CN115525872B - Two-step Bayesian estimation method for building scale population fused with position data - Google Patents
Two-step Bayesian estimation method for building scale population fused with position data Download PDFInfo
- Publication number
- CN115525872B CN115525872B CN202211262928.8A CN202211262928A CN115525872B CN 115525872 B CN115525872 B CN 115525872B CN 202211262928 A CN202211262928 A CN 202211262928A CN 115525872 B CN115525872 B CN 115525872B
- Authority
- CN
- China
- Prior art keywords
- population
- grid
- building
- probability
- residential
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
Abstract
The invention discloses a two-step Bayesian estimation method, a system, electronic equipment and a storage medium for building scale population fusing position data. Comprising the following steps: determining demographic-allocation assistance data; determining a signaling classification category and an auxiliary classification category of each grid; determining the probability of grid occurrence of each signaling classification category respectively, determining the probability of grid occurrence of each auxiliary classification category under different signaling classification categories respectively, and obtaining corresponding population posterior probability according to the auxiliary classification likelihood and the corresponding signaling classification prior probability; obtaining a grid expected population corresponding to the auxiliary classification category according to the population posterior probability and the population average value of each signaling classification category; determining an assigned population probability for the corresponding grid; and obtaining the predicted population of the corresponding grid according to the population probability and the population of the street where the corresponding grid is located. The population number of the building level can be predicted, the population estimation accuracy is improved, and the application range is wider.
Description
Technical Field
The invention relates to the technical field of electronic maps, in particular to a two-step Bayesian estimation method, a system, electronic equipment and a storage medium for building scale population fusing position data.
Background
Population data is one of the most important basic geographic data, accurate specific population data information is mastered, the method has important significance for solving various population related researches and applications, the refined population data which is needed to be used as the basic data in refined city management is urgent, and high-resolution population data is important for knowing dynamic environment so as to accurately conduct city and space planning; the possibility is provided for the development index with reliable design; the method is also beneficial to optimizing the intervention measures of communities, improving the accurate response capability to natural disasters, achieving the prevention in advance without the rain and the silk, and has important significance for practically solving the target problem.
The existing population space-time pattern prediction is characterized in that the population characteristics of a research area in time, space and space interaction dimensions are not comprehensively considered for the characteristic exploration of the population in time, space and space interaction dimensions on the basis of population aggregation area identification based on mobile phone signaling data, population distribution characteristic analysis, dynamic population distribution drawing, population activity space-time characteristic identification, day-night space dynamic distribution research and population space structure evolution characteristic exploration based on a metering model. In population prediction, population total amount or population structure of a large area is mainly focused on prediction, and also, a scholars perform population simulation prediction based on small-scale population spatialization, such as a population distribution surface modeling method, a 3G method (GIS-Geographic Information System-geographic information system, GP-Genetic Programming-genetic programming, GA-Genetic Algorithms-genetic algorithm) and a deep neural network, and a CA-Markov modeling method, but the key point of the population distribution surface modeling method is to determine a reasonable contribution value distribution surface, and the contribution value distribution surface needs to be obtained by combining auxiliary materials such as various historical information, so that in the case of insufficient auxiliary materials, the method is difficult to obtain the population distribution surface which is relatively close to a real situation. The intelligent algorithm has high automation degree and flexible model construction, but has poor result controllability and complex parameter setting. The CA-Markov model is a dynamic model with discrete time and discrete state based on future population simulation of population density level, and is not suitable for continuous simulation prediction. Most of the existing population spatialization techniques are prediction or evaluation from bottom to top, and observation and verification of population distribution are performed by integrating into larger space units (such as communities and streets), the method is based on the assumption that errors will be inevitably generated in the integrating unit in the flow, and the errors of population distribution in the internal space of the integrating unit will be wiped out. The traditional fine-grained population space allocation problem is mostly estimated or predicted based on single data, which may lead to poor robustness of the result first; secondly, the traditional population space distribution is still a bottom-up distribution mode, the accuracy of the model is verified through the accurate data of the upper layer, but the bottom-up mode has a great defect that only the accuracy of the upper layer is considered, the difference of the space distribution under fine granularity is completely covered, the prior art adopts a one-step in-place mode, the population on the building level is directly predicted, and the prediction result under the mode is often insufficient in accuracy.
Disclosure of Invention
In order to solve the problems, the invention provides a two-step Bayesian estimation method, a system, electronic equipment and a storage medium for building scale population fusing position data.
In a first aspect, an embodiment of the present invention provides a two-step bayesian estimation method for a building scale population fusing position data, including the steps of:
dividing the selected area according to grids according to signaling data of the selected area;
determining the type of each building in each grid according to the interest surface data of the selected area;
dividing residential buildings in the grids according to different residential attributes, and determining the residential calculation area corresponding to each residential building in the corresponding grid and the residential calculation area of the corresponding grid;
obtaining population distribution auxiliary data of the corresponding grids according to the living calculation area of the grids and the nuclear density of interest points of living buildings;
classifying the grids according to population numbers of signaling data of each grid to obtain signaling classification categories of each grid;
classifying the grids according to the population distribution auxiliary data of each grid to obtain auxiliary classification categories of each grid;
Determining the probability of occurrence of grids of each signaling classification category respectively, and marking the probability as signaling classification prior probability;
determining the probability of occurrence of grids of each auxiliary classification category under different signaling classification categories respectively, and marking the probability as auxiliary classification likelihood;
obtaining corresponding population posterior probability according to the auxiliary classification likelihood and the corresponding signaling classification prior probability;
obtaining expected population of grids corresponding to the auxiliary classification category according to the population posterior probability and population average value of each signaling classification category;
determining the distribution population probability of the corresponding grid according to the proportion of the expected population of the grid to the expected population of the grid of the street where the corresponding grid is located;
and obtaining the predicted population of the corresponding grid according to the population probability and the population of the street where the corresponding grid is located.
In some specific embodiments, determining the type of each building in each grid from the surface of interest data for the selected area further comprises the steps of:
when a building positioned outside the interest surface exists in the grid, a buffer zone with a preset width is arranged outside the outline of the building, and the number and types of all interest points in the outline of the building and in the buffer zone are counted;
Respectively giving corresponding weights to each interest point type, wherein the weights of the interest points in the building outline are larger than those of the interest points of the same type in the buffer zone;
multiplying the number of each interest point by corresponding weight to obtain weighted sums of each interest point type in the building outline and the buffer zone respectively;
and marking the type of the interest point with the largest weighted sum as the type of the building.
In some specific embodiments, determining a occupancy calculation area corresponding to each occupancy building in a corresponding grid and a occupancy calculation area of the corresponding grid includes the steps of:
respectively giving corresponding weight to each living attribute;
multiplying the actual building area of each residential building in the grid by the weight of the corresponding residential attribute to obtain the residential calculation area corresponding to each residential building in the grid;
and adding the residence calculation areas corresponding to each residence building in the grids to obtain the residence calculation area of the corresponding grid.
In some specific embodiments, according to the living calculation area of the grid and the nuclear density of the interest points of the living building, the population distribution auxiliary data of the corresponding grid is obtained, and the method comprises the following steps:
And multiplying the living calculation area of the grid by the square root of the product of the nuclear density of the points of interest of the living building of the grid, and recording the square root as the auxiliary data of population distribution corresponding to the living attribute of the grid.
In some specific embodiments, before classifying the grids according to the population of signaling data of each grid, the method comprises the following steps:
comparing the acquired signaling data sample size with population of a selected area, and determining sample expansion parameters;
and according to the sample spreading parameters, the obtained signaling data sample size is subjected to sample spreading to obtain the population of the signaling data of each grid in the selected area.
In some specific embodiments, the classification method for classifying the grids according to the population of the signaling data of each of the grids or according to the population of the population allocation assistance data of each of the grids is a natural break point method.
In some specific embodiments, the number of signaling classification categories is not less than 80 and not greater than 120.
In some specific embodiments, the number of auxiliary classification categories is no less than 250 and no greater than 350.
In some specific embodiments, the two-step bayesian estimation method of building scale population fusing location data further comprises the steps of:
And distributing the predicted population number of the grid of the residential building to be tested according to the proportion of the residential calculation area of the residential building to be tested and the residential calculation area of the grid of the residential building to be tested, so as to obtain the predicted population number of the residential building to be tested.
In some specific embodiments, the two-step bayesian estimation method of building scale population fusing location data further comprises the steps of:
setting a part of the selected areas as verification areas, and acquiring the real population of the verification areas;
recording the absolute value of the difference between the predicted population of the verification area and the real population as a prediction error;
when the prediction error of the current prediction exceeds a preset threshold, marking the posterior population probability of the current prediction as the signaling classification prior probability of the next prediction, and predicting to obtain the predicted population number of the next prediction;
and when the prediction error of the current prediction does not exceed the preset threshold value, the verification is passed, and the prediction is terminated.
In a second aspect, an embodiment of the present invention provides a two-step bayesian estimation system for building scale population fusing location data, including:
The data processing module is used for dividing the selected area according to the signaling data of the selected area; determining the type of each building in each grid according to the interest surface data of the selected area; dividing residential buildings in the grids according to different residential attributes, and determining the residential calculation area corresponding to each residential building in the corresponding grid and the residential calculation area of the corresponding grid; obtaining population distribution auxiliary data of the corresponding grids according to the living calculation area of the grids and the nuclear density of interest points of living buildings; classifying the grids according to population numbers of signaling data of each grid to obtain signaling classification categories of each grid; classifying the grids according to the population distribution auxiliary data of each grid to obtain auxiliary classification categories of each grid;
the grid population prediction module is used for respectively determining the probability of grid occurrence of each signaling classification category and recording the probability as signaling classification prior probability; determining the probability of occurrence of grids of each auxiliary classification category under different signaling classification categories respectively, and marking the probability as auxiliary classification likelihood; obtaining corresponding population posterior probability according to the auxiliary classification likelihood and the corresponding signaling classification prior probability; obtaining expected population of grids corresponding to the auxiliary classification category according to the population posterior probability and population average value of each signaling classification category; determining the distribution population probability of the corresponding grid according to the proportion of the expected population of the grid to the expected population of the grid of the street where the corresponding grid is located; and obtaining the predicted population of the corresponding grid according to the population probability and the population of the street where the corresponding grid is located.
In some specific embodiments, the number of signaling classification categories is not less than 80 and not greater than 120.
In some specific embodiments, the number of auxiliary classification categories is no less than 250 and no greater than 350.
In some specific embodiments, the two-step bayesian estimation system of building scale population fusing location data further comprises:
and the building population prediction module is used for distributing the predicted population of the grid where the residential building to be detected is positioned according to the ratio of the residential calculation area of the residential building to be detected to the residential calculation area of the grid where the residential building to be detected is positioned, so as to obtain the predicted population of the residential building to be detected.
In some specific embodiments, the two-step bayesian estimation system of building scale population fusing location data further comprises:
the prediction population verification module is used for setting a part of the selected area as a verification area and acquiring the real population number of the verification area; recording the absolute value of the difference between the predicted population of the verification area and the real population as a prediction error; when the prediction error of the current prediction exceeds a preset threshold, marking the posterior population probability of the current prediction as the signaling classification prior probability of the next prediction, and predicting to obtain the predicted population number of the next prediction; and when the prediction error of the current prediction does not exceed the preset threshold value, the verification is passed, and the prediction is terminated.
Based on the same inventive concept, an embodiment of the present invention further provides an electronic device, which is characterized by comprising: the system comprises a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor realizes the two-step Bayesian estimation method of the building dimension population fusing position data when executing the computer program.
Based on the same inventive concept, the embodiment of the invention also provides a computer storage medium, which is characterized in that computer executable instructions are stored in the computer storage medium, and the computer executable instructions realize the two-step Bayesian estimation method of the building dimension population fusing the position data when being executed.
The technical scheme provided by the embodiment of the invention has the beneficial effects that at least:
the population distribution is controlled through two layers (street-grid, grid-building), so that the accuracy of population estimation is improved; according to the characteristics of different types of buildings, one or more residence attributes of the grids are determined, the actual situation can be more attached, and population distribution auxiliary data is determined based on residence calculation areas and interest point kernel densities of the residence attributes, so that the prediction accuracy is improved; the population prediction of the grid level is realized, and the method can be widely applied to population prediction of large, medium and small areas; meanwhile, the mobile phone signaling data at the grid level is used as the signaling classification prior probability, so that the subjective judgment of introducing dislocation is avoided, and the actual situation is more fitted; further, through verification, repeated iteration is carried out for optimizing, parameter setting which is more in line with a selected area is found, a more accurate population prediction effect is achieved, a top-down population redistribution model is adopted, so that absolute accuracy of population at an upper layer (street) is guaranteed, and accuracy verification is carried out by taking down data at a lower layer (community or building), so that accuracy of population space distribution at a fine scale is guaranteed; furthermore, the anti-interference performance of the result can be improved by replacing the open-source big data with a single data source, the population prediction area range can be enlarged through sample expansion processing, abnormal value interference is reduced, and the processed data volume is reduced; based on the weighted sum of the interest point types, the buildings outside the interest surface can be reasonably classified, and the prediction accuracy is improved; the natural breakpoint method is used, so that a better estimation result can be realized in the embodiment of the invention than other classification methods; the number of the signaling classification categories and the number of the auxiliary classification categories are in a better range, so that classification homogenization is avoided, and a more accurate prediction effect can be realized; the prediction error can be effectively reduced, and the prediction precision is improved.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be apparent from the description, or may be learned by practice of the embodiments of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims thereof as well as the appended drawings.
The technical scheme of the invention is further described in detail through the drawings and the embodiments.
Drawings
The accompanying drawings are included to provide a further understanding of the invention, and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain, without limitation, the embodiments of the invention. In the drawings:
FIG. 1 is a flowchart of a two-step Bayesian estimation method for building scale population fusing position data in an embodiment of the present invention;
FIG. 2 is a schematic diagram of regression analysis of population predictions using signaling sample expansion data in the prior art;
FIG. 3 is a schematic representation of regression analysis of population predictions from building area estimates using conventional methods;
FIG. 4 is a schematic diagram of regression analysis of a Bayesian model for population prediction in a selected region according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a predicted population for each residential building in a selected area (southwest area of the street) in accordance with an embodiment of the present invention;
FIG. 6 is a schematic diagram of the predicted population error for each residential building in a selected area in accordance with an embodiment of the present invention;
FIG. 7 is a schematic diagram of estimating a selected area building population error by building area in a prior art method;
FIG. 8 is a graph showing the error variance of population prediction according to building area according to the embodiment of the present invention and the prior art method;
FIG. 9 is a schematic diagram of a building according to an embodiment of the present invention estimating population by building area only;
FIG. 10 is a schematic diagram of a Bayesian model to estimate building population according to an embodiment of the present invention;
FIG. 11 is a Bayesian estimation population error histogram in accordance with an embodiment of the present invention;
FIG. 12 is a histogram of population error estimated by building area according to an embodiment of the present invention;
FIG. 13 is a histogram of error variation for two methods according to an embodiment of the present invention;
fig. 14 is a schematic structural diagram of an electronic device according to an embodiment of the invention.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the technical solution of the present invention may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.
In order to solve the problems in the prior art, the embodiment of the invention provides a two-step Bayesian estimation method, a system, electronic equipment and a storage medium for building scale population fusing position data.
Example 1
The first embodiment of the invention provides a two-step Bayesian estimation method for building scale population fusing position data, which comprises the following steps:
step S1: dividing the selected area according to grids according to signaling data of the selected area; determining the type of each building in each grid according to the interest surface data of the selected area; dividing residential buildings in the grids according to different residential attributes, and determining the residential calculation area corresponding to each residential building in the corresponding grid and the residential calculation area of the corresponding grid; and obtaining the population distribution auxiliary data of the corresponding grids according to the living calculation area of the grids and the nuclear density of the interest points of the living building.
The building types include residential buildings, and the residential attributes include ordinary houses, apartments, villas, dormitories, business and living dual-purpose, and the like.
In some specific embodiments, determining the type of each building in each grid from the selected area interest surface data further comprises the steps of:
When a building positioned outside the interest surface exists in the grid, a buffer zone with a preset width is arranged outside the outline of the building, and the number and types of all interest points in the outline of the building and in the buffer zone are counted; respectively giving corresponding weights to each interest point type, wherein the weights of the interest points in the building outline are larger than those of the interest points of the same type in the buffer zone; multiplying the number of each interest point by corresponding weight to obtain weighted sums of each interest point type in the building outline and the buffer zone respectively; and marking the type of the interest point with the largest weighted sum as the type of the building.
In some specific embodiments, determining a occupancy calculation area corresponding to each occupancy building in a corresponding grid and a occupancy calculation area of the corresponding grid includes the steps of:
respectively giving corresponding weight to each living attribute; multiplying the actual building area of each residential building in the grid by the weight of the corresponding residential attribute to obtain the residential calculation area corresponding to each residential building in the grid; and adding the residence calculation areas corresponding to each residence building in the grids to obtain the residence calculation area of the corresponding grid.
In some specific embodiments, according to the living calculation area of the grid and the nuclear density of the interest points of the living building, the population distribution auxiliary data of the corresponding grid is obtained, and the method comprises the following steps:
and multiplying the living calculation area of the grid by the square root of the product of the nuclear density of the points of interest of the living building of the grid, and recording the square root as the auxiliary data of population distribution corresponding to the living attribute of the grid.
Step S2: comparing the acquired signaling data sample size with population of a selected area, and determining sample expansion parameters; and according to the sample spreading parameters, the obtained signaling data sample size is subjected to sample spreading to obtain the population of the signaling data of each grid in the selected area.
The anti-interference performance of the result can be improved by replacing the open source big data with a single data source, and the population prediction area range can be enlarged and abnormal value interference can be reduced through sample expansion processing. The mobile phone signaling data mainly originates from telecom operators (mobile, telecom, unicom), and directly adopting grid (250 m×250 m) data of single operator mobile phone signaling data after desensitization can lead to that the number of mobile phone signaling in the grid at the street level is lower than the seventh population census population of the street. Because the telecom operators are selected to have randomness and obtain more accurate prior probability, the population of each street building body is expanded in a certain proportion based on the existing signaling data as the sharing rate, so that the signaling data of the mobile phone is consistent with the census data of the seventh population at the street level.
Step S3: classifying the grids according to population numbers of signaling data of each grid to obtain signaling classification categories of each grid; and classifying the grids according to the population distribution auxiliary data of each grid to obtain auxiliary classification categories of each grid.
In some specific embodiments, the classification method for classifying the grids according to the population of the signaling data of each of the grids or according to the population of the population allocation assistance data of each of the grids is a natural break point method. The results obtained using the natural breakpoint method classification were found to be relatively superior by trying a variety of classification methods, e.g., quantile classification, equal interval classification, geometric interval, etc.
In some specific embodiments, the number of signaling classification categories is not less than 80 and not greater than 120. In some specific embodiments, the number of auxiliary classification categories is no less than 250 and no greater than 350. Dividing the grids into different categories according to the population of the signaling data of each grid and the population of the signaling data of each grid, and obtaining population reassignment results with different goodness, wherein the number of the categories is not as high as the number of the categories is as high as possible because the samples are limited while avoiding the influence of abnormal values; also not as good as the smaller the number of classifications, the smaller the number of classifications will mask the potential correspondence between signaling class a and auxiliary class B, resulting in a homogeneous final result, which is less well behaved. Through multiple experiments, the number of signaling classification categories is 80 to 120, and when the number of auxiliary classification categories is 250 to 350, the result is better.
Step S4: determining the probability of occurrence of grids of each signaling classification category respectively, and marking the probability as signaling classification prior probability; determining the probability of occurrence of grids of each auxiliary classification category under different signaling classification categories respectively, and marking the probability as auxiliary classification likelihood; obtaining corresponding population posterior probability according to the auxiliary classification likelihood and the corresponding signaling classification prior probability; obtaining expected population of grids corresponding to the auxiliary classification category according to the population posterior probability and population average value of each signaling classification category; determining the distribution population probability of the corresponding grid according to the proportion of the expected population of the grid to the expected population of the grid of the street where the corresponding grid is located; and obtaining the predicted population of the corresponding grid according to the population probability and the population of the street where the corresponding grid is located.
Step S5: and distributing the predicted population number of the grid of the residential building to be tested according to the proportion of the residential calculation area of the residential building to be tested and the residential calculation area of the grid of the residential building to be tested, so as to obtain the predicted population number of the residential building to be tested.
Step S6: setting a part of the selected areas as verification areas, and acquiring the real population of the verification areas; recording the absolute value of the difference between the predicted population of the verification area and the real population as a prediction error; when the prediction error of the current prediction exceeds a preset threshold, marking the posterior population probability of the current prediction as the signaling classification prior probability of the next prediction, and predicting to obtain the predicted population number of the next prediction; and when the prediction error of the current prediction does not exceed the preset threshold value, the verification is passed, and the prediction is terminated.
For example, referring to FIG. 1, in one particular embodiment, the steps are described as:
(1) the existing interest surface data is utilized to spatially connect building outlines, and the buildings in the interest surface are endowed with category attributes of the interest surface, for example: the interest surface of a cell, all buildings inside are considered to be ordinary houses. Where spatial connection is a term commonly used in geographic information systems (Geographic Information System, abbreviated GIS), refers to connecting an attribute of one element class to an attribute of another element class according to a spatial relationship.
(2) Not all buildings are located inside the interest plane, and fuzzy matching is performed by using interest points for buildings for which analog attributes cannot be obtained through the step (1). Counting the types and kinds of interest points falling into the building outline, giving different weights to different types of interest points according to experience, generating a 50-meter buffer area by using the building, counting the types of interest points falling into the buffer area (according to the first principle of geographic big data, anything is related to other things, but similar things are more closely related), the weight of the interest points of the same type in the building outline is larger than that of the interest points in the buffer area, calculating the weighted sum (the number of the interest point types is equal to the weight), and giving the interest point type with the largest weighted sum to the building.
(3) The vector grid (250 x 250 m) according to the signaling data is connected by space, and the weighted sum of the building areas of various residential buildings in the grid is counted according to the grid (different types of buildings are given different weights, such as villas and dormitories, the residence area of the residential land of the villa type is far higher than the residence area of the dormitory type, the residential land of the higher residence area is given a smaller weight, otherwise, the residential land of the relatively crowded residence type is given a larger weight, and the actual building area of the building is multiplied by the weight to be used as the calculated area of the building, so as to weaken the interference between the residential building area and the population relation.
(4) Correcting the residence calculation area attribute of the grid obtained in the last step by using the residence type interest point kernel density, and obtaining (normalized residence calculation area multiplied by residence type interest point kernel density) 1/2 As auxiliary data for grid population distribution, denoted by B. In the embodiment shown in fig. 1, the density of the residential interest point kernels of a cell is obtained according to the interest Point (POI) data of a cell.
(5) Classifying the mobile phone signaling data (A) and the auxiliary data (B) corresponding to all grids according to a certain rule, wherein the mobile phone signaling data (A) and the auxiliary data (B) can be divided into a1, a2, a3, … and an; b1 B2, b3, …, bm; wherein n and m are positive integers, and form corresponding relations according to the grids, such as: a1-b2, a2-bm, an-b3 … …, allow for duplicate, one-to-many and many-to-one correspondence to occur in correspondence. Taking bayesian models as an example, a includes three signaling classifications a1, a2, a3, a1=1, a2=2, a3=3, b includes b1 and b2, b1=1, b2=2, 18 units of street population total, { a: b } represents the correspondence, each set of correspondences may form a set {1:1, 3:2, 2:2, 2:1, 3:2, 2:2, 1:2}.
The average value of each class in A is calculated simultaneously, and is expressed by C (C1, C2, C3, …, cn), wherein C corresponds to A one by one. A represents the result data, here represented by signaling data. an, bn represent respectively the nth group of data in the grouping of the attribute and result data according to a certain rule. C is a set of averages for each group of A, where cn represents the average of an group. For example, c1=1, c2=2, c3=3.
(6) And carrying out simple mathematical statistics on the classified mobile phone signaling data (A), and calculating the probability P (A) of occurrence corresponding to each class an as the signaling classification prior probability of population refinement space allocation. For example, P (a 1) is the probability of occurrence of a grid of the signaling class a1, P (a 1) =1/3, P (a 2) =1/3, P (a 3) =1/3.
(7) The probability P (b|a) of the auxiliary class B occurring at the occurrence of the signaling class a, auxiliary class likelihood, is determined from the grids, respectively, while the normalization constant P (B) is obtained from the formula P (B) =p (b|a) ×p (a). For example, a includes three signaling classifications a1, a2, and a3, and when the signaling classification a occurs, the probability P (b1|a) of the occurrence of the auxiliary classification b1 is expressed as the likelihood of the auxiliary classification b1, and in the above example, the likelihood of b1 or b2 under a1, a2, and a3 is:
P(1|1)＝2/3；P(2|1)＝1/3；
P(1|2)＝1/3；P(2|2)＝2/3；
P(1|3)＝0/3；P(2|3)＝3/3；
(8) and (3) calculating the signaling classification prior probability P (A), the auxiliary classification likelihood P (B|A) and the normalization constant P (B) obtained in the steps (6) and (7) according to the formula P (A|B) =P (A) ×P (B|A)/P (B), and obtaining the P (A|B) as population posterior probability, namely the probability of A occurrence when B occurs. In the above example, the normalization constants of b1 and b2 are respectively:
P(b1)＝P(A)*P(b1|A)＝P(a1)*P(b1|a1)+P(a2)*P(b1|a2)+P(a3)* P(b1|a3)＝1/3*(2/3+1/3+0)＝1/3；
P(b2)＝P(A)*P(b2|A)＝P(a1)*P(b2|a1)+P(a2)*P(b2|a2)+P(a3)* P(b2|a3)＝1/3*(2/3+1/3+1)＝2/3。
the corresponding population posterior probabilities are shown in table 1 below:
TABLE 1 posterior probability results
(9) From the posterior population probabilities obtained in the previous step, the population expectations under the attribute B are calculated as grid expectations population having the attribute B, respectively, by finding the expectation pattern e=c×p (a|b).
In the above-described example of the present invention,
E b1 ＝c1*P(a1|b1)+c2*P(a2|b1)+c3*P(a3|b1)＝2*1/3+2*1/3+3*0＝4/3；
E b2 ＝c1*P(a1|b2)+c2*P(a2|b2)+c3*P(a3|b2)＝1*1/6+2*2/6+3*3/6＝7/3。
according to the street where the grid is located, respectively summarizing and summing the expected population of the grid, and according to the formula: grid allocation population probability = expected population per grid/expected population of grids on the street on which the grid is located, resulting in a probability size for each grid to be allocated to the population.
In the above example, normalization is performed for each signaling class in a,
the population probability of each grid allocation is multiplied by the population data of the corresponding street obtained by the seventh census of the corresponding street population to obtain the population number to which each grid is reassigned. For example, the sum of expected demographics of the grids on the street where the grid is located is 18 units, the grid allocation demographics of the attribute b1 grid is multiplied by 18 to obtain the allocation demographics of the attribute b1 grid, and the grid allocation demographics of the attribute b2 grid is multiplied by 18 to obtain the allocation demographics of the attribute b2 grid.
And (3) carrying out population redistribution on the building level according to the size of the residence calculation area of all the buildings in the grid obtained in the step (3) as the partition weight. Specifically, the predicted population of the building is obtained by multiplying the predicted population of the grid of the residential building to be measured by the proportion of the calculated residential area of the building to the calculated residential area of the grid of the building.
Based on trueThe real building resident investigation data is checked with the population distribution data obtained in the last step, and if the absolute real population-estimated population is less than or equal to a threshold (set by people), the process is finished; otherwise, returning to the step (6), and repeating the steps (7) to (I) by taking the population posterior probability as the signaling classification prior probability >All steps until the threshold of true population-estimated population is met.
In a specific embodiment, the method of the present invention is used for predicting the western-style flag population, fig. 2 is a schematic diagram of regression analysis of population prediction using signaling sample expansion data in the existing method, the horizontal axis in fig. 2 represents the real population number, the vertical axis represents the population number of population prediction using signaling data, fig. 3 is a schematic diagram of regression analysis of population prediction using building area estimation in the conventional method, the horizontal axis in fig. 3 represents the real population number, and the vertical axis represents the population number estimated by building area. FIG. 4 is a schematic diagram of regression analysis of a Bayesian model on a predicted population of a selected region according to an embodiment of the present invention, wherein the horizontal axis in FIG. 4 represents the actual population number, and the vertical axis represents the population number estimated according to the Bayesian model; comparing fig. 2, fig. 3 and fig. 4, it can be seen that, the degree of fit of the population results of the 114 grids of the west three flags with the real population is closest to 1, which is the bayesian estimation in the embodiment of the present invention, the degree of fit of the straight line in fig. 3 is 0.7839, the degree of fit of the straight line in fig. 3 is 0.2161, the degree of fit of the straight line in fig. 4 is 1.2027, the degree of fit of the straight line in fig. 4 is 0.2027, and the average errors (calculation methods: error= |predicted value-true value|/true value) of the population of the grids are 16.39, 2.56 and 0.91 respectively, which can be seen obviously that the bayesian estimation provided in the embodiment of the present invention is superior to the conventional estimation method on the grid level.
Correspondingly, compared with the technical effects of the prior method in the embodiment of the invention, fig. 5 is a schematic diagram of the predicted population of each residential building in the selected area (the southern area of the western-style street) in the embodiment of the invention, and the predicted population numbers are different according to the outline of the building filled in the different manner in fig. 5; FIG. 6 is a schematic diagram of a predicted population error for each residential building for a selected area street according to an embodiment of the present invention, where the magnitude of the population prediction error corresponds to the type of building population as illustrated in the diagram of FIG. 6; FIG. 7 is a schematic diagram of estimating population errors of street buildings in a selected area according to building areas in the prior art, wherein the population prediction errors correspond to the types of filling of the buildings as shown in the legend in FIG. 7; FIG. 8 is a schematic diagram of the error change of the population predicted by building area according to the embodiment of the present invention, where the error change is equal to the absolute value of the Bayesian estimated building population error according to the embodiment of the present invention minus the absolute value of the estimated building population error according to the building area, the smaller the error change is smaller than 0, which indicates that the smaller the building population error estimated by the embodiment of the present invention, the better the algorithm is, whereas the worse the estimation effect is, and the 89.75% of the building body error change in FIG. 8 is not greater than 0 through statistics, which indicates that the error of the Bayesian estimated population at the building level is significantly reduced compared with the error of the estimated population distributed by building area only.
As can be seen from fig. 9 and 10, the fitting degree between the population results of all the building bodies of the west three flags and the real population numbers obtained by prediction according to the building area and the bayesian method is closest to 1, namely bayesian estimation, (bayes: fitting straight line 0.6508, building area: 0.2236), the prediction accuracy of the two methods at the building level is obviously higher, and the model accuracy is obviously improved. The population is estimated according to the building area, so that only one-time distribution of the building area is considered, and when checking is performed on a larger space scale (such as grids, cells and communities, a plurality of researches are checked on a larger scale, because data acquisition is simple and convenient, the grid in the embodiment is used for estimating the population of all buildings in the corresponding grid according to the building through a centralized meter, namely, the population sum of all buildings in the grid), the possible error is smaller, but the error is larger at a fine scale (the building). Also illustrated, the present embodiment improves the accuracy of building level population estimation by employing two levels (street-grid, grid-building) to control population distribution, with constraints on the grid level population overall.
Fig. 11 is a histogram corresponding to fig. 6, fig. 12 is a histogram corresponding to fig. 7, and fig. 13 is a histogram corresponding to fig. 8. As can be seen from fig. 11 and 12, the errors of the method of the present embodiment concentrate toward [ -40,5] the human mouth error value of the building body, compared to the human mouth error value of the human mouth error. As can be seen from fig. 13, 89.75% of the building body errors vary by no more than 0, and the errors of the bayesian estimated population at the building level are significantly reduced compared to the errors of the estimated population only by building area allocation. The error change is equal to the absolute value of the Bayesian estimated building population error minus the absolute value of the building population error estimated according to the building area, the smaller the error change is, the smaller the building population error estimated by Bayesian is, and the better the algorithm is; otherwise, the estimation effect is poor.
In the method of the embodiment, the population distribution is controlled through two layers (street-grid, grid-building), so that the accuracy of population estimation is improved; according to the characteristics of different types of buildings, one or more residence attributes of the grids are determined, the actual situation can be more attached, and population distribution auxiliary data is determined based on residence calculation areas and interest point kernel densities of the residence attributes, so that the prediction accuracy is improved; the population prediction of the grid level is realized, and the method can be widely applied to population prediction of large, medium and small areas; meanwhile, the mobile phone signaling data at the grid level is used as the signaling classification prior probability, so that the subjective judgment of introducing dislocation is avoided, and the actual situation is more fitted; further, through verification, repeated iteration is carried out for optimizing, parameter setting which is more in line with a selected area is found, a more accurate population prediction effect is achieved, a top-down population redistribution model is adopted, so that absolute accuracy of population at an upper layer (street) is guaranteed, and accuracy verification is carried out by taking down data at a lower layer (community or building), so that accuracy of population space distribution at a fine scale is guaranteed; furthermore, the anti-interference performance of the result can be improved by replacing the open-source big data with a single data source, the population prediction area range can be enlarged through sample expansion processing, abnormal value interference is reduced, and the processed data volume is reduced; based on the weighted sum of the interest point types, the buildings outside the interest surface can be reasonably classified, and the prediction accuracy is improved; the natural breakpoint method is used, so that a better estimation result can be realized in the embodiment of the invention than other classification methods; the number of the signaling classification categories and the number of the auxiliary classification categories are in a better range, so that classification homogenization is avoided, and a more accurate prediction effect can be realized; the prediction error can be effectively reduced, and the prediction precision is improved.
Those skilled in the art can change the above sequence without departing from the scope of the invention.
Example two
The second embodiment of the invention provides a two-step Bayesian estimation system for building scale population fusing position data, which comprises the following steps:
the data processing module is used for dividing the selected area according to the signaling data of the selected area; determining the type of each building in each grid according to the interest surface data of the selected area; dividing residential buildings in the grids according to different residential attributes, and determining the residential calculation area corresponding to each residential building in the corresponding grid and the residential calculation area of the corresponding grid; obtaining population distribution auxiliary data of the corresponding grids according to the living calculation area of the grids and the nuclear density of interest points of living buildings; classifying the grids according to population numbers of signaling data of each grid to obtain signaling classification categories of each grid; classifying the grids according to the population distribution auxiliary data of each grid to obtain auxiliary classification categories of each grid;
the grid population prediction module is used for respectively determining the probability of grid occurrence of each signaling classification category and recording the probability as signaling classification prior probability; determining the probability of occurrence of grids of each auxiliary classification category under different signaling classification categories respectively, and marking the probability as auxiliary classification likelihood; obtaining corresponding population posterior probability according to the auxiliary classification likelihood and the corresponding signaling classification prior probability; obtaining expected population of grids corresponding to the auxiliary classification category according to the population posterior probability and population average value of each signaling classification category; determining the distribution population probability of the corresponding grid according to the proportion of the expected population of the grid to the expected population of the grid of the street where the corresponding grid is located; and obtaining the predicted population of the corresponding grid according to the population probability and the population of the street where the corresponding grid is located.
In some specific embodiments, the number of signaling classification categories is not less than 80 and not greater than 120, and the number of auxiliary classification categories is not less than 250 and not greater than 350.
In some specific embodiments, a two-step bayesian estimation system for building scale population fusing location data further comprises:
and the building population prediction module is used for distributing the predicted population of the grid where the residential building to be detected is positioned according to the ratio of the residential calculation area of the residential building to be detected to the residential calculation area of the grid where the residential building to be detected is positioned, so as to obtain the predicted population of the residential building to be detected.
In some specific embodiments, a two-step bayesian estimation system for building scale population fusing location data further comprises:
the prediction population verification module is used for setting a part of the selected area as a verification area and acquiring the real population number of the verification area; recording the absolute value of the difference between the predicted population of the verification area and the real population as a prediction error; when the prediction error of the current prediction exceeds a preset threshold, marking the posterior population probability of the current prediction as the signaling classification prior probability of the next prediction, and predicting to obtain the predicted population number of the next prediction; and when the prediction error of the current prediction does not exceed the preset threshold value, the verification is passed, and the prediction is terminated.
In the embodiment, the population distribution is controlled through two layers (street-grid, grid-building), so that the accuracy of population estimation is improved; according to the characteristics of different types of buildings, one or more residence attributes of the grids are determined, the actual situation can be more attached, and population distribution auxiliary data is determined based on residence calculation areas and interest point kernel densities of the residence attributes, so that the prediction accuracy is improved; the population prediction of the grid level is realized, and the method can be widely applied to population prediction of large, medium and small areas; meanwhile, the mobile phone signaling data at the grid level is used as the signaling classification prior probability, so that the subjective judgment of introducing dislocation is avoided, and the actual situation is more fitted; further, through verification, repeated iteration is carried out for optimizing, parameter setting which is more in line with a selected area is found, a more accurate population prediction effect is achieved, a top-down population redistribution model is adopted, so that absolute accuracy of population at an upper layer (street) is guaranteed, and accuracy verification is carried out by taking down data at a lower layer (community or building), so that accuracy of population space distribution at a fine scale is guaranteed; furthermore, the anti-interference performance of the result can be improved by replacing the open-source big data with a single data source, the population prediction area range can be enlarged through sample expansion processing, abnormal value interference is reduced, and the processed data volume is reduced; based on the weighted sum of the interest point types, the buildings outside the interest surface can be reasonably classified, and the prediction accuracy is improved; the natural breakpoint method is used, so that a better estimation result can be realized in the embodiment of the invention than other classification methods; the number of the signaling classification categories and the number of the auxiliary classification categories are in a better range, so that classification homogenization is avoided, and a more accurate prediction effect can be realized; the prediction error can be effectively reduced, and the prediction precision is improved.
Based on the same inventive concept, an embodiment of the present invention further provides an electronic device, whose structure is shown in fig. 14, including: the system comprises a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor realizes the two-step Bayesian estimation method of the building dimension population fusing position data when executing the computer program.
Based on the same inventive concept, the embodiment of the invention also provides a computer storage medium, wherein the computer storage medium stores computer executable instructions, and the computer executable instructions realize the two-step Bayesian estimation method of the building scale population fusing the position data when being executed by a processor.
With respect to the system, the electronic device, and the computer storage medium in the above embodiments, the detailed description thereof has been described in detail in the embodiments related to the method, and will not be explained in detail here.
Any modification, supplement, equivalent replacement, etc. within the principle of the technical solution of the present invention should still fall within the scope of patent coverage of the technical solution of the present invention.
Claims (12)
1. The two-step Bayesian estimation method for the building scale population fusing the position data is characterized by comprising the following steps of:
Dividing the selected area according to grids according to signaling data of the selected area;
determining the type of each building in each grid according to the interest surface data of the selected area; when a building positioned outside the interest surface exists in the grid, a buffer zone with a preset width is arranged outside the outline of the building, and the number and types of all interest points in the outline of the building and in the buffer zone are counted; respectively giving corresponding weights to each interest point type, wherein the weights of the interest points in the building outline are larger than those of the interest points of the same type in the buffer zone; multiplying the number of each interest point by corresponding weight to obtain weighted sums of each interest point type in the building outline and the buffer zone respectively; the type of the interest point with the largest weighted sum is recorded as the type of the building;
dividing residential buildings in the grid according to different residential attributes, and respectively giving corresponding weights to each residential attribute; multiplying the actual building area of each residential building in the grid by the weight of the corresponding residential attribute to obtain the residential calculation area corresponding to each residential building in the grid; adding the residence calculation areas corresponding to each residence building in the grids to obtain residence calculation areas corresponding to the grids;
Multiplying the living calculation area of the grid by the square root of the product obtained by the nuclear density of the points of interest of the living building of the grid, and recording the square root as the auxiliary data of population distribution corresponding to the living attribute of the grid;
classifying the grids according to population numbers of signaling data of each grid to obtain signaling classification categories of each grid;
classifying the grids according to the population distribution auxiliary data of each grid to obtain auxiliary classification categories of each grid;
determining the probability of occurrence of grids of each signaling classification category respectively, and marking the probability as signaling classification prior probability;
determining the probability of occurrence of grids of each auxiliary classification category under different signaling classification categories respectively, and marking the probability as auxiliary classification likelihood;
obtaining corresponding population posterior probability according to the auxiliary classification likelihood and the corresponding signaling classification prior probability;
obtaining expected population of grids corresponding to the auxiliary classification category according to the population posterior probability and population average value of each signaling classification category;
determining the distribution population probability of the corresponding grid according to the ratio of the expected population of the grid to the sum of expected population of the grids of all the grids of the street where the corresponding grid is located;
Obtaining a corresponding grid prediction population according to the distribution population probability and the population of the street where the corresponding grid is located;
setting a part of the selected areas as verification areas, and acquiring the real population of the verification areas; recording the absolute value of the difference between the predicted population of the verification area and the real population as a prediction error; when the prediction error of the current prediction exceeds a preset threshold, marking the posterior population probability of the current prediction as the signaling classification prior probability of the next prediction, and predicting to obtain the predicted population number of the next prediction; and when the prediction error of the current prediction does not exceed the preset threshold value, the verification is passed, and the prediction is terminated.
2. The method of claim 1, wherein prior to classifying said grids based on the population of signaling data for each said grid, comprising the steps of:
comparing the acquired signaling data sample size with population of a selected area, and determining sample expansion parameters;
and according to the sample spreading parameters, the obtained signaling data sample size is subjected to sample spreading to obtain the population of the signaling data of each grid in the selected area.
3. The method of claim 1, wherein the classification method of classifying the grids according to population of signaling data of each of the grids or according to population of the population allocation assistance data of each of the grids is a natural break point method.
4. The method of claim 1, wherein the number of signaling classification categories is not less than 80 and not greater than 120.
5. The method of claim 1, wherein the number of auxiliary classification categories is not less than 250 and not greater than 350.
6. The method of claim 1, further comprising the step of:
and distributing the predicted population number of the grid of the residential building to be tested according to the proportion of the residential calculation area of the residential building to be tested and the residential calculation area of the grid of the residential building to be tested, so as to obtain the predicted population number of the residential building to be tested.
7. A two-step bayesian estimation system for building scale population fusing location data, comprising:
the data processing module is used for dividing the selected area according to the signaling data of the selected area; determining the type of each building in each grid according to the interest surface data of the selected area; when a building positioned outside the interest surface exists in the grid, a buffer zone with a preset width is arranged outside the outline of the building, and the number and types of all interest points in the outline of the building and in the buffer zone are counted; respectively giving corresponding weights to each interest point type, wherein the weights of the interest points in the building outline are larger than those of the interest points of the same type in the buffer zone; multiplying the number of each interest point by corresponding weight to obtain weighted sums of each interest point type in the building outline and the buffer zone respectively; the type of the interest point with the largest weighted sum is recorded as the type of the building; dividing residential buildings in the grid according to different residential attributes, and respectively giving corresponding weights to each residential attribute; multiplying the actual building area of each residential building in the grid by the weight of the corresponding residential attribute to obtain the residential calculation area corresponding to each residential building in the grid; adding the residence calculation areas corresponding to each residence building in the grids to obtain residence calculation areas corresponding to the grids; multiplying the living calculation area of the grid by the square root of the product obtained by the nuclear density of the points of interest of the living building of the grid, and recording the square root as the auxiliary data of population distribution corresponding to the living attribute of the grid; classifying the grids according to population numbers of signaling data of each grid to obtain signaling classification categories of each grid; classifying the grids according to the population distribution auxiliary data of each grid to obtain auxiliary classification categories of each grid;
The grid population prediction module is used for respectively determining the probability of grid occurrence of each signaling classification category and recording the probability as signaling classification prior probability; determining the probability of occurrence of grids of each auxiliary classification category under different signaling classification categories respectively, and marking the probability as auxiliary classification likelihood; obtaining corresponding population posterior probability according to the auxiliary classification likelihood and the corresponding signaling classification prior probability; obtaining expected population of grids corresponding to the auxiliary classification category according to the population posterior probability and population average value of each signaling classification category; determining the distribution population probability of the corresponding grid according to the ratio of the expected population of the grid to the sum of expected population of the grids of all the grids of the street where the corresponding grid is located; obtaining a corresponding grid prediction population according to the distribution population probability and the population of the street where the corresponding grid is located;
the prediction population verification module is used for setting a part of the selected area as a verification area and acquiring the real population number of the verification area; recording the absolute value of the difference between the predicted population of the verification area and the real population as a prediction error; when the prediction error of the current prediction exceeds a preset threshold, marking the posterior population probability of the current prediction as the signaling classification prior probability of the next prediction, and predicting to obtain the predicted population number of the next prediction; and when the prediction error of the current prediction does not exceed the preset threshold value, the verification is passed, and the prediction is terminated.
8. The system of claim 7, wherein the number of signaling classification categories is not less than 80 and not more than 120.
9. The system of claim 7, wherein the number of auxiliary classification categories is not less than 250 and not greater than 350.
10. The system as recited in claim 7, further comprising:
and the building population prediction module is used for distributing the predicted population of the grid where the residential building to be detected is positioned according to the ratio of the residential calculation area of the residential building to be detected to the residential calculation area of the grid where the residential building to be detected is positioned, so as to obtain the predicted population of the residential building to be detected.
11. An electronic device, comprising: a memory, a processor and a computer program stored on the memory and running on the processor, the processor implementing the two-step bayesian estimation method of building scale population incorporating location data according to any one of claims 1-6 when the computer program is executed.
12. A computer storage medium having stored therein computer executable instructions that when executed implement the two-step bayesian estimation method of building scale population incorporating location data according to any of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211262928.8A CN115525872B (en) | 2022-10-13 | 2022-10-13 | Two-step Bayesian estimation method for building scale population fused with position data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211262928.8A CN115525872B (en) | 2022-10-13 | 2022-10-13 | Two-step Bayesian estimation method for building scale population fused with position data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115525872A CN115525872A (en) | 2022-12-27 |
| CN115525872B true CN115525872B (en) | 2023-09-15 |
Family
ID=84700866
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211262928.8A Active CN115525872B (en) | 2022-10-13 | 2022-10-13 | Two-step Bayesian estimation method for building scale population fused with position data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115525872B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018082784A1 (en) * | 2016-11-07 | 2018-05-11 | Swiss Reinsurance Company Ltd. | System and method for predicting of absolute and relative risks for car accidents |
| CN108871286A (en) * | 2018-04-25 | 2018-11-23 | 中国科学院遥感与数字地球研究所 | The completed region of the city density of population evaluation method and system of space big data collaboration |
| CN110868689A (en) * | 2019-11-25 | 2020-03-06 | 智慧足迹数据科技有限公司 | Standing population determining method and device and electronic equipment |
| CN111212383A (en) * | 2018-11-01 | 2020-05-29 | 百度在线网络技术（北京）有限公司 | Method, device, server and medium for determining number of regional permanent population |
| CN111932036A (en) * | 2020-09-23 | 2020-11-13 | 中国科学院地理科学与资源研究所 | Fine spatio-temporal scale dynamic population prediction method and system based on position big data |
| CN114282934A (en) * | 2021-03-30 | 2022-04-05 | 华南理工大学 | Urban low-income crowd distribution prediction method and system based on mobile phone signaling data and storage medium |
| CN115049159A (en) * | 2022-08-12 | 2022-09-13 | 北京大学 | Population distribution prediction method and device, storage medium and electronic equipment |
-
2022
- 2022-10-13 CN CN202211262928.8A patent/CN115525872B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018082784A1 (en) * | 2016-11-07 | 2018-05-11 | Swiss Reinsurance Company Ltd. | System and method for predicting of absolute and relative risks for car accidents |
| CN108871286A (en) * | 2018-04-25 | 2018-11-23 | 中国科学院遥感与数字地球研究所 | The completed region of the city density of population evaluation method and system of space big data collaboration |
| CN111212383A (en) * | 2018-11-01 | 2020-05-29 | 百度在线网络技术（北京）有限公司 | Method, device, server and medium for determining number of regional permanent population |
| CN110868689A (en) * | 2019-11-25 | 2020-03-06 | 智慧足迹数据科技有限公司 | Standing population determining method and device and electronic equipment |
| CN111932036A (en) * | 2020-09-23 | 2020-11-13 | 中国科学院地理科学与资源研究所 | Fine spatio-temporal scale dynamic population prediction method and system based on position big data |
| CN114282934A (en) * | 2021-03-30 | 2022-04-05 | 华南理工大学 | Urban low-income crowd distribution prediction method and system based on mobile phone signaling data and storage medium |
| CN115049159A (en) * | 2022-08-12 | 2022-09-13 | 北京大学 | Population distribution prediction method and device, storage medium and electronic equipment |
Non-Patent Citations (1)
| Title |
|---|
| 基于手机信令数据的居住地人口分布辨识改进方法;黄伟 等;《城市交通》;第19卷(第1期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115525872A (en) | 2022-12-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| AU2018101946A4 (en) | Geographical multivariate flow data spatio-temporal autocorrelation analysis method based on cellular automaton | |
| CN103020978B (en) | SAR (synthetic aperture radar) image change detection method combining multi-threshold segmentation with fuzzy clustering | |
| CN110648014B (en) | Regional wind power prediction method and system based on space-time quantile regression | |
| Kang et al. | Making the most of spatial information in health: a tutorial in Bayesian disease mapping for areal data | |
| CN106485269B (en) | SAR image object detection method based on mixing statistical distribution and multi-part model | |
| Saffari et al. | A methodology for identifying critical links and estimating macroscopic fundamental diagram in large-scale urban networks | |
| CN109613182A (en) | Monitoring location site selecting method and device based on atmosphere pollution | |
| US20220067336A1 (en) | System and process for integrative computational soil mapping | |
| CN110956412A (en) | Flood dynamic assessment method, device, medium and equipment based on real-scene model | |
| CN106250933A (en) | Method, system and the FPGA processor of data clusters based on FPGA | |
| Mariani et al. | The arcview and GeoDa application in optimization of spatial regression estimate | |
| CN110174106A (en) | A kind of healthy walking paths planning method and terminal device based on PM2.5 | |
| US20230131579A1 (en) | Wireless signal propagation prediction method and apparatus | |
| CN110689055B (en) | Cross-scale statistical index spatialization method considering grid unit attribute grading | |
| CN116187835A (en) | Data-driven-based method and system for estimating theoretical line loss interval of transformer area | |
| CN111639882A (en) | Deep learning-based power utilization risk judgment method | |
| CN115600729A (en) | Grid load prediction method considering multiple attributes | |
| CN114205831B (en) | Method, device, storage medium and equipment for determining optimization scheme | |
| CN115525872B (en) | Two-step Bayesian estimation method for building scale population fused with position data | |
| CN109583741A (en) | Driving trace methods of risk assessment and system | |
| CN116258279B (en) | Landslide vulnerability evaluation method and device based on comprehensive weighting | |
| CN114880954A (en) | Landslide sensitivity evaluation method based on machine learning | |
| CN115859765A (en) | Method, device, equipment and storage medium for predicting city expansion | |
| TWM624436U (en) | Housing price appraisal equipment | |
| CN114118512A (en) | Grouping public medical facility configuration method based on dynamic population data model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information |
Address after: Room 517, 5th Floor, Building 2, No. 5, Fuchengmen North Street, Xicheng District, Beijing 100037 Applicant after: Beijing Urban Construction Transportation Design and Research Institute Co.,Ltd. Address before: Room 517, 5th Floor, Building 2, No. 5, Fuchengmen North Street, Xicheng District, Beijing 100037 Applicant before: Beijing Urban Construction Design and Research Institute Co.,Ltd. |
|
| CB02 | Change of applicant information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |