Crime hotspot detection method, system and equipment

By constructing a geo-weighted regression model and EM algorithm, the problem of insufficient efficiency, accuracy and adaptability of crime hot spot detection in the existing technology is solved, efficient and accurate detection of unrecorded criminal events is achieved, and a credible crime space is obtained Hot topics, help the police adjust their prevention and control strategies.

CN114969157BActive Publication Date: 2024-11-15CENT SOUTH UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210618003.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-06-01
Publication Date
2024-11-15
Estimated Expiration
2042-06-01

AI Technical Summary

Technical Problem

The existing methods of crime hot spot detection are poor in efficiency, accuracy and adaptability when facing the crime black numbers, and cannot effectively detect unrecorded criminal events, resulting in incorrect hot spot detection results.

Method used

By eliminating cases without coordinate information in the case record, a priori distribution of crime and POI distribution are generated, a geo-weighted regression model is constructed, the positional condition probability of crime record points is calculated, and a priori distribution of crime is corrected with grid coordinates. EM algorithm is used to mine representative crimes. The subdistribution of events that occur frequently serves as hot spot areas.

Benefits of technology

It improves the efficiency, accuracy and adaptability of criminal incident detection, eliminates the impact of spatial uncertainty, and obtains credible criminal space hotspots, which can more accurately identify the relationship between criminal behavior and urban environment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114969157B_ABST
    Figure CN114969157B_ABST
Patent Text Reader

Abstract

The disclosed embodiments provide a method, system and device for detecting hot spots of crime incidents, which belong to the field of data processing technology, and specifically include: eliminating cases without coordinate information in case records; dividing the spatial research area into multiple grids, and defining the coordinates of each grid; generating crime prior distribution and POI distribution according to the coordinates of crime record point data and POI data; constructing a geographically weighted regression model, calculating the conditional probability of the location of the crime record point, and correcting the crime prior distribution in combination with the coordinates of each grid and the geographically weighted regression model to obtain a complete crime distribution; mining sub-distributions representing frequent occurrences of crime incidents from the complete crime distribution as hot spot areas. Through the disclosed scheme, the criminal event prior distribution is corrected by the geographically weighted regression model, the influence of spatial uncertainty is eliminated to obtain a complete crime distribution, and credible crime spatial hot spots are mined, thereby improving the efficiency, accuracy and adaptability of crime event detection.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The embodiments of the present disclosure relate to the field of data processing technology, and in particular to a method, system and device for detecting hot spots of criminal incidents. Background Art

[0002] At present, with the acceleration of urbanization, urban population mobility is becoming more frequent, the structure is becoming more complex, social problems and contradictions are gradually emerging, and crime is particularly prominent and important, posing a serious threat to people's livelihood and social stability. A large number of studies in urban criminology have shown that crime is closely related to the geographical environment, and the occurrence of urban crime incidents usually forms a localized distribution pattern in space, namely "crime spatial hotspots". Therefore, the use of crime spatial hotspot detection technology to achieve precise crackdown and prevention and control of urban crime has always been a focus of attention in fields such as urban criminology and environmental geography.

[0003] Existing crime hotspot detection methods can be divided into mapping and data mining methods according to detection technology. Mapping is based on crime record data and basic geographic data. Through spatial analysis methods such as overlay analysis and density estimation, thematic maps are drawn to intuitively display the spatial distribution of crime hotspots. For example, the crime density distribution map is generated by kernel density estimation. Taking property crime as an example, the kernel density estimation and standard deviation ellipse model are used to find that its hotspot areas are mainly concentrated in large commercial areas and near railway stations. The street crime pattern under the constraint of road network is studied by network kernel density estimation and Getis-Ord GI* statistics. The results show that urban street crimes are mainly concentrated in the city center, and the specific details of their distribution are affected by urban functional areas such as commercial and residential areas. The data mining method uses unsupervised learning methods to quantitatively calculate the spatial location and morphology of crime hotspots. Among them, the density-based clustering algorithm is feasible in mining the crime hotspot patterns in Chicago and New York City. The results show that its clustering method can accurately detect crime hotspot areas with complex boundary morphology; or taking into account the scale effect of crime hotspots, hierarchical clustering is used to further explore the differentiation of crime hotspots at different spatial scales. Most existing crime hotspot detection methods are based on the assumption that crime record data is complete. However, in the real world, some individuals may be victims of actual crimes, but they do not choose to report the crime for subjective reasons such as personal privacy or private reconciliation, resulting in the crime not being recorded. This phenomenon is also known as the crime black number phenomenon in the field of criminology. The crime black number phenomenon results in some areas where no crime incidents are recorded, but there may have been crimes. The existing hotspot detection methods ignore this phenomenon, which will lead to incorrect hotspot detection results.

[0004] It can be seen that there is an urgent need for an efficient, accurate and adaptable crime time hotspot detection method. Summary of the invention

[0005] In view of this, the embodiments of the present disclosure provide a method, system and device for detecting hot spots of criminal incidents, which at least partially solve the problems of poor detection efficiency, accuracy and adaptability in the prior art.

[0006] In a first aspect, an embodiment of the present disclosure provides a method for detecting a hotspot of a crime incident, including:

[0007] Step 1, remove cases without coordinate information in the case records, and obtain the coordinates of the crime record point data and POI data;

[0008] Step 2, dividing the spatial research area into a plurality of grids, and defining the coordinates of each of the grids;

[0009] Step 3, generating crime prior distribution and POI distribution according to the coordinates of the crime record point data and the POI data;

[0010] Step 4, based on the co-occurrence relationship between criminal events and the built environment, a geographically weighted regression model is constructed according to the crime prior distribution and the POI distribution, the conditional probability of the crime record point location is calculated, and the crime prior distribution is corrected by combining the coordinates of each grid and the geographically weighted regression model to obtain a complete crime distribution;

[0011] Step 5: mining sub-distributions representing frequent crime incidents from the complete crime distribution as hot spots.

[0012] According to a specific implementation of the embodiment of the present disclosure, step 3 specifically includes:

[0013] Step 3.1, by formula Calculate the crime prior distribution f i (p) (s), where K(·) is the adaptive kernel function, dist(·,·) represents the Euclidean distance, Representative C i The total number of crimes of this type, Represents C i The instance of the jth crime event under the class crime type, Represents its geographical location coordinates;

[0014] Step 3.2, the research results on the correlation between different crime types and POI data, select and obtain the common N related to each type of crime p POI-like data;

[0015] Step 3.3, using the formula Calculate the POI distribution f (pk) (s), where pk,k∈[1,N p ] represents the POI type, represents the total number of POI instances of the pk-th type of POI, Represents the lth POI instance under this POI type, Represents its geographic location coordinates.

[0016] According to a specific implementation of the embodiment of the present disclosure, in step 4, the step of calculating the conditional probability of the crime record point location and correcting the crime prior distribution in combination with the coordinates of each grid and the geographically weighted regression model to obtain the complete crime distribution includes:

[0017] Step 4.1, for crime type C i ,i∈[1,N C ], for each crime record point Take the crime prior distribution f i (p) (s) and POI distribution f (pk) The probability values ​​of the corresponding positions in (s) are filtered and eliminated through ordinary least squares and student-t test.

[0018] Step 4.2: For each type of crime event C i , the probability value of the corresponding event instance location in the filtered POI distribution As an explanatory variable, the probability value of the corresponding position in the crime prior distribution As the dependent variable, the conditional probability is calculated by the geographically weighted regression model

[0019] Step 4.3: Based on the conditional probability and the geographically weighted regression model, predict the complete distribution of crime in the entire study area. i (s).

[0020] According to a specific implementation of the embodiment of the present disclosure, step 4.1 specifically includes:

[0021] Step 4.1.1, for each crime record instance point The prior probability value of crime and POI distribution probability value Building an OLS model in is the OLS parameter, is the residual, and the OLS parameters are solved by minimizing the residual sum of squares;

[0022] Step 4.1.2, for each regression parameter b k ,k∈[1,N p ]Construct T that follows t distribution kStatistics;

[0023] Step 4.1.3, according to the formula Calculate the t value;

[0024] Step 4.1.4, set the significance level threshold to α. If the probability P{|T k |>t} is greater than the significance level α, which means that POI type pk has no significant effect on the occurrence of crime incidents and it is eliminated from the next step of conditional probability calculation.

[0025] According to a specific implementation of the embodiment of the present disclosure, the conditional probability The expression is

[0026] in, is the geographical weighted parameter, which is solved by weighted least squares method The spatial weight function is selected as a Gaussian function, and the bandwidth is determined by the minimum AIC information criterion.

[0027] According to a specific implementation of the embodiment of the present disclosure, the crime complete distribution f i The expression of (s) is Among them, s∈S refers to the two-dimensional geographic coordinates of each grid in the geographic space area, and f (pk) (s) represents the probability value corresponding to the POI distribution at position s, Represents the geographic weighting parameter at location s.

[0028] According to a specific implementation of the embodiment of the present disclosure, step 5 specifically includes:

[0029] Step 5.1: Given the number of sub-distributions H, define the parameter solution process as the optimization solution process Then the optimization problem is solved by EM algorithm;

[0030] Step 5.2: construct Gaussian mixture models for different H values, and use the formula BIC = mln (N S )-2ln(L) to calculate its Bayesian information, where m is the number of model parameters, N S is the number of spatial grid instances, L is the likelihood function, and the H value corresponding to the minimum BIC is selected as the number of sub-distributions;

[0031] Step 5.3: define the sub-distribution with the highest probability average in the spatial grid cell as the hotspot sub-distribution

[0032] Step 5.4, The area range contained in the grid instance is used as the hot spot area.

[0033] According to a specific implementation of the embodiment of the present disclosure, the EM algorithm is composed of an E step and an M step, and the step 5.1 specifically includes:

[0034] Step 5.1.1, step E approximates the lower bound of the random function by solving the expected value:

[0035]

[0036] Step 5.1.2, step M uses maximum likelihood estimation to estimate the parameter p h ,μ h ,σ h Make an update

[0037]

[0038]

[0039]

[0040] Step 5.1.3, update the parameters through the iterative cycle of E step and M step until convergence to the likelihood function Reached maximum value.

[0041] In a second aspect, the present disclosure provides a crime hotspot detection system, including:

[0042] The elimination module is used to eliminate cases without coordinate information in the case records and obtain the coordinates of the crime record point data and POI data;

[0043] A partitioning module, used for dividing the spatial research area into a plurality of grids and defining the coordinates of each of the grids;

[0044] A generation module, for generating crime prior distribution and POI distribution according to the coordinates of the crime record point data and the POI data;

[0045] A construction module is used to construct a geographically weighted regression model based on the co-occurrence relationship between crime events and the built environment, according to the crime prior distribution and the POI distribution, calculate the conditional probability of the crime record point location, and correct the crime prior distribution in combination with the coordinates of each grid and the geographically weighted regression model to obtain a complete crime distribution;

[0046] The prediction module is used to mine sub-distributions representing frequent occurrences of criminal incidents from the complete crime distribution as hot spots.

[0047] In a third aspect, an embodiment of the present disclosure further provides an electronic device, the electronic device comprising:

[0048] at least one processor; and,

[0049] a memory communicatively connected to the at least one processor; wherein,

[0050] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the crime event hotspot detection method in the aforementioned first aspect or any implementation of the first aspect.

[0051] The crime hotspot detection scheme in the disclosed embodiment includes: step 1, eliminating cases without coordinate information in the case records, and obtaining the coordinates of the crime record point data and the POI data; step 2, dividing the spatial research area into multiple grids, and defining the coordinates of each of the grids; step 3, generating a crime prior distribution and a POI distribution according to the coordinates of the crime record point data and the POI data; step 4, based on the co-occurrence relationship between criminal events and the built environment, constructing a geographically weighted regression model according to the crime prior distribution and the POI distribution, calculating the conditional probability of the crime record point location, and correcting the crime prior distribution in combination with the coordinates of each of the grids and the geographically weighted regression model to obtain a complete crime distribution; step 5, mining sub-distributions representing frequent occurrences of criminal events from the complete crime distribution as hotspot areas.

[0052] The beneficial effects of the embodiments of the present disclosure are as follows: through the scheme of the present disclosure, the relationship between crime events and the urban built environment is learned through the geographically weighted regression model, the prior distribution of crime events is corrected, thereby eliminating the influence of spatial uncertainty, obtaining a complete distribution of crime, and using the EM algorithm to mine credible crime spatial hotspots, thereby improving the efficiency, accuracy and adaptability of crime event detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0054] Figure 1 A flowchart of a method for detecting hot spots of criminal incidents provided by an embodiment of the present disclosure;

[0055] Figure 2 A flowchart of a crime event spatial hotspot detection method taking into account data uncertainty, which is provided in a crime event hotspot detection method according to an embodiment of the present disclosure;

[0056] Figure 3A schematic diagram of a crime spatial hotspot detected by using the crime event hotspot detection method of the present invention provided in an embodiment of the present disclosure;

[0057] Figure 4 A schematic diagram of crime spatial hot spots detected by a method based on data completeness assumption provided in an embodiment of the present disclosure;

[0058] Figure 5 A schematic diagram of the structure of a crime hotspot detection system provided by an embodiment of the present disclosure;

[0059] Figure 6 A schematic diagram of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0060] The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0061] The following describes the embodiments of the present disclosure through specific examples, and those skilled in the art can easily understand other advantages and effects of the present disclosure from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all of the embodiments. The present disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present disclosure. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present disclosure.

[0062] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein may be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on the present disclosure, it should be understood by those skilled in the art that an aspect described herein may be implemented independently of any other aspect, and two or more of these aspects may be combined in various ways. For example, any number of aspects described herein may be used to implement the device and / or practice the method. In addition, other structures and / or functionalities other than one or more of the aspects described herein may be used to implement this device and / or practice this method.

[0063] It should also be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present disclosure. The drawings only show components related to the present disclosure rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0064] Additionally, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, it will be understood by those skilled in the art that the aspects described may be practiced without these specific details.

[0065] The disclosed embodiment provides a method for detecting hot spots of criminal incidents, which can be applied to the process of detecting crime time hot spots in urban crime combating and prevention scenarios.

[0066] See also Figure 1 , is a flow chart of a method for detecting hot spots of criminal incidents provided by an embodiment of the present disclosure. Figure 1 and Figure 2 As shown, the method mainly comprises the following steps:

[0067] Step 1, remove cases without coordinate information in the case records, and obtain the coordinates of the crime record point data and POI data;

[0068] For example, based on the real police data recorded by the Municipal Public Security Bureau in the main urban area of ​​a prefecture-level city in the central and southern part of the Yangtze River Delta region of my country in 2015, as an actual data set for the implementation of a specific invention, including injury cases (2055 cases), burglary cases (855 cases), electric vehicle theft cases (653 cases) and Internet fraud cases (523 cases), the implementation process of the present invention is explained.

[0069] You can check the tabular data of case records, clear cases with missing coordinate information, and obtain the coordinates of crime record point data and POI data.

[0070] Step 2, dividing the spatial research area into a plurality of grids, and defining the coordinates of each of the grids;

[0071] In the specific implementation, for the spatial research area S, define s∈S as a random variable representing the spatial position coordinates. Divide the spatial area S into N equal areas s grid gird(S), that is definition gird(S) i The spatial coordinate value of the upper right corner.

[0072] Step 3, generating crime prior distribution and POI distribution according to the coordinates of the crime record point data and the POI data;

[0073] Based on the above embodiment, step 3 specifically includes:

[0074] Step 3.1, by formula Calculate the crime prior distribution f i(p) (s), where K(·) is the adaptive kernel function, dist(·,·) represents the Euclidean distance, Representative C i The total number of crimes of this type, Represents C i The instance of the jth crime event under the class crime type, Represents its geographical location coordinates;

[0075] Step 3.2, the research results on the correlation between different crime types and POI data, select and obtain the common N related to each type of crime p POI-like data;

[0076] Step 3.3, using the formula Calculate the POI distribution f (pk) (s), where pk,k∈[1,N p ] represents the POI type, represents the total number of POI instances of the pk-th type of POI, Represents the lth POI instance under this POI type, Represents its geographic location coordinates.

[0077] In the specific implementation, based on the coordinates of the crime record point data and POI data, the crime prior distribution f is generated respectively. i (p) (s) and POI distribution f (pk) (s). The specific steps can be as follows:

[0078] (3.1) Through the formula Calculate the crime prior distribution f i (p) (s), that is, for a specific crime type C i ,i∈[1,N C ], the original spatial probability distribution with uncertainty described by its incomplete record. Where K(·) is the adaptive kernel function, dist(·,·) represents the Euclidean distance, Representative C i The total number of crimes of this type, Represents C i The instance of the jth crime event under the class crime type, Represents its geographic location coordinates.

[0079] (3.2) Based on the existing research results of urban criminology on the correlation between different crime types and POIs, select and obtain the common N related to each type of crime. p= 8 categories of POI data, including shopping services, science, education and culture services, sports and leisure services, accommodation services, commercial residences, medical care services, financial insurance services, and public security, procuratorial and judicial institutions.

[0080] (3.3) Using the formula Calculate POI distribution f (pk) (s), where pk,k∈[1,N p ] represents the POI type, represents the total number of POI instances of the pk-th type of POI, Represents the lth POI instance under this POI type, Represents its geographic location coordinates.

[0081] Step 4, based on the co-occurrence relationship between criminal events and the built environment, a geographically weighted regression model is constructed according to the crime prior distribution and the POI distribution, the conditional probability of the crime record point location is calculated, and the crime prior distribution is corrected by combining the coordinates of each grid and the geographically weighted regression model to obtain a complete crime distribution;

[0082] Furthermore, in step 4, the step of calculating the conditional probability of the crime record point location and correcting the crime prior distribution in combination with the coordinates of each grid and the geographically weighted regression model to obtain the complete crime distribution includes:

[0083] Step 4.1, for crime type C i ,i∈[1,N C ], for each crime record point Take the crime prior distribution f i (p) (s) and POI distribution f (pk) The probability values ​​of the corresponding positions in (s) are filtered and eliminated through ordinary least squares and student-t test.

[0084] Step 4.2: For each type of crime event C i , the probability value of the corresponding event instance location in the filtered POI distribution As an explanatory variable, the probability value of the corresponding position in the crime prior distribution As the dependent variable, the conditional probability is calculated by the geographically weighted regression model

[0085] Step 4.3: Based on the conditional probability and the geographically weighted regression model, predict the complete distribution of crime in the entire study area. i (s).

[0086] Furthermore, the step 4.1 specifically includes:

[0087] Step 4.1.1, for each crime record instance point The prior probability value of crime and POI distribution probability value Building an OLS model in is the OLS parameter, is the residual, and the OLS parameters are solved by minimizing the residual sum of squares;

[0088] Step 4.1.2, for each regression parameter b k ,k∈[1,N p ]Construct T that follows t distribution k Statistics;

[0089] Step 4.1.3, according to the formula Calculate the t value;

[0090] Step 4.1.4, set the significance level threshold to α. If the probability P{|T k |>t} is greater than the significance level α, which means that POI type pk has no significant effect on the occurrence of crime incidents and it is eliminated from the next step of conditional probability calculation.

[0091] Optionally, the conditional probability The expression is

[0092]

[0093] in, is the geographical weighted parameter, which is solved by weighted least squares method The spatial weight function is selected as a Gaussian function, and the bandwidth is determined by the minimum AIC information criterion.

[0094] Optionally, the complete crime distribution f i The expression of (s) is Among them, s∈S refers to the two-dimensional geographic coordinates of each grid in the geographic space area, and f (pk) (s) represents the probability value corresponding to the POI distribution at position s, Represents the geographic weighting parameter at location s.

[0095] In the specific implementation, considering the impact of spatial uncertainty, it is believed that in the crime prior distribution, only the location of the crime record point has an accurate crime probability. Based on the co-occurrence relationship between crime events and the built environment, a geographically weighted regression (GWR) model is constructed to calculate the conditional probability of the crime record point location, and the crime prior distribution is corrected to obtain the complete crime distribution. The specific steps are as follows:

[0096] (4.1) For crime type C i ,i∈[1,N C ], for each crime record point Take the crime prior distribution f i (p) (s) and POI distribution f (pk) The probability values ​​of the corresponding positions in (s) are filtered and eliminated through ordinary least squares (OLS) and student-t test. The specific steps are as follows.

[0097] (4.1.1) For each crime record instance point The prior probability value of crime and POI distribution probability value Building an OLS model in is the OLS parameter, The residual is the OLS parameter, which is solved by minimizing the residual sum of squares.

[0098] (4.1.2) For each regression parameter b k ,k∈[1,N p ]Construct T that follows t distribution k Statistics, that is in represents the estimated values ​​of the regression parameters, represents the standard deviation of the regression parameter estimates.

[0099] (4.1.3) Calculate the t value

[0100] (4.1.4) Set the significance level threshold to α = 0.015. If the probability P{|T k |>t} is greater than the significance level α, which means that POI type pk has no significant effect on the occurrence of crime incidents and it is eliminated from the next step of conditional probability calculation.

[0101] (4.2) For each type of crime incident C i , the probability value of the corresponding event instance location in the filtered POI distribution As an explanatory variable, the probability value of the corresponding position in the crime prior distribution As the dependent variable, the conditional probability is calculated through the GWR model That is, given the POI distribution f (pk) (s), the probability value of the crime event record location should be as shown in the following formula:

[0102]

[0103] in, is the geographical weighted parameter. The geographical weighted parameter is solved by weighted least squares method. The spatial weight function is selected as a Gaussian function, and the bandwidth is determined by the minimum AIC information criterion.

[0104] (4.3) Furthermore, based on the conditional probability and GWR model, the complete probability distribution of crime in the entire study area is predicted. i (s), i.e., the specific crime event C after correcting the uncertainty i The spatial probability distribution of is as follows:

[0105]

[0106] Among them, s∈S refers to the two-dimensional geographic coordinates in the geographic space area, f (pk) (s) represents the probability value corresponding to the POI distribution at position s, Represents the geographic weighting parameter at location s.

[0107] Step 5: mining sub-distributions representing frequent crime incidents from the complete crime distribution as hot spots.

[0108] Based on the above embodiment, step 5 specifically includes:

[0109] Step 5.1: Given the number of sub-distributions H, define the parameter solution process as the optimization solution process Then the optimization problem is solved by EM algorithm;

[0110] Step 5.2: construct Gaussian mixture models for different H values, and use the formula BIC = mln (N S )-2ln(L) to calculate its Bayesian information, where m is the number of model parameters, N S is the number of spatial grid instances, L is the likelihood function, and the H value corresponding to the minimum BIC is selected as the number of sub-distributions;

[0111] Step 5.3: define the sub-distribution with the highest probability average in the spatial grid cell as the hotspot sub-distribution

[0112] Step 5.4, The area range contained in the grid instance is used as the hot spot area.

[0113] Furthermore, the EM algorithm is composed of an E step and an M step, and step 5.1 specifically includes:

[0114] Step 5.1.1, step E approximates the lower bound of the random function by solving the expected value:

[0115]

[0116] Step 5.1.2, step M uses maximum likelihood estimation to estimate the parameter p h ,μ h ,σ h Make an update

[0117]

[0118]

[0119]

[0120] Step 5.1.3, update the parameters through the iterative cycle of E step and M step until convergence to the likelihood function Reached maximum value.

[0121] In the specific implementation, in order to mine the sub-distribution representing the frequent occurrence of criminal events from the complete spatial distribution of crime, that is, the hot spot sub-distribution Define f i (s) is a mixed Gaussian distribution containing hotspot sub-distributions:

[0122]

[0123] Where s represents the geographic coordinates of any location in the spatial region, H is the number of specified sub-distributions, and p h ,μ h ,σ h is the mixed distribution f i The parameters of each sub-distribution in (s) are solved by the EM algorithm. The specific calculation steps are as follows:

[0124] (5.1) Given the number of sub-distributions H, the parameter solution process is defined as the optimization solution process:

[0125]

[0126] Then the optimization problem is solved by the EM algorithm. The EM algorithm consists of the E step and the M step. The specific steps are as follows:

[0127] (5.1.1) In the E step, the lower bound of the approximate random function is obtained by solving the expected value:

[0128] (5.1.2) In the M step, the maximum likelihood estimation is used to estimate the parameter p h ,μ h ,σ h Make an update

[0129]

[0130]

[0131] (5.1.3) Update the parameters through the iterative cycle of E step and M step until convergence to the likelihood function Reached maximum value.

[0132] (5.2) To determine the size of the number of sub-distributions H, Gaussian mixture models are constructed for different H values, and the formula BIC = m ln (N S )-2ln(L) to calculate its Bayesian information content (BIC), where m is the number of model parameters, N S is the number of spatial grid instances, and L is the likelihood function. The H value corresponding to the minimum BIC is selected as the number of sub-distributions. The final number of sub-distributions selected is shown in Table 1.

[0133] Table 1

[0134]

[0135] (5.3) Then, define the hotspot distribution is the sub-distribution with the highest probability average in the spatial grid unit, that is: Where 1{·} is the indicator function, and the numerator represents the subdistribution N h The accumulated value of the instance probability, the denominator represents the probability of belonging to the sub-distribution N h The number of instances of .

[0136] (5.4) Finally, the hotspot area s hot For The area range contained in the grid instance is: like Figure 3 As shown, (a) represents the crime space hotspot of injury crime, (b) represents the crime space hotspot of burglary crime, (c) represents the crime space hotspot of electric vehicle theft crime, and (d) represents the crime space hotspot of online fraud crime.

[0137] At the same time, the present invention can also be compared with the existing traditional crime hotspot detection method based on the data completeness assumption to verify the effectiveness of the present invention. The comparison results are as follows: Figure 4As shown, (a) is a comparison of the crime space hot spots of injury, (b) is a comparison of the crime space hot spots of burglary, (c) is a comparison of the crime space hot spots of electric vehicle theft, and (d) is a comparison of the crime space hot spots of online fraud. Figure 3 and Figure 4 It can be found that, in general, the spatial hotspots of the six types of police incidents are concentrated in the central area of ​​the study area, that is, near the town government.

[0138] The spatial hotspots detected by this method are as follows: ① For injury and electric vehicle theft incidents, their spatial hotspots have similar and stable distribution trends, all concentrated near the most prosperous Junyue Commercial Plaza in the west of the town government. ② The hotspots of burglary incidents are mainly concentrated in the urban villages in the east, south and southwest of the town government. ③ The hotspots of online fraud incidents include the south of the town government, the urban villages in the southwest, and the commercial plazas and office buildings in the southeast.

[0139] Comparative analysis of hotspot detection results between the method of the present invention and the traditional method: Research related to the number of crimes indicates that the victim or law enforcement department will choose whether to report or file a case depending on the actual harm caused by the crime. In this case, for the two types of incidents of injury and electric vehicle theft, the losses in terms of life and property are relatively small, so there may be more cases of not reporting or not filing a case, that is, the uncertainty of the crime information is high, resulting in obvious differences in the scope and location of the hotspots detected by the method of this paper and the comparative method. Specifically, the hotspot range is more accurately located from the general commercial plaza next to the town government to the most prosperous and more crowded Grand Hyatt Plaza area in the west of the town government. This result is more in line with the theory of crime attraction. The economic losses caused by the two types of incidents of burglary and online fraud are relatively large, and the degree of uncertainty is relatively low. Therefore, the hotspots detected by the method of this paper are different from the comparative method only in terms of coverage, which is specifically reflected in: the southern and southwestern areas of the town government, which are also urban villages, are added, which is in line with the daily activities theory in criminology.

[0140] The spatial hotspot detection method proposed in the present invention is highly robust. When faced with criminal incidents with different uncertainty intensities, it can detect reasonable hotspot results and discover new laws that are stable and consistent with criminological theories. It helps to understand the deeper internal mechanism between criminal behavior and urban environment, and helps the police to adjust crime prevention and control deployment strategies in a targeted manner.

[0141] The crime hotspot detection method provided in this embodiment learns the correlation between crime events and the urban built environment through a geographically weighted regression model, corrects the prior distribution of crime events, thereby eliminating the influence of spatial uncertainty, obtains a complete distribution of crimes, and mines credible crime spatial hotspots through the EM algorithm, thereby improving the efficiency, accuracy and adaptability of crime event detection.

[0142] Corresponding to the above method embodiment, see Figure 5 The embodiment of the present disclosure also provides a crime event hotspot detection system 50, including:

[0143] A removal module 501 is used to remove cases without coordinate information in the case records, and obtain the coordinates of the crime record point data and POI data;

[0144] A division module 502, for dividing the spatial research area into a plurality of grids and defining the coordinates of each of the grids;

[0145] A generating module 503, for generating crime prior distribution and POI distribution according to the coordinates of the crime record point data and the POI data;

[0146] A construction module 504 is used to construct a geographically weighted regression model based on the co-occurrence relationship between crime events and the built environment, according to the crime prior distribution and the POI distribution, calculate the conditional probability of the crime record point location, and correct the crime prior distribution in combination with the coordinates of each grid and the geographically weighted regression model to obtain a complete crime distribution;

[0147] The prediction module 505 is used to mine sub-distributions representing frequent occurrences of criminal incidents from the complete crime distribution as hot spots.

[0148] Figure 5 The system shown can correspondingly execute the contents of the above method embodiment. For the parts not described in detail in this embodiment, refer to the contents recorded in the above method embodiment and will not be repeated here.

[0149] See also Figure 6 The embodiment of the present disclosure further provides an electronic device 60, which includes: at least one processor and a memory connected to the at least one processor. The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the crime event hotspot detection method in the aforementioned method embodiment.

[0150] The embodiment of the present disclosure also provides a non-transitory computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable the computer to execute the crime incident hotspot detection method in the aforementioned method embodiment.

[0151] The embodiments of the present disclosure also provide a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions, which, when executed by a computer, enable the computer to execute the crime event hotspot detection method in the aforementioned method embodiment.

[0152] Reference below Figure 6 , which shows a schematic diagram of the structure of an electronic device 60 suitable for implementing the embodiment of the present disclosure. The electronic device in the embodiment of the present disclosure may include, but is not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle terminals (such as vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 6 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0153] like Figure 6 As shown, the electronic device 60 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 60 are also stored. The processing device 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0154] Typically, the following devices may be connected to the I / O interface 605: input devices 606 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 608 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 609. The communication device 609 may allow the electronic device 60 to communicate with other devices wirelessly or by wire to exchange data. Although the electronic device 60 with various devices is shown in the figure, it should be understood that it is not required to implement or have all the devices shown. More or fewer devices may be implemented or have alternatively.

[0155] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device 609, or installed from a storage device 608, or installed from a ROM 602. When the computer program is executed by the processing device 601, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0156] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0157] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0158] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device can execute the relevant steps of the method embodiment.

[0159] Alternatively, the computer-readable medium carries one or more programs, and when the one or more programs are executed by the electronic device, the electronic device can perform the relevant steps of the method embodiment.

[0160] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0161] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0162] The units involved in the embodiments described in the present disclosure may be implemented by software or by hardware.

[0163] It should be understood that various parts of the present disclosure may be implemented in hardware, software, firmware, or a combination thereof.

[0164] The above is only a specific implementation of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present disclosure should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.

Claims

1. A method for detecting hot spots of criminal incidents, It is characterized in that include: Step 1, remove cases without coordinate information in the case records, and obtain the coordinates of the crime record point data and POI data; Step 2, dividing the spatial research area into a plurality of grids, and defining the coordinates of each of the grids; Step 3, generating crime prior distribution and POI distribution according to the coordinates of the crime record point data and the POI data; The step 3 specifically includes: Step 3.1, by formula Calculate the crime prior distribution f i (p) (s), where K(·) is the adaptive kernel function, dist(·,·) represents the Euclidean distance, Representative C i The total number of crimes of this type, Represents C i The instance of the jth crime event under the class crime type, Represents its geographical location coordinates; Step 3.2, the research results on the correlation between different crime types and POI data, select and obtain the common N related to each type of crime p POI-like data; Step 3.3, using the formula Calculate the POI distribution f (pk) (s), where pk,k∈[1,N p ] represents the POI type, represents the total number of POI instances of the pk-th type of POI, Represents the lth POI instance under this POI type, Represents its geographical location coordinates; Step 4, based on the co-occurrence relationship between criminal events and the built environment, a geographically weighted regression model is constructed according to the crime prior distribution and the POI distribution, the conditional probability of the crime record point location is calculated, and the crime prior distribution is corrected by combining the coordinates of each grid and the geographically weighted regression model to obtain a complete crime distribution; The step of calculating the conditional probability of the location of the crime record point and correcting the crime prior distribution in combination with the coordinates of each of the grids and the geographically weighted regression model to obtain a complete crime distribution includes: Step 4.1, for crime type C i ,i∈[1,N C ], for each crime record point Take the crime prior distribution f i (p) (s) and POI distribution f (pk) The probability values ​​of the corresponding positions in (s) are filtered and eliminated through ordinary least squares and student-t test. Step 4.2: For each type of crime event C i , the probability value of the corresponding event instance location in the filtered POI distribution As an explanatory variable, the probability value of the corresponding position in the crime prior distribution As the dependent variable, the conditional probability is calculated by the geographically weighted regression model Step 4.3: Based on the conditional probability and the geographically weighted regression model, predict the complete distribution of crime in the entire study area. i (s), where the complete crime distribution f i The expression of (s) is Among them, s∈S refers to the two-dimensional geographic coordinates of each grid in the geographic space area, and f (pk) (s) represents the probability value corresponding to the POI distribution at position s, represents the geographical weighting parameter at position s; Step 5: mining sub-distributions representing frequent crime incidents from the complete crime distribution as hot spots.

2. The method according to claim 1, Features , the step 4.1 specifically includes: Step 4.1.1, for each crime record instance point The prior probability value of crime and POI distribution probability value Building an OLS model in is the OLS parameter, is the residual, and the OLS parameters are solved by minimizing the residual sum of squares; Step 4.1.2, for each regression parameter b k ,k∈[1,N p ]Construct T that follows t distribution k Statistics; Step 4.1.3, according to the formula k∈[1,N p ]Calculate the t value; Step 4.1.4, set the significance level threshold to α. If the probability P{|T k |>t} is greater than the significance level α, which means that POI type pk has no significant effect on the occurrence of crime incidents and it is eliminated from the next step of conditional probability calculation.

3. The method according to claim 1, Features , the conditional probability The expression is in, k∈[1,N p ], is the geographical weighted parameter, which is solved by weighted least squares method The spatial weight function is selected as a Gaussian function, and the bandwidth is determined by the minimum AIC information criterion.

4. The method according to claim 1, Features , the step 5 specifically includes: Step 5.1: Given the number of sub-distributions H, define the parameter solution process as the optimization solution process Then the optimization problem is solved by EM algorithm; Step 5.2: construct Gaussian mixture models for different H values, and use the formula BIC = mln (N S )-2ln(L) to calculate its Bayesian information, where m is the number of model parameters, N S is the number of spatial grid instances, L is the likelihood function, and the H value corresponding to the minimum BIC is selected as the number of sub-distributions; Step 5.3: define the sub-distribution with the highest probability average in the spatial grid cell as the hotspot sub-distribution Step 5.4, The area range contained in the grid instance is used as the hot spot area.

5. The method according to claim 4, Features ,The EM algorithm consists of E step and M step, and step 5.1 specifically includes: Step 5.1.1, step E approximates the lower bound of the random function by solving the expected value: Step 5.1.2, step M uses maximum likelihood estimation to estimate the parameter p h ,μ h ,σ h Make an update Step 5.1.3, update the parameters through the iterative cycle of E step and M step until convergence to the likelihood function Reached maximum value.

6. A crime hotspot detection system, It is characterized in that include: The elimination module is used to eliminate cases without coordinate information in the case records and obtain the coordinates of the crime record point data and POI data; A partitioning module, used for dividing the spatial research area into a plurality of grids and defining the coordinates of each of the grids; A generation module, for generating crime prior distribution and POI distribution according to the coordinates of the crime record point data and the POI data; The specific process of the generation module includes: Step 3.1, by formula Calculate the crime prior distribution f i (p) (s), where K(·) is the adaptive kernel function, dist(·,·) represents the Euclidean distance, Representative C i The total number of crimes of this type, Represents C i The instance of the jth crime event under the class crime type, Represents its geographical location coordinates; Step 3.2, the research results on the correlation between different crime types and POI data, select and obtain the common N related to each type of crime p POI-like data; Step 3.3, using the formula Calculate the POI distribution f (pk) (s), where pk,k∈[1,N p ] represents the POI type, represents the total number of POI instances of the pk-th type of POI, Represents the lth POI instance under this POI type, Represents its geographical location coordinates; A construction module is used to construct a geographically weighted regression model based on the co-occurrence relationship between crime events and the built environment, according to the crime prior distribution and the POI distribution, calculate the conditional probability of the crime record point location, and correct the crime prior distribution in combination with the coordinates of each grid and the geographically weighted regression model to obtain a complete crime distribution; The specific process of building the module includes: Step 4.1, for crime type C i ,i∈[1,N C ], for each crime record point Take the crime prior distribution f i (p) (s) and POI distribution f (pk) The probability values ​​of the corresponding positions in (s) are filtered and eliminated through ordinary least squares and student-t test. Step 4.2: For each type of crime event C i , the probability value of the corresponding event instance location in the filtered POI distribution As an explanatory variable, the probability value of the corresponding position in the crime prior distribution As the dependent variable, the conditional probability is calculated by the geographically weighted regression model Step 4.3: Based on the conditional probability and the geographically weighted regression model, predict the complete distribution of crime in the entire study area. i (s), where the complete crime distribution f i The expression of (s) is Among them, s∈S refers to the two-dimensional geographic coordinates of each grid in the geographic space area, and f (pk) (s) represents the probability value corresponding to the POI distribution at position s, represents the geographical weighting parameter at position s; The prediction module is used to mine sub-distributions representing frequent occurrences of criminal incidents from the complete crime distribution as hot spots.

7. An electronic device, It is characterized in that The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the crime event hotspot detection method described in any one of claims 1 to 5.