Method for determining population distribution data and related device

Abstract

The application discloses a population distribution data determination method and related equipment, and relates to the technical field of data processing. The method comprises the following steps: acquiring a multi-source population data set of a target region; performing sub-regional category division on the target region according to a population aggregation degree, to obtain N preset sub-regions with different aggregation degrees; performing data division on the multi-source population data set based on the N preset sub-regions, to obtain N sub-regional data sets corresponding to the N preset sub-regions; using a preset algorithm corresponding to the aggregation degree to respectively perform data integration on the N sub-regional data sets, to obtain N sub-regional population distribution data; and performing data merging on the N sub-regional population distribution data, to obtain target population distribution data of the target region. According to the application, the region is divided according to the population aggregation degree, and the data integration is performed by using the adaptive algorithm, so that the accuracy and quality of the population data of different regions are optimized, and the accuracy of the population multi-source data fusion is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, more particularly, the present application relates to a population distribution data determination method and related equipment. BACKGROUND

[0002] With the rapid development of social economy and the continuous advancement of urbanization, population distribution data is increasingly important in the fields of urban planning, resource management, public services, etc. In order to achieve more accurate population prediction and resource allocation, obtaining high-precision population distribution data has become a focus in current research. Traditional methods of obtaining population distribution data usually rely on population census and administrative division statistical data. However, these methods have problems such as poor timeliness, low spatial resolution, poor regional applicability, etc., and cannot accurately reflect the population distribution characteristics of different regions within a certain region.

[0003] In related technologies, to improve the accuracy of population distribution data, most methods rely on a single data source or use simple linear models for prediction. These methods are difficult to accurately reflect the detailed population distribution patterns within a region under complex geographical and socio-economic conditions, resulting in limited accuracy of the final prediction results. That is, there is a technical problem of low accuracy and poor adaptability of population distribution data in related technologies. SUMMARY

[0004] A series of simplified concepts are introduced in the summary part of the present application, which will be further described in detail in the specific embodiment part. The summary part of the present application does not mean to attempt to limit the key features and necessary technical features of the claimed technical solution, nor does it mean to attempt to determine the protection scope of the claimed technical solution.

[0005] The population distribution data determination method and related equipment provided by the present application can optimize the accuracy and quality of population data in different regions by dividing the regions according to population aggregation and using adaptive algorithms for data integration, thereby improving the accuracy of population multi-source data fusion.

[0006] In a first aspect, the present application provides a population distribution data determination method, comprising: obtaining a multi-source population data set of a target region; dividing the target region into sub-regional categories according to population aggregation, obtaining N preset sub-regions with different aggregation degrees, where N is a natural number greater than 1; dividing the multi-source population data set based on the N preset sub-regions, obtaining N sub-regional data sets corresponding to the N preset sub-regions; using a preset algorithm corresponding to the aggregation degree, performing data integration on the N sub-regional data sets respectively, obtaining N sub-regional population distribution data; and merging the N sub-regional population distribution data, obtaining target population distribution data of the target region.

[0007] In some embodiments, the data partitioning of the multi-source population data set based on the N types of preset sub-regions to obtain N sub-region data sets corresponding to the N types of preset sub-regions comprises: data partitioning of the multi-source population data set and an auxiliary variable data set of the target region based on the N types of preset sub-regions to obtain N sub-region population data sets and N sub-region auxiliary data sets, wherein the auxiliary variable data set comprises an impervious surface data set and a mobile phone positioning data set; and merging each sub-region population data set with a corresponding sub-region auxiliary data set based on the N types of preset sub-regions to obtain N sub-region data sets corresponding to the N types of preset sub-regions.

[0008] In some embodiments, the sub-region category division of the target region according to population aggregation degrees to obtain N types of preset sub-regions with different aggregation degrees comprises: sub-region division of the target region according to a preset administrative division to obtain M initial sub-regions, wherein M is a natural number greater than 1; calculating M population aggregation degrees corresponding to the M initial sub-regions according to M population data and M area data corresponding to the M initial sub-regions; and dividing the M initial sub-regions into N types of preset sub-regions with different aggregation degrees based on the M population aggregation degrees and N-1 preset aggregation degree thresholds.

[0009] In some embodiments, the calculation of M population aggregation degrees corresponding to the M initial sub-regions according to M population data and M area data corresponding to the M initial sub-regions comprises: ratio calculation of each population number and corresponding area data based on the M initial sub-regions to obtain M first ratios corresponding to the M initial sub-regions; ratio calculation of the sum of the M population numbers and the sum of the M area data to obtain a second ratio; and ratio calculation of the M first ratios and the second ratio in sequence to obtain M population aggregation degrees corresponding to the M initial sub-regions.

[0010] In some embodiments, the N is 3, the preset aggregation degree threshold includes a first threshold and a second threshold, wherein the first threshold is greater than or equal to a lower quintile of the aggregation degree and less than or equal to an upper tertile of the aggregation degree, the lower quintile of the aggregation degree is a value at a fifth percentile point in an ascending order of the M population aggregation degrees, and the upper tertile of the aggregation degree is a value at a third percentile point in the ascending order of the M population aggregation degrees; the second threshold is greater than or equal to an upper quintile of the aggregation degree and less than or equal to a lower tertile of the aggregation degree, the upper quintile of the aggregation degree is a value at a fifth percentile point in a descending order of the M population aggregation degrees, and the lower tertile of the aggregation degree is a value at a third percentile point in the descending order of the M population aggregation degrees.

[0011] In some embodiments, the method further includes: performing cross-validation and error index evaluation of the candidate algorithm modeling for data corresponding to the aggregation degree in population distribution training data, obtaining algorithm accuracy corresponding to the candidate algorithm, and determining the candidate algorithm with the algorithm accuracy greater than a preset accuracy as the preset algorithm corresponding to the aggregation degree, wherein a spatial resolution of the population distribution training data is lower than a spatial resolution of the target population distribution data.

[0012] In some embodiments, the candidate algorithm includes a generalized regression algorithm, a variance weighted algorithm, and a random forest algorithm.

[0013] In a second aspect, the present application further provides a population distribution data determination apparatus, including: a data acquisition unit configured to acquire a multi-source population data set of a target region; a region division unit configured to divide the target region into preset sub-regions according to population aggregation degrees, to obtain N preset sub-regions with different aggregation degrees, wherein N is a natural number greater than 1; a data division unit configured to divide the multi-source population data set based on the N preset sub-regions, to obtain N sub-region data sets corresponding to the N preset sub-regions; a data integration unit configured to use a preset algorithm corresponding to the aggregation degree to integrate the N sub-region data sets respectively, to obtain N sub-region population distribution data; and a data merging unit configured to merge the N sub-region population distribution data, to obtain target population distribution data of the target region.

[0014] In a third aspect, the present application further provides an electronic device, including a memory and a processor, wherein the processor is configured to execute a computer program stored in the memory to realize the steps of the population distribution data determination method of the first aspect.

[0015] In a fourth aspect, the present application also provides a computer readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the population distribution data determination method of the first aspect.

[0016] In a fifth aspect, the present application also provides a computer program product comprising a computer program or computer executable instructions, which, when executed by a processor, implement the population distribution data determination method provided by the embodiments of the present application.

[0017] To sum up, the present application can select appropriate processing strategies for different types of population distribution areas, such as densely populated areas, uniform areas, and sparse areas, so that the data processing is more in line with the actual characteristics of the area, avoiding the deviation caused by the difference in regional characteristics, thereby improving the accuracy of population data in different areas. In addition, different preset algorithms are used for data integration for areas with different degrees of aggregation, which can ensure that the data fusion process can be optimized according to the characteristics of various types of areas, thereby improving the accuracy of the integrated population distribution data. The use of multi-source population data sets for detailed integration and processing of population distribution in the target region can not only integrate data from different sources, but also optimize the fusion results according to the characteristics of each sub-region, thereby improving the quality and detail reflecting ability of the integrated data. To sum up, the population distribution data determination method provided by the present application divides the region according to the population aggregation degree and uses adaptive algorithms for data integration, which optimizes the accuracy and quality of population data in different regions and improves the accuracy of multi-source population data fusion. BRIEF DESCRIPTION OF DRAWINGS

[0018] Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The accompanying drawings are included to provide a description of the preferred embodiments and are not intended to limit the scope of the present application. Moreover, the same reference numerals are used throughout the various drawings to designate identical or similar components. In the drawings:

[0019] Figure 1 A flowchart of a population distribution data determination method provided by an embodiment of the present application;

[0020] Figure 2 A schematic diagram of the composition structure of a population distribution data determination device provided by an embodiment of the present application;

[0021] Figure 3 A schematic diagram of the composition structure of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0022] The terms "first", "second", "third", "fourth" etc. (if any) in the description, claims, and drawings of the present application are used for distinguishing between similar objects, not necessarily described in the described order, and are therefore used to identify the embodiments being described. It is understood that these terms are interchangeable under appropriate circumstances such that the embodiments of the present application described herein are capable of operating in other sequences than described or otherwise understood from the present description. Also, the use of the terms "is" and "have" and any of their conjugations herein is not intended to refer exclusively to direct and / or physical ownership, but can also refer to indirect and / or functional ownership. For example, a process, method, system, product, or apparatus that includes several steps or units is not necessarily limited to the steps or units explicitly listed, but can also include other steps or units not explicitly listed, or inherent to such process, method, system, product, or apparatus.

[0023] In the present application, "module" or "unit" refers to a computer program or a part of a computer program with specific functions, and works in cooperation with other related parts to achieve predetermined goals. These modules or units can be implemented by software, hardware (such as processing circuitry or memory), or a combination of both. One or more processors or memories can implement one or more modules or units. At the same time, each module or unit can also be part of a larger module or unit.

[0024] The technical solutions in the present application will be described in detail below in conjunction with the accompanying drawings in the embodiments. It should be noted that the described embodiments are only a part of the present application, not all embodiments. In the following description, "some embodiments" mentioned is only a subset of all possible embodiments, which can be the same or different subset, and different embodiments can be combined with each other without conflict.

[0025] Figure 1 is a flowchart of a population distribution data determination method provided by an embodiment of the present application. For example, see Figure 1 The population distribution data determination method provided by the embodiments of the present application can include the following steps 101 to 105:

[0026] Step 101, obtaining a multi-source population data set of a target region;

[0027] In some examples, the target region refers to a geographical area of interest in the population distribution data analysis process, which can be a city, a province, a country or an entire continent, etc. The boundary and scope of the target region can be determined by geographic information system (GIS) data, administrative division data, online map services, etc. The multi-source population data set refers to population-related data from multiple different channels and sources, including population size, density, distribution, socio-economic characteristics, etc. These data can be obtained through annual population census data of the national statistical bureau, satellite remote sensing data, mobile positioning data, social media data, urban traffic data, etc. For example, the World Population (WorldPop), the Global Population of the World (GPW), the United Nations Environment Programme / Global Resource Information Database (UNEP / GRID), the LandScan Global Population Distribution Data, the Global Human Settlement Population (GHS-POP), and the China Population Spatial Distribution Kilometer Grid Dataset, etc.

[0028] By implementing step 101, the advantages of different data sources can be fully utilized, and various types of data such as official statistical data, satellite data, mobile data, etc. can be integrated to enhance the completeness and diversity of the data, providing rich raw data for subsequent data fusion and analysis, and ensuring that the results are more accurate and reliable.

[0029] Step 102, dividing the target region into sub-regional categories according to the population concentration degree, obtaining N types of preset sub-regions with different concentration degrees, wherein N is a natural number greater than 1.

[0030] In some examples, the population aggregation degree refers to the population density in a sub-region, which can be measured by the number of people per square kilometer or the ratio of population to land area. A region with a high aggregation degree usually indicates a high population density, and a region with a low aggregation degree usually indicates a low population density. For example, the aggregation degree of the central administrative district of a city is high, indicating a high population density, while the aggregation degree of the suburbs or rural areas is low. According to the population aggregation degree, the target region can be divided into several regional subsets, each corresponding to a different population aggregation degree category for differentiated data processing. The N preset sub-regions refer to N different types of preset sub-regions divided according to different aggregation degrees, each having similar population aggregation characteristics. For example, N = 3, and the target region is divided into three types of preset sub-regions, which can be divided into high, medium, and low aggregation degree regions according to the three levels of population aggregation degree.

[0031] By implementing step 102, the region is divided according to the population aggregation degree, and the data is processed differently according to the regional characteristics, avoiding the "one-size-fits-all" processing method, making the data analysis more in line with the actual situation, and reasonably considering the population distribution characteristics of different regions, which can effectively reduce errors and biases caused by regional characteristic differences.

[0032] Step 103, data division of the multi-source population data set based on N preset sub-regions, obtaining N sub-region data sets corresponding to N preset sub-regions;

[0033] In some examples, the N preset sub-regions in the previous step can be used for classification, and each data in the multi-source population data set of the target region is matched with the corresponding sub-region to obtain N sub-region data sets.

[0034] For example, assume that the multi-source population data set of the target region includes resident data of multiple administrative districts, and these administrative districts are divided into high, medium, and low aggregation degree regions according to population density. After data division, the data of the high aggregation degree region is divided into a separate sub-region data set, and the data of the medium and low aggregation degree regions each corresponds to a sub-region data set.

[0035] By implementing step 103, the multi-source population data set is divided according to the preset sub-region, ensuring that each sub-region data set can focus on the characteristics of the region, improving the regional adaptability of the data, and avoiding the mixing and inconsistency caused by regional differences in overall data processing.

[0036] Step 104, using a preset algorithm corresponding to the aggregation degree to integrate the N sub-region data sets respectively, obtaining N sub-region population distribution data;

[0037] In some examples, the preset algorithm is a suitable algorithm selected for different levels of aggregation to perform data processing and integration. The preset algorithm can be based on statistical methods, machine learning, etc., and can provide the best data processing strategy according to the aggregation level of the sub-region. For example, for high-aggregation urban areas, a complex machine learning algorithm such as random forest algorithm can be used, while for sparsely populated suburban areas, linear regression or simple weighted average method can be selected to integrate data. Data integration is the collection of multiple source sub-region data sets into a single sub-region population data for subsequent analysis or result generation. For example, for a high-aggregation area of a city, population data from A channel, B channel and C channel can be integrated to accurately estimate the population data of the area. N sub-region population distribution data refers to the statistical results of the population distribution of each preset sub-region after data integration. These data reflect the number, density and other related characteristics of the internal population of N types of preset sub-regions.

[0038] Through the implementation of step 104, data integration is performed using suitable preset algorithms according to the aggregation level of different regions, which better reflects the special needs and data characteristics of each region, thereby improving the accuracy of each sub-region population distribution data and ensuring that the integrated data is highly consistent with the actual situation.

[0039] Step 105, merging N sub-region population distribution data to obtain target population distribution data of the target region;

[0040] In some examples, data merging is the combination of integrated sub-region population distribution data to form complete population distribution data of the entire target region. The merging process accurately splices the statistical results of each region according to the geographical location and related characteristics of each sub-region, ensuring seamless connection and overall consistency of the data. The target population distribution data is the complete data set after merging, reflecting the population distribution of each region in the entire target region.

[0041] For example, in urban planning and infrastructure construction, the target population distribution data of the target region can be used by the government to reasonably plan residential areas, commercial areas, and transportation networks, thereby optimizing the allocation of urban resources; for public service resource allocation, the data can guide the layout of medical, educational, and police facilities to ensure balanced and efficient services. In emergency response and disaster management, the target population distribution data of the target region can be used to evaluate disaster-affected areas and prioritize the allocation of rescue resources, thereby improving emergency response efficiency. In addition, business decision-makers can use population distribution data for market analysis to select the best store location or promotion area, thereby improving market coverage and profit space; the transportation department can analyze traffic flow based on the target population distribution data of the target region, optimize traffic routes and public transportation layout, effectively alleviate congestion, and improve travel efficiency. In summary, target population distribution data can provide precise decision support for cities, public services, businesses, and transportation, promoting the sustainable development of society and economy.

[0042] Through the implementation of step 105, the final result is more complete and comprehensive, and can reflect the actual population distribution of the entire target region, providing high-quality data support for subsequent applications.

[0043] In summary, the embodiments of the present application divide the target region into sub-regions according to population aggregation, which can select appropriate processing strategies for different types of population distribution areas, such as densely populated areas, uniform areas, and sparse areas, so that data processing is more in line with the actual characteristics of the region, avoiding biases caused by regional characteristics differences, thereby improving the accuracy of population data in different regions; and for regions with different aggregation degrees, different preset algorithms are used for data integration, which can ensure that the data fusion process can be optimized according to the characteristics of each type of region, thereby improving the accuracy of the integrated population distribution data; using multiple source population data sets to integrate the population distribution of the target region in detail can not only integrate data from different sources, but also optimize the fusion results according to the characteristics of each sub-region, improving the quality and detail reflecting ability of the integrated data. In summary, the population distribution data determination method provided by the embodiments of the present application divides the region according to population aggregation and uses adaptive algorithms for data integration, optimizing the accuracy and quality of population data in different regions and improving the accuracy of population multi-source data fusion.

[0044] In some embodiments, the foregoing step 103 can include: based on the foregoing N types of preset sub-regions, performing data division on the multi-source population data set and the auxiliary variable data set of the target region to obtain N sub-regional population data sets and N sub-regional auxiliary data sets, wherein the auxiliary variable data set can include an impervious surface data set and a mobile phone positioning data set; based on the N types of preset sub-regions, merging each sub-regional population data set with the corresponding sub-regional auxiliary data set to obtain N sub-regional data sets corresponding to the N types of preset sub-regions.

[0045] In some examples, the auxiliary variable data set is a set of additional information used to supplement and enhance the population distribution data, which can be used to further analyze the population characteristics of the sub-regions, and can provide environmental factors or behavioral data that affect population distribution; for example, the auxiliary variable data set can include data related to land use, traffic flow, infrastructure construction, etc., such as the impervious surface data set and the mobile phone positioning data set; the impervious surface data set refers to data reflecting the coverage of impervious surfaces in a certain region, such as the coverage of artificial structures such as buildings and roads, through which the urbanization level and land use type of the region can be inferred, which is closely related to population distribution characteristics such as population density; the mobile phone positioning data set includes data from mobile devices, which can be used to analyze people's movement trajectories and stay times, and through the GPS information of the mobile phone, the activity range and density distribution of the crowd in different time periods can be accurately captured, and the population distribution of certain sub-regions can be inferred. The N sub-regional data sets refer to the data sets corresponding to each sub-region obtained from the multi-source population data set and the auxiliary variable data set through the foregoing data division step, and each sub-regional data set contains population distribution data and related auxiliary data in the region.

[0046] Through the implementation of the foregoing embodiments, when performing data division, the addition of auxiliary variables can further improve the fineness of data processing, and through combination with the population data set, a more comprehensive sub-regional data set can be obtained, which helps to optimize the data fusion result, thereby improving the quality of the population distribution data of the target region.

[0047] In some embodiments, the foregoing step 102 can include: dividing the target region into M initial sub-regions according to a preset administrative division, wherein M is a natural number greater than 1; calculating M population concentrations corresponding to the M initial sub-regions according to M population data and M area data corresponding to the M initial sub-regions; and based on the M population concentrations and N-1 preset concentration thresholds, dividing the M initial sub-regions into N types of preset sub-regions with different concentration degrees.

[0048] In some examples, the preset administrative division refers to an administrative division unit set in advance in actual regional division, such as a province, a city, a county, and the like, and a regional division made based on geographical and administrative management factors. The M initial sub-regions are preliminary sub-regions obtained by dividing the target region according to the preset administrative division, the boundary of each initial sub-region is consistent with the administrative division, and the number of these sub-regions is M, M being a natural number greater than 1; for example, the preset administrative division of a city can divide it into multiple districts, such as district A, district B, district C, and the like, and each district is an initial sub-region. The M population data are total population quantity data in each sub-region corresponding to the M initial sub-regions, the population data can be obtained from the National Bureau of Statistics, or the population data of any single source in a multi-source population data set; the M area data are area information of each sub-region corresponding to the M initial sub-regions, and the area data can be obtained through a map service or a GIS system. The N-1 preset aggregation degree thresholds refer to standards for dividing the target region into N types of preset sub-regions according to population aggregation degree, the preset aggregation degree thresholds can divide these initial sub-regions into different categories, and the preset aggregation degree thresholds can be set based on certain statistical standards, such as quantile of aggregation degree, and the like; for example, when N is 3, the preset aggregation degree thresholds can take two numerical values of 0.5 and 2.

[0049] Through implementation of the above embodiments, initial sub-region division is performed based on a preset administrative division, and then refined according to population aggregation degree, and classification is performed using different aggregation degree thresholds, so that different types of regions can be defined more finely, and the characteristics of each type of region in the data processing and analysis process are reasonably considered.

[0050] In some embodiments, the aforementioned calculation of the M population aggregation degrees corresponding to the M initial sub-regions based on the M population data and the M area data corresponding to the M initial sub-regions can include: based on the M initial sub-regions, performing ratio calculation on each population quantity and the corresponding area data to obtain M first ratios corresponding to the M initial sub-regions; performing ratio calculation on the sum of the M population quantities and the sum of the M area data to obtain a second ratio; and performing ratio calculation on the M first ratios and the second ratio in sequence to obtain the M population aggregation degrees corresponding to the M initial sub-regions.

[0051] In some examples, the M first ratios are calculated by calculating the ratio of the population number to the corresponding area, i.e. the population density of each initial sub-region, for the M initial sub-regions. For example, assuming that an initial sub-region has an area of 10 square kilometers and a population of 50,000, the first ratio is 5,000 people per square kilometer. The M first ratios are calculated in this way for all M initial sub-regions. The second ratio is calculated by calculating the ratio of the total population number to the total area number for all initial sub-regions, i.e. the global population density. For example, assuming that the total population of the M initial sub-regions is 10 million and the total area is 500 square kilometers, the second ratio is 20,000 people per square kilometer. The M population concentrations corresponding to the M initial sub-regions are calculated by sequentially calculating the ratio of each of the M first ratios to the second ratio. For example, assuming that the first ratio is 5,000 people per square kilometer and the second ratio is 20,000 people per square kilometer, the concentration of the sub-region is 0.25. If the first ratio of a certain initial sub-region is 30,000 people per square kilometer and the second ratio is 20,000 people per square kilometer, the concentration of the initial sub-region is 1.5. In this way, the concentration of each initial sub-region is calculated, and the concentration of each initial sub-region reflects the degree of change in the population density of the region relative to the overall population density.

[0052] Through implementation of the above embodiments, the M population concentrations calculated can assign a reasonable concentration value to each sub-region, providing a more accurate basis for subsequent regional division and data integration. Combining the population density and the area of the region can more accurately reflect the population concentration of each region, improving the accuracy of regional division.

[0053] In some embodiments, the aforementioned N is 3, and the preset concentration threshold can include a first threshold and a second threshold. The first threshold is greater than or equal to the lower quintile of the concentration and less than or equal to the lower tertile of the concentration. The lower quintile of the concentration is a value at the one-fifth quantile point in the ascending order of the M population concentrations, and the lower tertile of the concentration is a value at the one-third quantile point in the ascending order of the M population concentrations. The second threshold is greater than or equal to the upper tertile of the concentration and less than or equal to the upper quintile of the concentration. The upper quintile of the concentration is a value at the one-fifth quantile point in the descending order of the M population concentrations, and the upper tertile of the concentration is a value at the one-third quantile point in the descending order of the M population concentrations.

[0054] In some examples, the lower quintile of the population aggregation degree refers to the data value at the fifth position after all M population aggregation degrees are arranged in ascending order, representing the smallest 20% of the population aggregation degree values, which can be used to reflect the lower range of the population aggregation degree of the region; for example, assuming that the values of the M population aggregation degrees arranged in ascending order are: [200, 300, 500, 700, 900, 1000], then the lower quintile of the population aggregation degree is the value at the fifth position, which is approximately the first data point, i.e., 200. The lower tertile of the population aggregation degree refers to the data value at the third position after all M population aggregation degrees are arranged in ascending order, representing the smallest 33% of the population aggregation degree values of the region; for example, continuing to assume that the values of the M population aggregation degrees arranged in ascending order are: [200, 300, 500, 700, 900, 1000], then the lower tertile of the population aggregation degree is the second data point, i.e., 300. The upper tertile of the population aggregation degree refers to the data value at the third position after all M population aggregation degrees are arranged in descending order; for example, assuming that the values of the M population aggregation degrees arranged in descending order are: [1000, 900, 700, 500, 300, 200], then the upper tertile of the population aggregation degree is the second data point, i.e., 900. The upper quintile of the population aggregation degree refers to the data value at the fifth position after all M population aggregation degrees are arranged in descending order; for example, assuming that the values of the M population aggregation degrees arranged in descending order are: [1000, 900, 700, 500, 300, 200], then the upper quintile of the population aggregation degree is the first data point, i.e., 1000. The first threshold value is between the lower quintile of the population aggregation degree and the lower tertile of the population aggregation degree, which is used as the boundary for dividing the low population aggregation degree region; the second threshold value is between the upper tertile of the population aggregation degree and the upper quintile of the population aggregation degree, which is used as the boundary for dividing the high population aggregation degree region.

[0055] Through the implementation of the above embodiments, it can be ensured that each sub-region is divided according to certain standards, further reducing human bias and improving the objectivity and consistency of data processing, thereby enhancing the reliability of the results; by dividing the target region into three different aggregation degree levels, such as the population sparse area, the population uniform area, and the population dense area, it is helpful to avoid overly complex or overly rough classification strategies while ensuring the degree of detail of data classification; if N=2 is selected, although it is simple, it may not be able to fully capture the unique characteristics of the medium aggregation degree region, thereby affecting the accuracy of the data; and selecting N=4 may be too detailed, resulting in unnecessary complexity in data processing in some regions, increasing the computational burden and execution cost of the algorithm.

[0056] In some embodiments, the method can further comprise: performing cross-validation and error index evaluation of the candidate algorithm modeling for data corresponding to the degree of concentration in the population distribution training data, obtaining algorithm accuracy corresponding to the candidate algorithm, and determining the candidate algorithm with algorithm accuracy greater than a preset accuracy as the preset algorithm corresponding to the degree of concentration.

[0057] In some examples, a plurality of candidate machine learning or statistical models can be selected first and applied to a data set composed of any degree of concentration in the population distribution training data to build the model; then the models are trained and tested cyclically by dividing the data set into several small subsets to ensure the generalization ability of the model on different data partitions; finally, the prediction accuracy of the model is measured using indicators such as Mean Square Error (MSE) or Mean Absolute Error (MAE); for example, assuming that random forest algorithm and generalized regression algorithm are selected as candidate algorithms, by cross-validation, the training data set with high population density in the population distribution training data is divided into 5 subsets, 4 of which are used to train the model in turn, and the remaining 1 subset is used for testing. This process can continue for 5 times, with a different subset being verified each time, and the performance data of all candidate algorithms is finally obtained. Algorithm accuracy can be quantified by the error index of the algorithm model, and the higher the accuracy, the more accurate the fitting and prediction of the candidate algorithm on the data. The preset accuracy is a predefined standard value for comparison with the accuracy of the candidate algorithm, and if the accuracy of the candidate algorithm is higher than the preset value, the algorithm is considered to be qualified and can be used for subsequent processing; for example, the preset accuracy can be set to 2.0, which means that the error of any candidate algorithm needs to be less than or equal to 2.0 before the candidate algorithm can be used as a preset algorithm for actual data processing. The population distribution training data is historical or simulated geographic area population distribution data, which usually contains specific population numbers and other related variables within the area; spatial resolution refers to the spatial accuracy of the data, which refers to the fineness of each unit area in the population distribution data; the spatial resolution of the population distribution training data is relatively rough, while the target population distribution data requires higher accuracy; for example, assuming that the spatial resolution of the target population distribution data is 1 km2, while the resolution of the training data can be 5 km2, or the spatial resolution of the target population distribution data is county-level city, while the resolution of the training data can be province.

[0058] Through the implementation of the above embodiments, the applicability and accuracy of the selected algorithm can be ensured, the needs of different application scenarios can be met, and the reliability of the data integration result can be further improved.

[0059] In some embodiments, the foregoing alternative algorithms can include a generalized regression algorithm, a variance weighted algorithm, and a random forest algorithm.

[0060] In some examples, the generalized regression (GR) algorithm assumes a linear relationship between different model products and calibration data, and uses a multivariate linear regression equation for fitting. In order to avoid estimation bias caused by strong covariance between input data, i.e., the variance inflation factor is greater than 10, a ridge regression method can be used to solve the equation. The variance weighted (VW) algorithm is an extension of the Bates-Granger algorithm, mainly including two steps of bias removal and weighted linear average. The calculation of the weight is based on the uncertainty between the calibration data and the input data, which is represented by the variance and covariance. In this algorithm, the relationship between the input data and the bias is weighted and averaged to reduce the impact of errors on the final result. In order to ensure the accuracy of the algorithm, a partition modeling strategy is used, and the extrapolated results outside the partition are used to fill in the missing data to improve the integrity and accuracy of the data. The random forest (RF) algorithm is a nonlinear, nonparametric modeling method that generates multiple decision trees through the bagging method. Each tree randomly selects a feature combination for splitting during training, and the final output result is obtained by averaging the prediction results of all trees. The random forest algorithm can effectively handle high-dimensional data and reduce the risk of overfitting through ensemble learning.

[0061] It should be noted that the generalized regression algorithm can effectively handle problems with multicollinearity, ensuring accurate fitting of population data in the presence of covariance, and can be applied to relatively linear and regular areas. The variance weighted algorithm can effectively reduce the impact of data uncertainty through bias removal and weighted linear averaging, so that the difference between the input data and the calibration data is compensated, thereby improving the accuracy and stability of the data, and can be applied to areas with large deviations. The random forest algorithm combines the results of multiple decision trees through the bagging method, avoiding overfitting, and can handle complex nonlinear relationships, and is suitable for areas with complex structures and diverse data.

[0062] Through the implementation of the above embodiments, different algorithms can handle different types of data characteristics. By comparing the performance of each algorithm, the optimal algorithm can be selected to ensure that the final population distribution data has high precision and reliability.

[0063] Furthermore, as an implementation of the aforementioned method embodiments, this application also provides a population distribution data determination apparatus for implementing the aforementioned method embodiments. This apparatus embodiment corresponds to the aforementioned method embodiments. For ease of reading, this population distribution data determination apparatus embodiment will not repeat the details of the aforementioned method embodiments one by one, but it should be understood that the apparatus in this application embodiment can correspondingly implement all the contents of the aforementioned method embodiments. For example... Figure 2 As shown, the population distribution data determination device 20 includes: a data acquisition unit 201, a region division unit 202, a data integration unit 203, a data merging unit 204, and a data merging unit 205. The data acquisition unit 201 is used to acquire multi-source population datasets for the target region. The region division unit 202 is used to divide the target region into sub-region categories according to population concentration, obtaining N preset sub-regions with different concentration levels, where N is a natural number greater than 1. The data integration unit 203 is used to divide the multi-source population datasets based on the N preset sub-regions, obtaining N sub-region datasets corresponding to the N preset sub-regions. The data integration unit 204 is used to integrate the N sub-region datasets using a preset algorithm corresponding to the concentration level, obtaining N sub-region population distribution data. The data merging unit 205 is used to merge the N sub-region population distribution data to obtain the target population distribution data for the target region.

[0064] In some embodiments, the region partitioning unit 202 is further configured to partition the multi-source population dataset and the auxiliary variable dataset of the target region based on N types of preset sub-regions, to obtain N sub-region population datasets and N sub-region auxiliary datasets, wherein the auxiliary variable datasets include impermeable surface datasets and mobile phone positioning datasets; and based on the N types of preset sub-regions, merge each sub-region population dataset with the corresponding sub-region auxiliary dataset to obtain N sub-region datasets corresponding to the N types of preset sub-regions.

[0065] In some embodiments, the region division unit 202 is further configured to divide the target region into sub-regions according to a preset administrative division to obtain M initial sub-regions, where M is a natural number greater than 1; calculate the M population clustering degree corresponding to the M initial sub-regions based on the M population data and M area data corresponding to the M initial sub-regions; and divide the M initial sub-regions into N preset sub-regions with different clustering degrees based on the M population clustering degree and N-1 preset clustering degree thresholds.

[0066] In some embodiments, the region division unit 202 is further configured to calculate a ratio of each population quantity to corresponding area data based on the M initial sub-regions, to obtain M first ratios corresponding to the M initial sub-regions; calculate a ratio of a sum of the M population quantities to a sum of the M area data, to obtain a second ratio; and sequentially calculate a ratio of the M first ratios to the second ratio, to obtain M population concentrations corresponding to the M initial sub-regions.

[0067] In some embodiments, N is 3, and the preset concentration threshold includes a first threshold and a second threshold, where the first threshold is greater than or equal to a lower quintile of the population concentrations and less than or equal to an upper tertile of the population concentrations, the lower quintile of the population concentrations is a value at a fifth percentile point in an ascending order of the M population concentrations, and the upper tertile of the population concentrations is a value at a third percentile point in the ascending order of the M population concentrations; and the second threshold is greater than or equal to an upper quintile of the population concentrations and less than or equal to an upper quintile of the population concentrations, the upper quintile of the population concentrations is a value at a fifth percentile point in a descending order of the M population concentrations, and the upper tertile of the population concentrations is a value at a third percentile point in the descending order of the M population concentrations.

[0068] In some embodiments, the data integration unit 204 is further configured to perform cross-validation and error index evaluation of the candidate algorithm modeling on data corresponding to the concentration degree in the population distribution training data, to obtain an algorithm accuracy corresponding to the candidate algorithm, and determine the candidate algorithm with an algorithm accuracy greater than a preset accuracy as the preset algorithm corresponding to the concentration degree, where a spatial resolution of the population distribution training data is lower than a spatial resolution of the target population distribution data.

[0069] In some embodiments, the candidate algorithm includes a generalized regression algorithm, a variance weighted algorithm, and a random forest algorithm.

[0070] The present application also provides a computer readable storage medium having computer executable instructions or computer programs stored therein, which, when executed by a processor, cause the processor to perform any step of the population distribution data determination method provided by the present application.

[0071] In some embodiments, the computer readable storage medium can be a memory such as a random access memory (RAM), a read-only memory (ROM), a flash memory, a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM), etc.; or can be various devices including one or any combination of the above memories.

[0072] In some embodiments, the computer-executable instructions can take the form of programs, software, software modules, scripts, or code, written in any suitable programming language (including compiled or interpreted languages, or declarative or procedural languages), and can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0073] In some embodiments, the computer-executable instructions can, but need not, correspond to files in a file system, can be stored in files employed by the operating system (for example, stored in one or more scripts in a HyperText Markup Language (HTML) document employed by an Internet browser application), can be stored in a single file dedicated to the program in question, or can be stored across multiple coordinated files (for example, files storing one or more modules, subprograms, or code portions).

[0074] In some embodiments, the computer-executable instructions can be deployed across one electronic device, or across multiple electronic devices located at one site, or distributed across multiple sites and interconnected by a communication network.

[0075] As shown in Figure 3 The present application also provides an electronic device 30, comprising a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor 320, wherein the processor 320 implements any step of the population distribution data determining method when executing the computer program 311.

[0076] The present application also provides a computer program product, comprising a computer program or computer-executable instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer program or computer-executable instructions from the computer-readable storage medium, and the processor executes the computer program or computer-executable instructions, so that the electronic device performs any step of the population distribution data determining method.

[0077] The above embodiments are only used to illustrate the technical solutions of the present application, but not limit the technical solutions; even though the above technical solutions are described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements to some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for determining population distribution data, characterized in that, include: Obtain multi-source population datasets for a target region. The target region refers to the geographical area of ​​interest in the population distribution data analysis process, such as a city, a province, a country, or the entire Earth continent. The boundaries and scope of the target region are determined through geographic information system data, administrative division data, online map services, etc. The multi-source population datasets are population-related data from multiple different channels and sources. The population-related data includes multi-dimensional data such as population size, density, distribution, and socioeconomic characteristics. The target region is divided into sub-regions according to population density, resulting in N preset sub-regions with different degrees of density, where N is a natural number greater than 1. Based on the N preset sub-regions, the multi-source population dataset and the auxiliary variable dataset of the target region are divided into N sub-region population datasets and N sub-region auxiliary datasets. The auxiliary variable datasets include impermeable surface datasets and mobile phone location datasets. The auxiliary variable datasets are additional information sets used to supplement and enhance population distribution data. The impermeable surface datasets reflect the impermeable surface coverage in a certain area, including the coverage of artificial structures such as buildings and roads. The urbanization level and land use type of the area can be inferred from the impermeable surface data. The mobile phone location datasets include data from mobile devices, which are used to analyze people's movement trajectories and dwell time. Through the GPS information of mobile phones, the activity range and density distribution of people in different time periods can be accurately captured. Based on the N types of preset sub-regions, the population dataset of each sub-region is merged with the corresponding sub-region auxiliary dataset to obtain N sub-region datasets corresponding to the N types of preset sub-regions. Using a preset algorithm corresponding to the degree of aggregation, the data of the N sub-region datasets are integrated to obtain population distribution data of the N sub-regions; The population distribution data of the N sub-regions are merged to obtain the target population distribution data of the target region. The method for determining population distribution data further includes: performing cross-validation and error index evaluation on the data in the population distribution training data that corresponds to the degree of clustering, obtaining the algorithm accuracy corresponding to the candidate algorithm, and determining the candidate algorithm whose algorithm accuracy is greater than the preset accuracy as the preset algorithm corresponding to the degree of clustering, wherein the spatial resolution of the population distribution training data is lower than the spatial resolution of the target population distribution data.

2. The method for determining population distribution data according to claim 1, characterized in that, The step of dividing the target region into sub-regions according to population density to obtain N preset sub-regions with different degrees of density includes: The target region is divided into sub-regions according to a preset administrative division, resulting in M ​​initial sub-regions, where M is a natural number greater than 1; Based on the M population data and M area data corresponding to the M initial sub-regions, the M population clustering degrees corresponding to the M initial sub-regions are calculated. Based on the M population clustering degrees and N-1 preset clustering degree thresholds, the M initial sub-regions are divided into N preset sub-regions with different clustering degrees.

3. The method for determining population distribution data according to claim 2, characterized in that, The step of calculating the M population clustering degrees corresponding to the M initial sub-regions based on the M population data and M area data corresponding to the M initial sub-regions includes: Based on the M initial sub-regions, the ratio of each population size to the corresponding area data is calculated to obtain the M first ratios corresponding to the M initial sub-regions; The second ratio is obtained by calculating the ratio of the sum of the M population counts to the sum of the M area data. The ratios of the M first ratios and the second ratios are calculated sequentially to obtain the population concentration of M people corresponding to the M initial sub-regions.

4. The method for determining population distribution data according to claim 2, characterized in that, The value of N is 3. The preset clustering threshold includes a first threshold and a second threshold. The first threshold is greater than or equal to the lower quintile of the clustering degree and less than or equal to the lower tertiary of the clustering degree. The lower quintile of the clustering degree is the value located at the 1 / 5 quartile in the ascending order of the clustering degree of the M populations, and the lower tertiary of the clustering degree is the value located at the 1 / 3 quartile in the ascending order of the clustering degree of the M populations. The second threshold is greater than or equal to the upper tertiary of the clustering degree and less than or equal to the upper quintile of the clustering degree. The upper quintile of the clustering degree is the value located at the 1 / 5 quartile in the descending order of the clustering degree of the M populations, and the upper tertiary of the clustering degree is the value located at the 1 / 3 quartile in the descending order of the clustering degree of the M populations.

5. The method for determining population distribution data according to claim 1, characterized in that, The alternative algorithms include generalized regression, variance-weighted algorithm, and random forest algorithm.

6. A population distribution data determination device, characterized in that, include: The data acquisition unit is used to acquire multi-source population datasets of a target region. The target region refers to the geographical area of ​​interest in the population distribution data analysis process, such as a city, a province, a country, or the entire Earth continent. The boundaries and scope of the target region are determined through geographic information system data, administrative division data, online map services, etc. The multi-source population dataset is population-related data from multiple different channels and sources. The population-related data includes multi-dimensional data such as population size, density, distribution, and socioeconomic characteristics. The region division unit is used to divide the target region into sub-region categories according to the population concentration, resulting in N preset sub-regions with different concentration levels, where N is a natural number greater than 1; A data partitioning unit is used to partition the multi-source population dataset and the auxiliary variable dataset of the target region based on the N preset sub-regions, resulting in N sub-region population datasets and N sub-region auxiliary datasets. The auxiliary variable datasets include impermeable surface datasets and mobile phone location datasets. The auxiliary variable datasets are additional information sets used to supplement and enhance population distribution data. The impermeable surface dataset reflects the impermeable surface coverage in a certain area, including the coverage of artificial structures such as buildings and roads. The urbanization level and land use type of the area can be inferred from the impermeable surface data. The mobile phone location dataset includes data from mobile devices used to analyze people's movement trajectories and dwell time. Using the GPS information of mobile phones, the activity range and density distribution of people in different time periods can be accurately captured. Based on the N preset sub-regions, the population dataset of each sub-region is merged with the corresponding sub-region auxiliary dataset to obtain N sub-region datasets corresponding to the N preset sub-regions. The data integration unit is used to integrate the data of the N sub-region datasets respectively using a preset algorithm corresponding to the degree of aggregation, so as to obtain the population distribution data of the N sub-regions; A data merging unit is used to merge the population distribution data of the N sub-regions to obtain the target population distribution data of the target region. The data integration unit is further configured to perform cross-validation and error index evaluation on the data in the population distribution training data corresponding to the degree of clustering, to obtain the algorithm accuracy corresponding to the candidate algorithm, and to determine the candidate algorithm whose algorithm accuracy is greater than the preset accuracy as the preset algorithm corresponding to the degree of clustering, wherein the spatial resolution of the population distribution training data is lower than the spatial resolution of the target population distribution data.

7. An electronic device, comprising: A memory and a processor, characterized in that the processor, when executing a computer program stored in the memory, implements the steps of the population distribution data determination method as described in any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the population distribution data determination method as described in any one of claims 1-5.

Images