilayer perceptron model and a spatial interaction computation sub-model connected in series. The training process of the joint prediction model includes: Obtain the historical training dataset for the reference year; wherein, the historical training dataset includes several historical data samples, each historical data sample includes the educational resource attribute data of each school in the reference year, the number of students waiting to enroll in each residential area in the reference year, the spatial coordinate data of each residential area in the reference year, and the actual number of students enrolled in each school in the reference year as labels. The historical training dataset is randomly divided into several batches of training samples according to a preset batch size; Each batch of training samples is sequentially input into the joint prediction model for iterative training until a preset number of training rounds is reached, completing the training of the joint prediction model. In each iteration, the joint prediction model uses a multilayer perceptron model to generate predicted attractiveness features for each school in a reference year based on the educational resource attribute data in the current batch of samples. Using a spatial interaction computation sub-model, it generates predicted enrollment numbers for each school in the reference year based on the predicted attractiveness features, combined with spatial coordinate data and the number of students awaiting enrollment in the current batch of samples. A preset loss function is used to calculate the difference between the predicted enrollment numbers for each school in the reference year and the actual enrollment numbers in the current batch of samples, generating a loss function value. Finally, a preset optimizer updates the learnable network parameters in the multilayer perceptron model based on the loss function value.

[0038] Specifically, the system first extracts key feature vectors from the acquired educational resource attribute data of each school. This data includes the number of full-time teachers, the number of county-level or higher-level backbone teachers, and the number of classes in each school. These indicators objectively reflect the school's hardware and software strengths. The system then inputs these feature vectors into a pre-set Multilayer Perceptron (MLP) model. This model, pre-trained, contains an input layer, several hidden layers, and an output layer. Through weighted connections and activation function mappings between layers, it transforms the high-dimensional resource attribute features into one-dimensional scalar values, i.e., school attractiveness features. This feature value intuitively represents the degree to which a school is attractive to parents and students, without considering distance factors.

[0039] In a preferred embodiment, to ensure that the multilayer perceptron model accurately reflects real-world school selection logic, the training of the multilayer perceptron model is based on a joint prediction model. The joint prediction model architecturally includes a cascaded feature extraction sub-model and a spatial interaction computation sub-model. The feature extraction sub-model, which is the multilayer perceptron model to be trained, is responsible for extracting attractiveness from resource data; the spatial interaction computation sub-model solidifies the spatial probability computation logic based on the Huff Model, responsible for converting attractiveness into observable enrollment numbers. This architectural design solves the technical challenge of directly supervising training because "school attractiveness" itself lacks a ground truth label.

[0040] The specific training process of the joint prediction model is as follows: First, obtain the historical training dataset for the reference year (e.g., the year with the most recent complete enrollment data). The historical training dataset includes several historical data samples. Each historical data sample contains educational resource attribute data of each school in the reference year, which serves as the model input, and the number of students waiting to enroll and the spatial coordinate data of each residential area in the reference year, which serve as the computational context. Simultaneously, obtain the actual number of students enrolled in each school in the reference year as supervision labels.

[0041] During the training execution phase, the system randomly divides the historical training dataset into several batches of training samples according to a preset batch size. For each batch, the system sequentially inputs the training samples into the joint prediction model for forward propagation iterative training. In each iteration, the data stream first enters the feature extraction sub-model. This sub-model, based on the sample in the current batch, selects the first feature extraction sample from the current batch. Number of full-time teachers in the school Number of full-time teachers at the county level and above and number of classes The neural network outputs the predicted attractiveness characteristics of the school for a reference year. The calculation logic is shown in the following formula: In the formula, For the first The predicted attractiveness characteristics of the schools; The nonlinear mapping function represents the multilayer perceptron model; Indicates the first The number of full-time teachers in the school; Indicates the first The number of full-time teachers at the county level or above in each school; Indicates the first The number of classes in the school; These are the learnable network parameters in the multilayer perceptron model. By implementing this step, the present invention can map the multi-dimensional hard indicators of educational resources in schools into a single-dimensional comprehensive attractiveness index, providing key input variables for subsequent probability calculations based on the gravity model.

[0042] Subsequently, the predicted attractiveness features The data is passed to the spatial interactive computing sub-model. This sub-model does not contain learnable parameters but serves as a mathematical expression layer of physical laws. It combines spatial coordinate data (used to calculate distances) from the current batch of samples with the number of students awaiting enrollment in each settlement to perform physical-level supply and demand allocation calculations. Specifically, the spatial interactive computing sub-model first... The first settlement and the first The spatial coordinates of the schools are used to calculate the commuting distance between them. Using a preset distance attenuation factor The nonlinear decay calculation of the commuting distance is performed, and combined with the school's predicted attractiveness characteristics. Calculate the first The residential area was selected as the first School admission probability The calculation process follows the principles of the Huff gravity model, and its formula is as follows: In the formula, Indicates the first The residential area was selected as the first The probability of admission to the school; The output of the feature extraction sub-model represents the first... The predicted attractiveness characteristics of the schools; Indicates the first From the first residential area to the first The commuting distance to each school; Indicates the distance attenuation factor; This represents the set of schools available in the current batch; and They represent the first in the set. The attractiveness characteristics and distance of each school are analyzed. After obtaining the enrollment probability matrix, the spatial interaction calculation sub-model further combines the number of prospective students in each settlement and generates the predicted number of students enrolled in each school for a reference year through a weighted summation. The calculation formula is as follows: In the formula, The first output of the model represents the... The projected number of students enrolled in each school; This represents the set of settlements within the current batch; Indicates the first The number of residents in each settlement awaiting enrollment; This is the enrollment probability calculated above.

[0043] Finally, the system calculates the difference between the predicted number of students enrolled as output by the joint prediction model and the actual number of students enrolled in the current batch of samples using a preset loss function (such as the mean squared error loss function), and generates a loss function value. It can be represented as: In the formula, This represents the number of schools in the current batch. The first output of the model The projected number of students enrolled in each school; For the first The system identifies the actual number of students enrolled in each school. Based on this loss function value, the system uses a preset optimizer (such as Adam or SGD) to calculate the gradient according to the loss function value. Then, using the gradient backpropagation algorithm, the gradient flow is transmitted through a fixed spatial interactive computation sub-model and back to the feature extraction sub-model. This process is applied to the learnable network parameters in the multilayer perceptron model. Update the model. Repeat the above process until the preset number of training rounds is reached or the model converges. By implementing the above training and prediction steps, this invention can construct an end-to-end mapping relationship between "educational resource input" and "actual enrollment results," which not only achieves a scientific quantification of school attractiveness but also ensures that the model prediction results highly match historical actual data, significantly improving the interpretability and accuracy of degree pressure prediction.

[0044] Step S5: Calculate the commuting distance from each residential point to each school based on the spatial coordinate data of each school and the spatial coordinate data of each residential point; calculate the probability of each residential point choosing each school based on the school attraction characteristics and the commuting distance. In a preferred embodiment, calculating the probability of each settlement selecting each school based on the school attractiveness characteristics and the commuting distance includes: A nonlinear attenuation calculation is performed on the travel distance using a preset distance attenuation factor to obtain a space impedance factor that characterizes the resistance to space movement. The school attractiveness characteristics are coupled with the spatial impedance factor to generate the school attraction intensity of each residential area. The attractiveness of schools to each residential area is normalized to determine the probability of each residential area choosing a school for enrollment.

[0045] Specifically, the system first calculates the commuting distance between residential points (such as the center point of a residential community) and schools based on the underlying data of a Geographic Information System (GIS). This distance can be calculated using the Euclidean distance between the two points or by combining road network data to calculate the actual commuting distance. After obtaining the distance data, the system proceeds to calculate the enrollment probability. In a preferred embodiment, this calculation process is based on an improved gravity model logic and includes three sub-steps. First, the system performs a nonlinear attenuation operation on the commuting distance using a preset distance attenuation factor to obtain a spatial impedance factor characterizing the resistance to spatial movement. This distance attenuation factor is typically an empirical parameter used to control the rate at which increased distance negatively impacts enrollment intention (e.g., the magnitude of the decrease in intention for every 1 kilometer increase in distance). Second, the system couples the school attractiveness characteristics with the spatial impedance factor to generate the enrollment attraction intensity of each school for each residential point. This intensity value comprehensively reflects the combined effect of the school's educational quality (push) and geographical location (resistance). Finally, the enrollment attraction intensity of each school for each residential point is normalized to determine the enrollment probability of each residential point choosing each school. Normalization ensures that for any given settlement, the sum of the probabilities of choosing all available schools is 1, which conforms to the basic axioms of probability theory.

[0046] The above calculation logic can be clearly expressed by the following mathematical formula. First, calculate the... The first residential area to the first The attractiveness of schools to students This intensity is directly proportional to the school's attractiveness and inversely proportional to a power of the commuting distance, calculated as follows: in, Indicates the first The first residential area to the first The attractiveness of each school to students; This represents the first step generated by the preceding steps. The school's attractiveness characteristics; Indicates the first From the first residential area to the first The commuting distance to each school; This indicates the preset distance attenuation factor.

[0047] Subsequently, the probability of enrollment was calculated based on the attractiveness of the school. The calculation formula is as follows: In the formula, Indicates the first The residential area was selected as the first The probability of admission to the school; This represents the set of all available schools within the region; Indicates the first The first residential point in the set The intensity of school attraction. By implementing the above spatial interaction calculation, this invention can transform the abstract school attraction and physical distance barriers into quantifiable student mobility probabilities, thereby accurately depicting the complex competition and distribution patterns of students within a city. This solves the technical shortcomings of traditional methods that rely solely on administrative zoning for static segmentation and ignore residents' active school selection behavior.

[0048] Step S6: Calculate the predicted number of students enrolled in each school each year based on the number of students waiting to enroll in each residential area and the enrollment probability, and generate the school-level degree pressure index for each year based on the predicted number of students enrolled and the maximum carrying capacity of the school calculated based on the educational resource attribute data.

[0049] In a preferred embodiment, generating an annual school-level degree pressure index based on the predicted number of enrolled students and the school's maximum capacity calculated according to the educational resource attribute data includes: Obtain the land area, number of teachers, and number of classes of each school from the educational resource attribute data; Using preset construction standard parameters and preset weighting factors, the land area, the number of teachers, and the number of classes are weighted and calculated to obtain the maximum carrying capacity of each school; The ratio of the predicted number of students enrolled to the maximum capacity is calculated to obtain the annual school-level degree pressure index.

[0050] Specifically, for each school, the system aggregates the potential enrollment demand for that school from all residential areas. Using the number of residents waiting to enroll each year calculated in step S3, and the probability of each resident choosing that school calculated in step S5, the predicted number of students enrolled in that school in a specific predicted year is estimated by summing the products. This calculation process essentially calculates the mathematical expectation of enrollment demand, which objectively reflects the total number of students a school may ultimately enroll in under a multi-school competition environment. The calculation formula is as follows: In the formula, Indicates the first The projected number of students enrolled in the school in the projected year; This represents the set of settlements within the target city; Indicates the first The residential area was selected as the first The probability of admission to the school; Indicates the first Residential areas in the predicted year The number of students awaiting enrollment.

[0051] In a preferred embodiment, to scientifically assess the pressure on school capacity, the system needs to construct a quantitative model of the school's maximum capacity. The process of generating an annual school-level capacity pressure index based on the predicted number of students and the maximum capacity calculated from the educational resource attribute data specifically includes the following steps: First, the system extracts the land area, number of teachers, and number of classes for each school from the educational resource attribute data. These indicators represent the hard constraints on the school in terms of space, teachers, and staffing. Next, using preset construction standard parameters and preset weighting factors, the land area, number of teachers, and number of classes are weighted and calculated to obtain the maximum capacity of each school. The construction standard parameters are set according to the standardized school construction specifications promulgated by the local education department, including per-student land area standards, teacher-student ratio standards, and standard class sizes. Through weighted calculation, the contribution of different resource elements to the school's carrying capacity can be comprehensively considered, avoiding the one-sidedness of a single indicator. The formula for calculating the maximum capacity is as follows: In the formula, Indicates the first The maximum capacity of the school; Indicates the first The area occupied by the school; Indicates the first The number of full-time teachers in the school; Indicates the first The number of classes in the school; This indicates the pre-set standard for per capita land area; This indicates the pre-set standard for the teacher-student ratio (number of teachers / number of students). This indicates the preset standard class size (number of students per class). , , These represent the weighting factors of land area, teacher-student ratio, and class size in the carrying capacity assessment, respectively, and satisfy the following conditions: Typically, because the land area occupied by a school is a rigid constraint and difficult to adjust in the short term, its weight... It can be set to a larger value (such as 0.6), while the number of teachers and classes has a certain degree of adjustability, and the weight can be reduced accordingly.

[0052] Finally, the system calculates the ratio of the predicted number of enrolled students to the maximum capacity, obtaining the annual Degree Load Index (DLI). This index directly reflects the supply and demand tension of school educational resources. Its calculation formula is as follows: In the formula, Indicates the first Schools in the predicted year The university-level degree pressure index; Indicates the first Schools in the predicted year The projected number of students enrolled; Indicates the first The maximum capacity of a school. By generating this index, this invention can help education administrators quickly identify specific schools that may experience "overcrowding" (DLI significantly greater than 1) or "idle resources" (DLI significantly less than 1) in the future. This provides a quantitative basis for making precise decisions regarding teacher recruitment plans, school district adjustment schemes, and school building expansion plans, achieving a leap from vague qualitative judgments to precise quantitative analysis.

[0053] Based on the above method embodiments, the present invention provides corresponding apparatus embodiments.

[0054] like Figure 3 As shown, an embodiment of the present invention provides a school-level degree pressure prediction device, including: a data acquisition module, a macro-population extrapolation module, a micro-allocation module, a feature extraction module, a spatial interaction module, and a pressure assessment module; The data acquisition module is used to acquire the baseline year gender and age structure of the target city, natural population change parameters for all predicted years, population density control thresholds for each administrative region, educational resource attribute data for each school, spatial coordinate data for each school, and spatial coordinate data for each settlement; wherein, the predicted year includes the baseline year. The macro-population extrapolation module is used to determine the regional migration rate for each predicted year based on the gender and age structure of the baseline year and the population density control threshold; and to extrapolate the total number of school-age children in each administrative region year by year based on the gender and age structure of the baseline year and the natural population change parameters for all predicted years, combined with the regional migration rate. The micro-allocation module is used to allocate the total number of school-age children to the corresponding residential areas year by year according to the preset population proportion of each residential area in its administrative region, so as to obtain the number of children waiting to enroll in school in each residential area year by year. The feature extraction module is used to input the educational resource attribute data into a preset multilayer perceptron model, so that the multilayer perceptron model generates school attractiveness features based on the educational resource attribute data. The spatial interaction module is used to calculate the commuting distance from each residential point to each school based on the spatial coordinate data of each school and the spatial coordinate data of each residential point; and to calculate the probability of each residential point choosing each school for enrollment based on the school attractiveness characteristics and the commuting distance. The stress assessment module is used to calculate the predicted number of students enrolled in each school each year based on the number of students waiting to enroll in each residential area and the enrollment probability, and to generate an annual school-level degree pressure index based on the predicted number of students enrolled and the maximum capacity of the school calculated based on the educational resource attribute data.

[0055] It should be noted that the embodiments of the device described above correspond to the embodiments of the present invention described above, and can realize the school-level degree pressure prediction method described in any one of the above embodiments of the present invention. Furthermore, the embodiments of the device described above are merely illustrative. The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between modules indicates that they have a communication connection, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without creative effort.

[0056] Based on the above-described method embodiments of the present invention, a corresponding embodiment of an electronic device is provided.

[0057] An embodiment of the present invention provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the school-level degree pressure prediction method according to any one of the present invention, or, when the processor executes the computer program, it implements the functions of each module in the above-described device embodiments.

[0058] For example, the computer program may be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the terminal device.

[0059] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0060] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0061] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, applications required for at least one function, etc.; the data storage area may store data created based on the use of the mobile phone, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital card (SD card), flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0062] Based on the above method embodiments, the present invention provides corresponding storage medium embodiments; Another embodiment of the present invention provides a storage medium including a stored computer program, wherein, when the computer program is executed, the device where the storage medium is located controls the execution of any one of the above-described school-level degree pressure prediction methods of the present invention.

[0063] The aforementioned storage medium is a computer-readable storage medium, and the computer program includes computer program code, which may be in the form of source code, object code, executable file, or certain intermediate forms. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0064] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0065] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for college degree stress prediction, the method comprising: The method comprises the following steps: obtaining the baseline year gender-age structure of the target city, the population natural change parameters of all prediction years, the population density control threshold of each administrative district, the educational resource attribute data of each school, the spatial coordinate data of each school, and the spatial coordinate data of each residential area; wherein the prediction years include the baseline year; determining the regional migration rate of each prediction year according to the baseline year gender-age structure and the population density control threshold; deducing the total number of school-age population of each administrative district year by year according to the baseline year gender-age structure and the population natural change parameters of all prediction years, in combination with the regional migration rate; distributing the total number of school-age population to the corresponding residential area year by year according to the preset population proportion of each residential area in the administrative district to obtain the number of students to be enrolled in each residential area year by year; inputting the educational resource attribute data into a preset multi-layer perception machine model to enable the multi-layer perception machine model to generate school attraction features according to the educational resource attribute data; calculating the commuting distance from each residential area to each school according to the spatial coordinate data of each school and the spatial coordinate data of each residential area; calculating the probability of enrollment of each school according to the school attraction features and the commuting distance; calculating the predicted number of students in each school year by year according to the number of students to be enrolled in each residential area year by year and the probability of enrollment, and generating a school-level degree pressure index year by year based on the predicted number of students and the maximum carrying capacity of the school calculated according to the educational resource attribute data.

2. The college degree pressure prediction method of claim 1, wherein, The method comprises the following steps: According to the baseline year gender-age structure, the total population of each administrative district in the baseline year is obtained; According to the population density control threshold, the maximum carrying population of each administrative district is determined; According to the total population in the baseline year and the maximum carrying population, the total population of each administrative district is determined; According to the total population, the preset prediction time, and the total population in the baseline year, the estimated total population of each administrative district in each prediction year is calculated; According to the total population, the preset prediction time, and the estimated total population of each prediction year, the regional migration rate of each prediction year is calculated.

3. The college degree pressure prediction method of claim 2, wherein, The method comprises the following steps: For each administrative district, the ratio of the total population of the current administrative district to the preset prediction time is calculated to obtain the annual migration scale of the current administrative district; If the total population of the current administrative district in the baseline year is greater than the maximum carrying capacity of the current administrative district, the total population of the current administrative district in the baseline year is taken as the basis, and the annual migration scale is subtracted year by year to obtain the estimated total population of the current administrative district in each prediction year; If the total population of the current administrative district in the baseline year is less than or equal to the maximum carrying capacity of the current administrative district, the total population of the current administrative district in the baseline year is taken as the basis, and the annual migration scale is added year by year to obtain the estimated total population of the current administrative district in each prediction year.

4. The college degree pressure prediction method of claim 3, wherein, The total number of school-age population of each administrative region in each year is derived according to the gender-age structure of the base year, the population natural change parameters of all predicted years, and the regional migration rate, including: The school-age population prediction operation is repeated until the current predicted year reaches the prediction termination year, and the total number of school-age population of each administrative region in each year is generated; The school-age population prediction operation includes: According to the population natural change parameters of the current predicted year, a gender-specific survival rate matrix of the current predicted year and a gender-specific birth matrix of the current predicted year are constructed; the initial current predicted year is the base year; According to the gender-specific survival rate matrix of the current predicted year, the gender-specific birth matrix of the current predicted year, and the gender-specific age structure of the current predicted year, the natural increase population of the current predicted year is calculated; According to the gender-specific age structure of the current predicted year, the total population of the current predicted year is counted, and the total population of the current predicted year, the regional migration rate of the current predicted year, and the preset gender-specific migration mode vector are multiplied to obtain the mechanical migration population of the current predicted year; The natural increase population of the current predicted year and the mechanical migration population of the current predicted year are superimposed to obtain the gender-specific age structure of the next predicted year; The school-age population corresponding to the school system is extracted from the gender-specific age structure of the next predicted year as the total number of school-age population of each administrative region in the next predicted year; The next predicted year is updated as a new current predicted year, and the gender-specific age structure of the next predicted year is updated as the gender-specific age structure of the new current predicted year.

5. The college degree pressure prediction method of claim 4, wherein, The training of the multi-layer perception model is based on a joint prediction model; the joint prediction model includes a multi-layer perception model and a spatial interaction calculation sub-model connected in series; The training process of the joint prediction model includes: Obtaining a historical training data set of a reference year; the historical training data set includes a plurality of historical data samples, each historical data sample including education resource attribute data of each school in the reference year, a number of students to be enrolled in each residential area in the reference year, spatial coordinate data of each residential area in the reference year, and actual number of students in each school in the reference year as a label; The historical training data set is randomly divided into a plurality of batches of training samples according to a preset batch size; The batches of training samples are sequentially input into the joint prediction model for iterative training until a preset number of training rounds is reached, and the training of the joint prediction model is completed; wherein, in each iteration, the joint prediction model generates predicted attraction features of each school in the reference year according to the education resource attribute data in the current batch of samples through a multilayer perceptron model; generates predicted student numbers of each school in the reference year according to the predicted attraction features, in combination with the spatial coordinate data and the number of students to be enrolled in the current batch of samples through a spatial interaction calculation sub-model; calculates the difference between the predicted student numbers of each school in the reference year and the actual student numbers in the current batch of samples through a preset loss function to generate a loss function value; and updates the learnable network parameters in the multilayer perceptron model according to the loss function value using a preset optimizer.

6. The college degree pressure prediction method of claim 5, wherein, The school attraction features and the commuting distance are used to calculate the enrollment probability of each school in each residential area, including: A nonlinear attenuation operation is performed on the commuting distance using a preset distance attenuation factor to obtain a spatial impedance factor for representing spatial movement resistance; The school attraction features and the spatial impedance factor are coupled to calculate the school attraction strength of each school in each residential area; The school attraction strength of each school corresponding to each residential area is normalized to determine the enrollment probability of each school in each residential area.

7. The college degree pressure prediction method of claim 6, wherein, The school-level degree pressure index is generated year by year based on the predicted student numbers and the maximum carrying capacity of the school calculated from the education resource attribute data, including: The land area, the number of teachers, and the number of classes of each school are obtained from the education resource attribute data; The land area, the number of teachers, and the number of classes are weighted and calculated using a preset construction standard parameter and a preset weight factor to obtain the maximum carrying capacity of each school; The ratio of the predicted student number to the maximum carrying capacity is calculated to obtain the school-level degree pressure index year by year.

8. A college degree stress prediction device, characterized by, It includes: a data acquisition module, a macro population deduction module, a micro distribution module, a feature extraction module, a spatial interaction module, and a pressure evaluation module; The data acquisition module is configured to obtain the baseline year gender-age structure of a target city, population natural change parameters for all prediction years, population density control thresholds for each administrative district, education resource attribute data for each school, spatial coordinate data for each school, and spatial coordinate data for each residential area; wherein, the prediction years include the baseline year; The macro population deduction module is configured to determine the regional migration rate for each prediction year according to the baseline year gender-age structure and the population density control thresholds; and deduce the total number of school-age population in each administrative district year by year according to the baseline year gender-age structure, the population natural change parameters for all prediction years, and the regional migration rate; The micro distribution module is configured to distribute the total number of school-age population to the corresponding residential areas year by year according to a preset population proportion of each residential area in the administrative district to obtain the number of students to be enrolled in each residential area year by year; The feature extraction module is configured to input the education resource attribute data into a preset multi-layer perception machine model, so that the multi-layer perception machine model generates school attractiveness features according to the education resource attribute data; The space interaction module is configured to calculate a commuting distance from each residential area to each school according to the space coordinate data of each school and the space coordinate data of each residential area, and calculate an enrollment probability of each school selected by each residential area according to the school attractiveness features and the commuting distance; The pressure evaluation module is configured to calculate a predicted number of students in each school in each year according to the number of students to be enrolled in each residential area in each year and the enrollment probability, and generate a school-level degree pressure index in each year based on the predicted number of students in each school in each year and a maximum carrying capacity of each school calculated according to the education resource attribute data.

9. An electronic device, comprising: A computer program product includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the computer program is executed by the processor to implement the school-level degree pressure prediction method according to any one of claims 1 to 7.

10. A storage medium, characterized by The storage medium includes a stored computer program, wherein the computer program controls a device where the storage medium is located to implement the school-level degree pressure prediction method according to any one of claims 1 to 7 when the computer program is running.

Images