A method and system for mining spatial correlations across different types of geographic objects in general

Abstract

The application discloses a kind of cross overall different type geographical object space correlation mining method and system, the method includes obtaining the city road data in research area and the POI data of first geographical object total geographical object total and second geographical object total and is preprocessed;The POI data of first geographical object total and second geographical object total is divided into different category subsets according to priori knowledge;Zero hypothesis is set and the statistical index of binary spatial correlation is defined;From the global and local measurement level, the observation count of different category subsets in another geographical object total is carried out in spatial neighborhood, and the statistical index of binary spatial correlation is calculated;Significance test, correlation analysis and visual display are carried out.Compared with the traditional binary spatial correlation mining method, the application is suitable for mining the binary spatial correlation between the cross overall, large-scale data set, and can reveal the potential contact and dependent relationship between any two different geographical object totals.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of binary spatial association mining, and particularly relates to a method and system for mining spatial association between different types of geographical objects across populations. BACKGROUND

[0002] Binary spatial association mining is of great significance for understanding the spatial dependence between different types of geographical objects. For example, in urban planning, analyzing the spatial association between fires and land use facilities helps to assess the fire risk around urban facilities, so as to formulate more effective fire prevention policies. Although binary spatial association mining methods have attracted extensive attention and application, however, most of the existing methods are not suitable for mining binary spatial association between different types of geographical objects across populations (for example, the association between facility types and crime types, the association between fire types and land use facility types). The existing binary spatial association mining methods still face two major problems when mining spatial relationships between different types of geographical objects across populations.

[0003] Firstly, the zero hypothesis selection is not appropriate. For the binary spatial association mining task between different types of geographical objects across populations, the existing mining methods may overestimate the binary spatial association results due to the inappropriate zero hypothesis. Secondly, although the existing association indicators can indicate whether there is an association between two types of geographical objects, they cannot accurately express the strength of the binary spatial association (compared to random distribution). This poses a great challenge to the existing methods for mining association between different types of geographical objects across populations. SUMMARY

[0004] The main purpose of the present application is to overcome the shortcomings and deficiencies of the prior art, and to provide a method and system for mining spatial association between different types of geographical objects across populations.

[0005] In order to achieve the above purpose, the technical scheme adopted by the present application is as follows:

[0006] In one aspect of the present application, a method for mining spatial association between different types of geographical objects across populations is provided, comprising the following steps:

[0007] Obtaining urban road data in a study area, and POI data of a first geographical object population and a second geographical object population; P Performing debugging and road network matching on the obtained data;Calculating the road network distance between the POI data and constructing a road network model;The first geographical object population and the second geographical object population are different geographical object populations to be mined for spatial association. Q P Q

[0008] The first geographical object population is divided into a plurality of geographical object sub-populations, and the second geographical object population is divided into a plurality of geographical object sub-populations;​P Second Geographic Object Overall Q The POI data is divided into different category subsets based on prior knowledge; the overall first geographic object is set as the first geographic object. P With the second geographic object as a whole Q The populations are independent of each other, and the spatial distribution within each geographic object population follows the null hypothesis of random labeling; the statistical index of binary spatial association is defined as the ratio of the actual observed value to the expected value under the null hypothesis; different spatial proximity definitions are used to define the populations of the first geographic objects. P Second Geographic Object Overall Q Construct spatial neighborhoods for different categories of subsets or POI data;

[0009] From the global and local measurement levels, we observe and count different categories of subsets in another geographic object population within the spatial neighborhood corresponding to the selected geographic object population, and calculate the statistical indicators of the binary spatial association between different categories of subsets in the selected geographic object population and different categories of subsets in another geographic object population.

[0010] The statistical indicators are tested for significance based on the observation and counting results to determine the significance of the binary spatial association mining results; the significant binary spatial association mining results are then subjected to association analysis and visualization.

[0011] As a preferred technical solution, the overall first geographic object P Second Geographic Object Overall Q The POI data is divided into different category subsets based on prior knowledge, specifically:

[0012] The first geographic object as a whole P Second Geographic Object Overall Q Classified according to different subtypes: P ={ P 1,..., P A ,..., P Z}, A ∈{1,2,…, Z}; Q ={ Q 1,..., Q a ,..., Q z}, a ∈{1,2,…, z};in, P A and Q a For different category subsets, Z andz is the total number of category subsets corresponding to the geographical object population, A and a is the index of the category subset; each category subset includes a number of POI data, denoted as: P A = P A-1 , P A-M , P A-O , Q a = Q a-1 , Q a-m , Q a-o , wherein M ∈{1, 2, …, O}, m ∈{1, 2, …, o}, O and o is the total number of POI data corresponding to the category subset, M and m is the POI data index corresponding to the category subset.

[0013] As a preferred technical solution, for the global metric hierarchy, the corresponding spatial neighborhood is determined according to all POI data of any one category subset in the first geographical object population P and the second geographical object population Q by a pre-set spatial proximity definition method, specifically:

[0014] P A → G A ;

[0015] Q a → G a ;

[0016] The observation count is performed on all category subsets of the other geographical object population in the obtained spatial neighborhood, specifically:

[0017] G A = Q a : N obv(Qa) | N obv(Qa) ∈{ C obv( Q 1), C obv ( Q 2),…, C obv ( Q z )}};

[0018] G a ={ P A : N obv(PA) | N obv(PA) ∈{ C obv ( P 1), C obv ( P 2), …, C obv ( P Z )}};

[0019] wherein, the preset spatial proximity definition mode comprises spatial distance, geographical weighted spatial distance, k near neighbor and geographical weighted k near neighbor; N obv(Qa) 、 N obv(PA) are respectively the results of observation counting on all category subsets of another geographical object population in spatial neighborhood G A 、 G a Q a 、 P A C obv is an observation counting operation.

[0020] As a preferred technical solution, at the global measurement level, the statistical indicators of binary spatial association between a category subset in the selected geographical object population and different category subsets in another geographical object population are calculated, specifically:

[0021] , G ={ G A , G a};

[0022] wherein, GSI is the spatial proximity definition mode​​k the statistical indicator of the binary spatial association between a certain subset of categories in the selected geographical object population and another subset of categories in the other geographical object population when G is taken G A the formula represents the statistical indicator of the observed counts in the spatial neighborhood G A of all subsets of categories in the second geographical object population Q Q a when N 1 represents the number of geographical objects in the subset of categories Q a 2 represents the number of geographical objects in the second geographical object population; when N is taken Q a the formula represents the statistical indicator of the observed counts in the spatial neighborhood G a of all subsets of categories in the first geographical object population G A when G a A P P A N A P A N P N G is the number of geographical objects in G i is the multiplication sign x i is a 0-1 variable that takes the value 1 when G the specific geographical object i corresponding subset of categories is P A or Q a

[0023] ;

[0024] where GWGSI is the spatial proximity defined as the geographical weighted k nearest neighbor or the geographical weighted spatial distance under the global metric hierarchy; w i is the geographical weight assigned to the specific geographical object i given by the kernel function.

[0025] ​​​​As a preferred technical solution, for the local metric level, a corresponding spatial neighborhood is determined according to a selected POI data of any one category subset of the first geographic object population and the second geographic object population by a preset spatial proximity definition manner, specifically as follows: P Q

[0026] P A-M G A-M

[0027] Q a-m G a-m

[0028]

[0029] G A-M Q a n obv(Qa) n obv(Qa) c obv Q 1), c obv Q 2),…, c obv Q z

[0030] G a-m P A n obv(PA) n obv(PA) c obv P 1), c obv P 2),…, c obv P Z

[0031] k k wherein the preset spatial proximity definition manner includes spatial distance, geographical weighted spatial distance,​​​​​​​​​​​​​​​​​​​​​​​​​n obv(Qa) , n obv(PA) In the spatial neighborhood G A-M , G a-m The results of observational counting of all category subsets of another geographic object population. c obv For observation counting operations.

[0032] As a preferred technical solution, at the local metric level, statistical indicators are calculated regarding the binary spatial association between specific POI data in a subset of a selected geographic object population and different subsets of another geographic object population. Specifically:

[0033] , G’ ={ G A-M , G a-m};

[0034] in, LSI The spatial proximity definition method under the local metric level is as follows k Nearest neighbor or spatial distance is a statistical indicator of the binary spatial association between a specific POI data set within a subset of a category and a subset of categories within another geographic object population; when G’ Pick G A-M The time formula is expressed in the spatial neighborhood. G A-M The overall second geographical object Q All category subsets Q a The statistical indicators obtained from observation and counting, at this time N 1 indicates a subset of categories Q a The number of geographic objects in the data. N 2 represents the totality of the second geographic object. Q The number of geographical objects in the middle; when G’ Pick G a-m The time formula is expressed in the spatial neighborhood. G a-m The overall situation of the first geographical object in China P All category subsets P A The statistical indicators obtained from observation and counting, at this time N 1 indicates a subset of categories P A The number of geographic objects in the data. N 2 represents the total number of the first geographic objects.P the number of geographic objects in N G’ is G’ the number of geographic objects in x j is a 0-1 variable, when G’ a specific geographic object in j the corresponding category subset is P A or Q a is 1;

[0035] ;

[0036] wherein, GWLSI is a statistical indicator of binary spatial association when the spatial proximity is defined as geographic weighted k nearest neighbors or geographic weighted spatial distance; w j is the geographic weight assigned to a specific geographic object j , given by a kernel function.

[0037] As a preferred technical solution, the statistical indicator is subjected to a significance test according to the observed count result, specifically:

[0038] the probability distribution of hypergeometric distribution X~H ( N 2 , N 1 , N ) is calculated: P [ C = x ] :

[0039] , N ={ N G , N G’} ;

[0040] wherein, N G and N G’ respectively represent the number of geographic objects in the spatial neighborhood G and G’ , when N takes N G The formula represents a significance test of the statistical indicator according to the observed count result under the global metric hierarchy, when N takes N G’The formula represents significance test of statistical indicators according to the observation count result under the local metric level; when it is necessary to perform significance test on the statistical indicators of binary spatial correlation between a certain category subset P A or specific POI data P A-M and a category subset in another geographical object population Q a , the significance test of the statistical indicators of binary spatial correlation between the two is performed, N 1 represents the number of geographical objects in the category subset Q a , and N 2 represents the number of geographical objects in the second geographical object population Q ; when it is necessary to perform significance test on the statistical indicators of binary spatial correlation between a certain category subset Q a or specific POI data Q a-m and a category subset in another geographical object population P A , the significance test of the statistical indicators of binary spatial correlation between the two is performed, N 1 represents the number of geographical objects in the category subset P A , and N 2 represents the number of geographical objects in the first geographical object population P . x represents the observation count result; the appearing brackets on the right side of the formula equal sign represent permutation combination calculation, that is, the possible number of times of taking k geographical objects from n geographical objects. I J

[0041] As a preferred technical solution, the correlation analysis on the significant binary spatial correlation mining result is specifically:

[0042] when the value of the corresponding statistical indicator is greater than the set threshold value and the significance test P [ C ≥ x ] is less than the significance threshold value, it is considered that the null hypothesis is rejected at the significance level of the significance threshold value, and there is positive correlation;

[0043] when the value of the corresponding statistical indicator is less than the set threshold value and the significance test P [ C ≤ x ] is less than the significance threshold value, it is considered that the null hypothesis is rejected at the significance level of the significance threshold value, and there is negative correlation;

[0044] wherein, P [ C ≥ x ] and​​P [ C ≤ x ] values are obtained according to the hypergeometric distribution X~H ( N 2 , N 1 , N ) probability distribution P [ C = x ].

[0045] As a preferred technical solution, urban road data in a research area is obtained through geographic information system software and open source data, and POI data of a first geographic object population P and a second geographic object population Q ; the urban road data is simplified, topologically checked, loop-removed and debugged, and the POI data of the first geographic object population P and the second geographic object population Q is debugged; the first geographic object population P is a city facility population, and the second geographic object population Q is a crime event population.

[0046] Another aspect of the present application also provides a system for mining spatial correlations between different types of geographic objects across populations, which is applied to the above-mentioned method for mining spatial correlations between different types of geographic objects across populations, and comprises a data acquisition and preprocessing module, a statistical index calculation module and an analysis module.

[0047] The data acquisition and preprocessing module obtains urban road data in a research area, and POI data of a first geographic object population P and a second geographic object population Q ; the obtained data is debugged and matched with a road network; road network distances between POI data are calculated, and a road network model is constructed; the first geographic object population P and the second geographic object population Q are different geographic object populations to be subjected to spatial correlation mining; the POI data of the first geographic object population P and the second geographic object population Q are divided into different category subsets according to prior knowledge;

[0048] The statistical index calculation module is used for setting that the first geographic object population P and the second geographic object population Q are independent of each other, and the spatial distribution within each geographic object population follows a zero hypothesis of random labeling; a statistical index of binary spatial correlation is defined as a ratio of an actual observation value to an expected value under the zero hypothesis; the first geographic object population PSecond Geographic Object Overall Q Construct spatial neighborhoods for different categories of subsets or POI data; from the global and local metric levels, observe and count different categories of subsets in another geographic object population within the spatial neighborhood corresponding to the selected geographic object population, and calculate statistical indicators of the binary spatial association between different categories of subsets or POI data in the selected geographic object population and different categories of subsets in another geographic object population.

[0049] The analysis module is used to perform significance tests on the statistical indicators based on the observation count results, determine the significance of the binary spatial association mining results, and perform association analysis and visualization on the significant binary spatial association mining results.

[0050] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0051] (1) Based on urban road data, urban facilities and crime incident POI data, this invention provides a spatial association mining method between different types of geographic objects by combining spatial analysis methods. This invention uses the above method to reveal the potential connections and dependencies between urban facility types and crime incident types. Compared with traditional binary spatial association mining methods, this invention can be applied to mining binary spatial associations between large-scale datasets across the whole, providing certain benefits to related industries or research such as urban planning and crime analysis.

[0052] (2) As a general and universal method, this invention is not limited to mining spatial association patterns between urban facilities and crime events. It can be applied to mining spatial association patterns between all urban road data and POI data from two different geographic object populations. This invention can reveal the potential connections and dependencies between any two different geographic object populations, providing new guidance and insights for existing binary spatial association mining methods. Attached Figure Description

[0053] Figure 1 This is a flowchart of a spatial association mining method for different types of geographic objects across a general population, as described in an embodiment of the present invention.

[0054] Figure 2 This is a structural diagram of a spatial association mining system for different types of geographic objects across a general population, as described in an embodiment of the present invention. Detailed Implementation

[0055] In the following, the technical solutions in the embodiments of the present application will be described clearly and completely in conjunction with the drawings in the embodiments of the present application, so that those skilled in the art can better understand the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0056] Embodiments

[0057] Reference is made to Figure 1 , Figure 1 is a flowchart of a method for mining spatial correlations between different types of geographic objects in the present embodiment. In the present embodiment, urban facilities and crime events are taken as different geographic object groups to be mined for spatial correlations. As a general and universal method, those skilled in the art can apply the method of the present application to mining spatial correlation patterns between all urban road data and POI data from two different geographic object groups (denoted as a first geographic object group and a second geographic object group, in addition to urban facilities-crime events, other different geographic object groups can be taken as research objects, such as fire-land use type, commercial service facility type-human activity event type, residential building type-office building type), without creative work. The embodiments of the present application are not limited to urban facilities and crime events, and should be included in the protection scope of the present application.

[0058] Specifically, the following steps are included:

[0059] S1: Obtain urban road data and point of interest (POI) data of urban facilities and crime events in a research area; perform pre-processing procedures such as debugging and road network matching on the POI data; construct a road network model and calculate road network distances between POI data.

[0060] Further, step S1 includes the following steps:

[0061] S11: Obtain urban road data and POI data of urban facilities and crime events in the research area through geographic information system professional software and open source databases; perform pre-processing procedures such as simplification, topology checking, loop removal and debugging on the urban road data to obtain pre-processed urban road data;

[0062] S12: Perform debugging processing on the POI data of urban facilities and crime events, and match the cleaned POI data to the pre-processed urban road to calculate and record road network distances between each two POI, and complete the construction of the road network model.

[0063] S2: According to prior knowledge, the urban facilities and crime event data contained in the POI data are divided into different category subsets; the null hypothesis and statistical index definition strategy are set; the respective spatial neighborhood of different types of urban facilities and crime events in different category subsets or POI data is constructed by selecting different spatial proximity definition methods.

[0064] Further, step S2 includes the following steps:

[0065] S21: According to prior knowledge, the urban facilities overall (data set, denoted as P ) and crime overall (data set, denoted as Q ) contained in the POI data are classified according to different subtypes, that is P ={ P 1,..., P A ,..., P Z}, A ∈{1,2,…, Z}; Q ={ Q 1,..., Q a ,..., Q z}, a ∈{1,2,…, z}; wherein, P A and Q a are different category subsets, Z and z are the total number of category subsets of the corresponding geographical object overall, A and a are the indexes of the category subsets; each category subset includes a number of data, denoted as: P A ={ P A-1 ,..., P A-M ,..., P A-O}, Q a ={ Q a-1 ,..., Q a-m ,..., Q a-o}, wherein, M ∈{1,2,…, O}, m ∈{1,2,…, o}O and o total number of POI data of the corresponding category subset, M and m POI data index of the corresponding category subset;

[0066] S22: Form a city facility overall P and a crime event overall Q independent of each other, but the spatial distribution within each geographical object overall follows the zero hypothesis of random marking, and is set to the actual observed value C obv ratio of the observed value to the expected value under the zero hypothesis C exp Definition strategy of the statistical index of binary spatial association. It is stipulated that the statistical index Statistical Indicator , SI Definition strategy mathematical expression is:

[0067] ;

[0068] S23: Set four different spatial proximity definition methods: spatial distance, geographical weighted spatial distance, k near neighbor and geographical weighted k near neighbor, and construct the respective spatial neighborhood for different types of city facility and / or crime event data by selecting different spatial proximity definition methods.

[0069] S3: From the global measurement level and the local measurement level, determine the spatial neighborhood for each category subset or POI data, and count the observations of different types of city facilities or crime event data in the spatial neighborhood; according to the pre-set spatial proximity definition method, calculate the statistical index of binary spatial association between different city facility types and different crime event types.

[0070] Further, step S3 includes the following steps:

[0071] S31: For the global measurement level, determine the corresponding spatial neighborhood according to all POI data of any one category subset in the city facility overall P or the crime event overall Q , specifically:

[0072] P A → G A ;

[0073] Q a → G a ;

[0074] counting the observation of all category subsets of another geographical object population in the obtained spatial neighborhood, in particular:

[0075] G A Q a N obv(Qa) N obv(Qa) C obv Q 1), C obv Q 2),…, C obv Q z

[0076] G a P A N obv(PA) N obv(PA) C obv P 1), C obv P 2),…, C obv P Z

[0077] wherein the pre-defined spatial neighborhood definition comprises a spatial distance, a geographically weighted spatial distance, k a k-neighborhood and a geographically weighted k k-neighborhood; N obv(Qa) N obv(PA) are the results of counting the observation of all category subsets of another population in the spatial neighborhood G A G a Q a P A C obv is the observation counting operation;

[0078] ​​​​​​​​​​​​​​​​​​​​​S32: At the global metric level, calculate the statistical indicators of the binary spatial association between a subset of a certain category in a selected geographic object population and different subsets of another geographic object population, specifically:

[0079] , G ={ G A , G a};

[0080] in, GSI The spatial proximity definition method under the global metric level is as follows k Neighborhood or spatial distance is a statistical indicator of the binary spatial association between a subset of a category and a subset of a category in another geographic object population; when G Pick G A The time formula is expressed in the spatial neighborhood. G A China's overall crime situation Q All category subsets Q a The statistical indicators obtained from observation and counting, at this time N 1 indicates a subset of categories Q a The number of geographic objects in the data. N 2 indicates the overall number of criminal events. Q The number of geographical objects in the middle; when G Pick G a The time formula is expressed in the spatial neighborhood. G a China's overall urban infrastructure P All category subsets P A The statistical indicators obtained from observation and counting, at this time N 1 indicates a subset of categories P A The number of geographic objects in the data. N 2 indicates the overall urban infrastructure P The number of geographic objects in the data; N G for G The number of geographic objects in the text, where * indicates multiplication. x i For 0-1 variables, when G Specific geographic objects i The corresponding category subset is P A or Q a The value is 1 at time.

[0081] ;

[0082] wherein, GWGSI is the global metric level spatial proximity definition method is the geographical weighting k statistical indicator of binary spatial association when the spatial proximity is defined as the geographical weighting spatial distance; w i is the geographical weight assigned to the specific geographical object by the kernel function; i

[0083] S33: for the local metric level, the corresponding spatial neighborhood is determined by a pre-set spatial proximity definition method according to a selected POI data of any one category subset of the city facility population or the crime event population, specifically: P Q

[0084] P A-M → G A-M ;

[0085] Q a-m → G a-m ;

[0086] the observation count of all category subsets of another geographical object population in the obtained spatial neighborhood is performed, specifically:

[0087] G A-M = Q a : n obv(Qa) | n obv(Qa) ∈ c obv ( Q 1), c obv ( Q 2),…, c obv ( Q z );

[0088] G a-m = P A : n obv(PA) | n obv(PA) ∈ c obv ( P 1), c ​​​obv P 2),…, c obv P Z}};

[0089] wherein the preset spatial proximity definition manner comprises spatial distance, geographical weighted spatial distance, k near neighbor and geographical weighted k near neighbor; n obv(Qa) , n obv(PA) are observation counting results of all category subsets in the spatial neighborhood G A-M , G a-m of another geographical object population, c obv is an observation counting operation;

[0090] S34: calculating a statistical index of binary spatial association between specific POI data in a certain category subset in the selected geographical object population and a different category subset in another geographical object population at a local measurement level, specifically:

[0091] , G’ ={ G A-M , G a-m};

[0092] wherein, LSI is a statistical index of binary spatial association between specific POI data in a certain category subset and a category subset in another geographical object population when the spatial proximity definition manner at the local measurement level is k near neighbor or spatial distance; when G’ G A-M , the formula represents a statistical index obtained by performing observation counting on all category subsets in the spatial neighborhood G A-M of the crime event population Q Q a , wherein N 1 represents the number of geographical objects in the category subset Q a , and N 2 represents the number of geographical objects in the crime event population Q ; when G’ G a-m , the formula represents a statistical index obtained by performing observation counting on all category subsets in the spatial neighborhood​​​​​G a-m China's overall urban infrastructure P All category subsets P A The statistical indicators obtained from observation and counting, at this time N 1 indicates a subset of categories P A The number of geographic objects in the data. N 2 indicates the overall urban infrastructure P The number of geographic objects in the data; N G’ for G’ The number of geographic objects in the text, where * indicates multiplication. x j For 0-1 variables, when G’ Specific geographic objects j The corresponding category subset is P A or Q a The value is 1 at this time.

[0093] ;

[0094] in, GWLSI The spatial proximity definition method at the local metric level is geographically weighted. k Statistical indicators of binary spatial associations based on nearest neighbor or geographically weighted spatial distance; w j For specific geographic objects j The assigned geographic weights are given by the kernel function.

[0095] S4: Design a parameterized method based on the observation counts of different category subsets in the spatial neighborhood, perform significance testing on the calculated statistical indicators, and determine the significance of the binary spatial association mining results; perform association analysis and visualization on the significant binary spatial association mining results.

[0096] Furthermore, step S4 includes the following steps:

[0097] S41: Based on the observation counts of POI data from different category subsets in the spatial neighborhood, design a parameterization method based on hypergeometric distribution to perform significance testing on the calculated statistical indicators and determine the significance of the binary spatial association mining results. The formula for the parameterization method is shown below:

[0098] Calculate the hypergeometric distribution X~H ( N 2 , N 1 , N probability distribution of ) P [ C = x]:

[0099] , N = { N G , N G’};

[0100] wherein, N G and N G’ denote the number of geographic objects in the spatial neighborhood G and G’ respectively, when N takes the value N G the formulae represent the significance test of the statistical indicator based on the observed count result at the global metric level, when N takes the value N G’ the formulae represent the significance test of the statistical indicator based on the observed count result at the local metric level; when the significance test of the statistical indicator of the binary spatial association between a certain category subset P A or a specific POI data P A-M and a category subset Q a in another geographic object population is needed, N 1 denotes the number of geographic objects in the category subset Q a , N 2 denotes the number of geographic objects in the population of criminal events Q ; when the significance test of the statistical indicator of the binary spatial association between a certain category subset Q a or a specific POI data Q a-m and a category subset P A in another geographic object population is needed, N 1 denotes the number of geographic objects in the category subset P A , N 2 denotes the number of geographic objects in the population of urban facilities P ; x is the observed count result, i.e. the number of certain geographic objects observed in the spatial neighborhood; the brackets appearing on the right side of the equal sign in the formulae represent the permutation and combination calculation, i.e. the possible number of taking I geographic objects out of J geographic objects;

[0101] ​S42: correlation analysis is performed on the significant binary spatial correlation mining results, specifically as follows:

[0102] When the value of the corresponding statistical index is greater than the set threshold value and the significance test P [ C ≥ x ] is less than the significance threshold value, it is considered that the null hypothesis is rejected at the significance level of the significance threshold value, and there is a positive correlation;

[0103] When the value of the corresponding statistical index is less than the set threshold value and the significance test P [ C ≤ x ] is less than the significance threshold value, it is considered that the null hypothesis is rejected at the significance level of the significance threshold value, and there is a negative correlation;

[0104] Wherein, P [ C ≥ x ] and P [ C ≤ x ] values are obtained according to the probability distribution of hypergeometric distribution X~H ( N 2 , N 1 , N ) is P [ C = x ];

[0105] In a preferred embodiment, the set threshold value is 1, and the significance threshold value is 0.05;

[0106] S43: The obtained statistically significant significant local hierarchical binary spatial correlation is visualized by geographic information system professional software, and the analysis result of the potential relationship and dependence between the city facility type and the crime event type is obtained.

[0107] Referring to Figure 2 , in another embodiment of the present application, a spatial correlation mining system across different types of geographic objects is provided, comprising the following modules:

[0108] The data acquisition and preprocessing module 10 is used to acquire city road data and POI data of city facilities and crime events, and to perform preprocessing procedures such as debugging and road network matching on the POI data; a road network model is constructed, the road network distance between the POI data is calculated, and the city facility population and the crime event population contained in the POI data are classified into subsets according to prior knowledge;

[0109] The statistical index calculation module 20 sets a suitable null hypothesis and a semantic statistical index definition strategy, constructs a respective spatial neighborhood for different category subsets or POI data in the urban facility population and the crime event population by selecting different spatial neighborhood definition manners; from the global measurement level and the local measurement level, calculates the statistical index of the binary spatial association between different category subsets or POI data and different category subsets in another geographical object population according to the preset spatial neighborhood definition manner in the spatial neighborhood;

[0110] The analysis module 30 is used for implementing the significance test and visualization analysis of the statistical index, designs a parameterized method based on the hypergeometric distribution according to the neighborhood count of different category subsets in the spatial neighborhood, performs the significance test on the calculated statistical index, determines the significance of the binary spatial association mining result, and then performs the association analysis and visualization display on the significant binary spatial association mining result by using the geographic information system professional software.

[0111] Further, in the data acquisition and preprocessing module, the POI data of the urban road data, the urban facilities and the crime events are acquired, and the POI data is subjected to the preprocessing procedures such as debugging and road network matching. In this way, it can be ensured that the acquired data can comprehensively reflect the spatial distribution of the urban facilities and the crime events on the urban road network. At the same time, the road network model is constructed, and the road network distance between the POI data is calculated. The road network distance plays an important role in human activities in urban areas, because human activities are usually limited to the street network structure. The use of the road network distance takes into account the non-uniformity of the urban space and the complex road network structure, and has good practical significance.

[0112] It should be noted that the system provided by the above embodiment is only exemplified by the division of the above functional modules, and in actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure is divided into different functional modules to complete all or part of the functions described above. The system can be applied to the above-mentioned cross-population spatial association mining method between different types of geographical objects.

[0113] It should be understood that parts of the present application can be realized by hardware, software, firmware or their combination. In the above embodiments, a plurality of steps or methods can be realized by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if realized by hardware, and as in another embodiment, it can be realized by any one or a combination of the following technologies known in the art: discrete logic circuit with logic gate circuit for implementing logic function on data signal, application specific integrated circuit with suitable combination logic gate circuit, programmable gate array (PGA), field programmable gate array (FPGA) and the like.

[0114] The above embodiments are the preferred embodiments of the present application, but the embodiments of the present application are not limited to the above embodiments, and any changes, modifications, substitutions, combinations, simplifications, etc. made without departing from the spirit and principles of the present application should be equivalent replacement manners and should be included in the protection scope of the present application.

Claims

1. A method for mining spatial associations among different types of geographic objects across a general population, characterized in that, The method comprises the following steps: Acquire urban road data within the study area, as well as the overall first geographic object. P Second Geographic Object Overall Q POI data; perform error correction and road network matching on the acquired data; calculate the road network distance between POI data and construct a road network model; the first geographic object overall... P With the second geographic object as a whole Q The total number of different geographic objects to be spatially correlated; POI data of the first population of geographical objects P and the second population of geographical objects Q are classified into different category subsets according to prior knowledge; the first population of geographical objects P and the second population of geographical objects Q are set to be mutually independent, and the spatial distribution within each population of geographical objects follows the zero hypothesis of random labeling; The statistical indicator defining the binary spatial association is the ratio of the observed value to the expected value under the null hypothesis; by choosing different ways of defining spatial proximity, the first population of geographic objects P and the second population of geographic objects Q are divided into different subsets of classes or POI data to construct respective spatial neighborhoods; counting observations of different category subsets in another geographical object population in a spatial neighborhood corresponding to the selected geographical object population from the global measurement level and the local measurement level, and calculating statistical indicators of binary spatial associations between different category subsets in the selected geographical object population or POI data and different category subsets in another geographical object population; performing significance test on the statistical indicators according to the counting results of the observations to determine the significance of the binary spatial association mining results; performing association analysis and visual display on the significant binary spatial association mining results.

2. The method of claim 1, wherein the method further comprises: The POI data of the first geographical object population P and the second geographical object population Q are divided into different category subsets according to prior knowledge, specifically: a first population of geographical objects P and a second population of geographical objects Q sorted by different subtypes: P ={ P 1,..., P A ,..., P Z}, A ∈{1,2,…, Z}; Q ={ Q 1,..., Q a ,..., Q z}, a ∈{1,2,…, z}; wherein, P A and Q a are different subsets of categories, Z and z are the total number of subsets of categories for the corresponding population of geographical objects, A and a are indices of the subsets of categories; each subset of categories comprises a number of POI data, denoted as: P A = { P A-1 ,..., P A-M ,..., P A-O}, Q a = { Q a-1 ,..., Q a-m ,..., Q a-o}, wherein, M ∈ { 1, 2,..., O}, m ∈ { 1, 2,..., o}, O and o are the total number of POI data for the corresponding subset of categories, M and m are the indices of the POI data for the corresponding subset of categories.​​ 3. The method of claim 2, wherein the method further comprises: For the global metric level, the corresponding spatial neighborhood is determined according to all POI data of any one category subset of the first geographic object population P and the second geographic object population Q by a pre-set spatial proximity definition method, specifically as follows: P A → G A ； Q a → G a ； counting observations of all category subsets in another geographical object population in the obtained spatial neighborhood, specifically: G A ={ Q a : N obv(Qa) | N obv(Qa) ∈{ C obv ( Q 1), C obv ( Q 2),…, C obv ( Q z )}}； G a ={ P A : N obv(PA) | N obv(PA) ∈{ C obv ( P 1), C obv ( P 2), …, C obv ( P Z )}}； The preset spatial neighborhood definition manner includes a spatial distance, a geographical weighted spatial distance, k proximity and geographical weighting k proximity; N obv(Qa) , N obv(PA) is a result of observing counting all category subsets of another geographical object population in a spatial neighborhood G A , G a Q a , P A , C obv is an observation counting operation.​ 4. The method of claim 3, wherein the method further comprises: at the global measurement level, calculating statistical indicators of binary spatial associations between a certain category subset in the selected geographical object population and different category subsets in another geographical object population, specifically: ， G ={ G A , G a}； wherein, GSI is the definition of spatial proximity at the global level of measurement k is a statistical indicator of the binary spatial association between a subset of categories in a certain geographical object population and a subset of categories in another geographical object population; when G is taken as G A the formula represents the statistical indicator obtained by counting the observations in the spatial neighborhood G A of the second geographical object population Q for all subsets of categories Q a , where N 1 represents the number of geographical objects in the subset of categories Q a , and N 2 represents the number of geographical objects in the second geographical object population Q ; when G is taken as G a the formula represents the statistical indicator obtained by counting the observations in the spatial neighborhood G a of the first geographical object population P for all subsets of categories P A , where N 1 represents the number of geographical objects in the subset of categories P A , and N 2 represents the number of geographical objects in the first geographical object population P ; N G is the number of geographical objects in G , * represents multiplication, x i is a 0-1 variable, which takes the value 1 when G a certain geographical object in i corresponds to P A or Q a ; ； wherein, GWGSI The definition of spatial proximity under the global metric hierarchy is a geographical weighting k The statistical indicator of binary spatial association when the proximity is a geographically weighted spatial distance. w i For a specific geographical object i The geographical weight assigned to it is given by the kernel function.

5. The method of claim 2, wherein the method further comprises: For the local metric level, the corresponding spatial neighborhood is determined according to a selected POI data of any one category subset of the first geographic object population P and the second geographic object population Q by a pre-set spatial proximity definition manner, specifically as follows: P A-M → G A-M ； Q a-m → G a-m ； counting observations of all category subsets in another geographical object population in the obtained spatial neighborhood, specifically: G A-M ={ Q a : n obv(Qa) | n obv(Qa) ∈{ c obv ( Q 1), c obv ( Q 2),…, c obv ( Q z )}}； G a-m ={ P A : n obv(PA) | n obv(PA) ∈{ c obv ( P 1), c obv ( P 2),…, c obv ( P Z )}}； The preset spatial neighborhood definition manner includes a spatial distance, a geographical weighted spatial distance, k proximity and geographical weighting k proximity; n obv(Qa) , n obv(PA) are respectively observation counts of all category subsets of another geographical object population in a spatial neighborhood G A-M , G a-m , c obv is an observation counting operation.

6. The method of claim 5, wherein the step of mining spatial correlations between different types of geographic objects comprises the steps of: determining a set of spatial correlations between the different types of geographic objects; and determining a set of spatial correlations between the different types of geographic objects based on the set of spatial correlations. at the local measurement level, calculating statistical indicators of binary spatial associations between specific POI data in a certain category subset in the selected geographical object population and different category subsets in another geographical object population, specifically: ， G’ ={ G A-M , G a-m}； wherein, LSI The definition of spatial proximity under the local metric hierarchy is k The spatial proximity or spatial distance is a statistical indicator of the binary spatial association between a particular POI data in a category subset and a category subset in another geographical object population; when G’ = 1, G A-M The formula represents the statistical indicator of the observation count of all category subsets in the spatial neighborhood G A-M of the second geographical object population Q Q a The formula represents the statistical indicator of the observation count of all category subsets in the spatial neighborhood N Q a of the first geographical object population N Q G’ = 1, G a-m The formula represents the statistical indicator of the observation count of all category subsets in the spatial neighborhood G a-m of the second geographical object population P P A The formula represents the statistical indicator of the observation count of all category subsets in the spatial neighborhood N P A The formula represents the statistical indicator of the observation count of all category subsets in the spatial neighborhood N P N G’ The formula represents the statistical indicator of the observation count of all category subsets in the spatial neighborhood G’ x j The formula represents the statistical indicator of the observation count of all category subsets in the spatial neighborhood G’ j P A The formula represents the statistical indicator of the observation count of all category subsets in the spatial neighborhood Q a The formula represents the statistical indicator of the observation count of all category subsets in the spatial neighborhood​​​​​​​​​​​ ； wherein, GWLSI The definition of spatial proximity at the local metric level is a geographical weighting k The statistical indicator of the binary spatial association when the proximity or the geographical weighted spatial distance; w j For a specific geographical object j The geographical weight assigned to it is given by the kernel function.

7. The method of claim 1, wherein the method further comprises: performing significance test on the statistical indicators according to the counting results of the observations, specifically: Computing hypergeometric distribution X~H ( N 2 , N 1 , N ) probability distribution P [ C = x ] ， N ={ N G , N G’}； wherein, N G and N G’ denote the number of geographic objects in the spatial neighborhood G and G’ , respectively, when N takes the value N G The formulae denote the significance test of the statistical indicator based on the observed count result at the global metric level when N takes the value N G’ The formulae denote the significance test of the statistical indicator based on the observed count result at the local metric level when P A or the specific POI data P A-M and a subset of categories Q a in another geographic object population, N 1 denotes the number of geographic objects in the subset of categories Q a , and N 2 denotes the number of geographic objects in the second geographic object population Q ; when the significance test of the statistical indicator of the binary spatial association between a subset of categories Q a or the specific POI data Q a-m and a subset of categories P A in another geographic object population, N 1 denotes the number of geographic objects in the subset of categories P A , and N 2 denotes the number of geographic objects in the first geographic object population P ; x denotes the observed count result; the appearing brackets on the right side of the formulae denote the permutation and combination calculation, i.e. the possible number of taking I geographic objects out of J geographic objects.

8. The method of claim 7, wherein the method further comprises: the association analysis on the significant binary spatial association mining results is specifically: When the value of the corresponding statistical indicator is greater than the set threshold and the significance test is successful... P [ C ≥ x If the value is less than the significance threshold, then the null hypothesis is rejected at the significance level of the significance threshold, indicating a positive association. When the value of the corresponding statistical indicator is less than a set threshold and the significance test is P [ C ≤ x ] less than the significance threshold, then the null hypothesis is rejected at the significance level of the significance threshold, and there is a negative association. wherein P [ C ≥ x ] and P [ C ≤ x ] values are obtained from the probability distribution of hypergeometric distribution X~H ( N 2 , N 1 , N ) distribution P [ C = x ].

9. The method of claim 1-8, wherein, obtaining urban road data within the study area, as well as POI data for a first geographic object population and a second geographic object population, by means of geographic information system software and open source data P and a second geographic object population Q ​ simplifying, topology checking, loop removal and debugging the city road data, debugging the POI data of the first and second geographical object populations P and second geographical object populations Q ​ The first geographical object aggregate P is an aggregate of urban facilities. The second geographical object aggregate Q is an aggregate of crime events.

10. A system for mining spatial correlations across different types of geographic objects in general, characterized in that, The method for mining spatial associations between different types of geographical objects across populations according to any one of claims 1-8 comprises a data collection and preprocessing module, a statistical indicator calculation module, and an analysis module; The data acquisition and preprocessing module obtains urban road data within the study area, as well as the overall data of the first geographic object. P Second Geographic Object Overall Q POI data; perform error correction and road network matching on the acquired data; calculate the road network distance between POI data and construct a road network model; the first geographic object overall... P With the second geographic object as a whole Q The population of different geographic objects to be spatially correlated is defined; for the first geographic object population... P Second Geographic Object Overall Q The POI data is divided into different category subsets based on prior knowledge; The statistical index calculation module is configured to set a first geographical object population P independent of a second geographical object population Q and the spatial distribution within each geographical object population follows the zero hypothesis of random marking. The statistical indicator defining the binary spatial association is the ratio of the observed value to the expected value under the null hypothesis; by choosing different ways of defining spatial proximity, the first population of geographic objects P and the second population of geographic objects Q are divided into different subsets of classes or POI data to construct respective spatial neighborhoods; counting observations of different category subsets in another geographical object population in a spatial neighborhood corresponding to the selected geographical object population from the global measurement level and the local measurement level, and calculating statistical indicators of binary spatial associations between different category subsets in the selected geographical object population or POI data and different category subsets in another geographical object population; the analysis module is used for performing significance test on the statistical indicators according to the counting results of the observations to determine the significance of the binary spatial association mining results; performing association analysis and visual display on the significant binary spatial association mining results.

Images