Method, device, electronic device and storage medium for identifying abnormal time and space points

Abstract

The present invention discloses a method for identifying abnormal spatiotemporal points, comprising the following steps: receiving a monitoring request to select a spatiotemporal region to be monitored; generating a spatiotemporal grid corresponding to the spatiotemporal region; traversing the grid nodes of the spatiotemporal grid in a preset spatiotemporal kernel density estimation model using a grid search method, and outputting an estimated value; the grid nodes include paid order spatiotemporal data and unpaid order spatiotemporal data; the paid order spatiotemporal data and the unpaid order spatiotemporal data include latitude and longitude data and time data; and comparing the estimated value with a preset abnormal spatiotemporal threshold to identify abnormal spatiotemporal points. The present invention introduces time data, establishes constraints on the proximity of the time dimension, and then establishes a spatiotemporal kernel density estimation model, effectively identifying abnormal spatiotemporal points that are clustered in a specific time period and specific time region, thereby addressing the shortcomings of the spatial and temporal fragmentation analysis method and effectively identifying passengers' behavior of evading orders.

Description

Technical Field

[0001] The present invention belongs to the technical field of big data processing, and specifically relates to a method, device, electronic device and storage medium for identifying abnormal time and space points. Background Art

[0002] In the ride-hailing sector, the typical business model is: a passenger requests a ride, the driver receives the request and generates an online ride order. When the ride-hailing vehicle delivers the passenger to the order destination, the passenger pays the platform for the order, which then remits the fare to the driver. This order-now-pay-later model allows many passengers to exploit loopholes and avoid paying. The platform typically advances the fare to the driver and then reminds passengers to pay promptly via text messages, phone calls, and app notifications. However, a significant number of unpaid orders remain.

[0003] In existing related technologies, the accounts of passengers who run away without paying are usually monitored. However, passengers who run away without paying usually only use the ride-hailing platform's service once and discard the account after use. Since passengers can successfully register on the online ride-hailing platform only through their mobile phone number, and the platform fails to obtain the passengers' personal information, it is difficult to effectively monitor the passengers' behavior of running away without paying. Summary of the Invention

[0004] The purpose of the present invention is to solve the above-mentioned technical problems and provide a method, device, electronic device and storage medium for identifying abnormal time and space points.

[0005] In order to solve the above problems, the present invention is implemented according to the following technical solutions:

[0006] In a first aspect, the present invention provides a method for identifying abnormal spatiotemporal points, comprising the following steps:

[0007] Receive a monitoring request and select a monitoring time and space area;

[0008] Generating a space-time grid corresponding to the space-time region;

[0009] Traversing the grid nodes of the spatiotemporal grid in a preset spatiotemporal kernel density estimation model using a grid search method and outputting an estimated value; the preset spatiotemporal kernel density estimation model includes a paid order kernel density estimation model and an unpaid order kernel density estimation model; the grid nodes include paid order spatiotemporal data and unpaid spatiotemporal data; the paid order spatiotemporal data and unpaid spatiotemporal data include latitude and longitude data and time data;

[0010] The estimated value is compared with a preset abnormal time-space threshold to identify abnormal time-space points.

[0011] In combination with the first aspect, the present invention further provides a first implementation manner of the first aspect, wherein after generating the space-time grid corresponding to the space-time region, the method further includes:

[0012] A resolution selection command is received, and the space-time grid of the space-time region is re-divided according to the resolution selection command.

[0013] In combination with the first aspect, the present invention further provides a second implementation of the first aspect, wherein comparing the estimated value with a preset abnormal spatiotemporal threshold to identify abnormal spatiotemporal points is specifically:

[0014] The estimated value of the kernel density estimation model for paid orders and the estimated value of the kernel density estimation model for unpaid orders are compared through a preset formula, and a real value with monotonic meaning is output; if the real value is greater than the preset abnormal spatiotemporal threshold, the grid node is an abnormal spatiotemporal point.

[0015] In combination with the first aspect, the present invention further provides a third implementation of the first aspect, wherein the expression of the preset formula is:

[0016]

[0017] Where K(x,y,t) is a real value, KDE up is the estimated value of the kernel density estimation model for unpaid orders, KDE p is the estimated value of the kernel density estimation model for paid orders, KDE up The input is KDE up (x,y,t), KDE p The input is KDE p (x,y,t).

[0018] In combination with the first aspect, the present invention further provides a fourth implementation of the first aspect, wherein the preset spatiotemporal kernel density estimation model is obtained by the following method, comprising:

[0019] Obtain the spatiotemporal data of paid or unpaid orders from the passenger side and build a spatiotemporal dataset;

[0020] By presetting the distribution function, spatial bandwidth and temporal bandwidth, kernel density estimation is performed on each paid order spatiotemporal data or unpaid spatiotemporal data in the spatiotemporal dataset to generate a kernel density estimation function corresponding to the spatiotemporal dataset.

[0021] In combination with the first aspect, the present invention further provides a fifth implementation of the first aspect, wherein the preset distribution function is a Gaussian kernel function, and the expression of the kernel density estimation function is:

[0022]

[0023] Where KDE(x, y, t) is the kernel density estimate, n is the number of paid or unpaid order spatiotemporal data in the spatiotemporal dataset, h1 is the spatial bandwidth, h2 is the temporal bandwidth, (x i ,y i ) is the longitude and latitude data of the i-th paid order spatiotemporal data or the unpaid spatiotemporal data, t i is the time data in the spatiotemporal data of the i-th paid order or the spatiotemporal data of the unpaid order.

[0024] In a second aspect, the present invention provides a device for identifying abnormal spatiotemporal points, comprising:

[0025] A spatiotemporal region selection module, configured to receive a monitoring request and select a spatiotemporal region to be monitored;

[0026] A space-time grid generation module, configured to generate a space-time grid corresponding to the space-time region;

[0027] A traversal module, the traversal module being configured to traverse the grid nodes of the spatiotemporal grid in a preset spatiotemporal kernel density estimation model using a grid search method and output an estimated value; the preset spatiotemporal kernel density estimation model comprising a paid order kernel density estimation model and an unpaid order kernel density estimation model; the grid nodes comprising paid order spatiotemporal data and unpaid spatiotemporal data; the paid order spatiotemporal data and unpaid spatiotemporal data comprising latitude and longitude data and time data;

[0028] An identification module is used to compare the estimated value with a preset abnormal time-space threshold and identify abnormal time-space points.

[0029] In conjunction with the second aspect, the present invention further provides a first implementation of the second aspect, further comprising:

[0030] The spatiotemporal grid re-dividing module is used to receive a resolution selection command and re-divide the spatiotemporal grid of the spatiotemporal area according to the resolution selection command.

[0031] In a third aspect, the present invention provides an electronic device comprising 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. When the at least one processor executes the instructions, it specifically performs a method for identifying abnormal space-time points as described in any one of the first aspects.

[0032] In a fourth aspect, the present invention provides a storage medium storing a computer program, characterized in that when the computer program is executed by a processor, it specifically performs a method for identifying abnormal space-time points as described in any one of the first aspects.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] The kernel density estimation model for paid orders outputs the estimated value of paid orders, which reflects the probability of occurrence of paid orders at the grid node. The kernel density estimation model for unpaid orders outputs the estimated value of unpaid orders, which reflects the probability of occurrence of unpaid orders at the grid node. By introducing time data and establishing the proximity constraint of the time dimension, a spatiotemporal kernel density estimation model is established to calculate the spatiotemporal distribution density of the behaviors of unpaid and paid orders, and then analyze the frequency of occurrence of unpaid orders and paid orders, effectively identifying abnormal spatiotemporal points that are clustered in specific time periods and specific time areas, making up for the defects of the spatial and temporal separation analysis method, and effectively identifying the behavior of passengers escaping orders. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings, wherein:

[0036] Figure 1 It is a flow chart of a method for identifying abnormal spatiotemporal points of the present invention;

[0037] Figure 2 It is a structural schematic diagram of an abnormal space-time point identification device of the present invention. DETAILED DESCRIPTION

[0038] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0039] The preferred embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although preferred embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to make the present disclosure more thorough and complete, and to fully convey the scope of the present disclosure to those skilled in the art.

[0040] As used herein, the term "including" and its variations represent open inclusion, i.e., "including but not limited to." Unless otherwise stated, the term "or" means "and / or." The term "based on" means "based at least in part on." The terms "an example embodiment" and "an embodiment" mean "at least one example embodiment." The term "another embodiment" means "at least one additional embodiment." The terms "first," "second," etc. may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0041] The typical operating model in the ride-hailing sector is: a passenger requests a ride, the driver receives the request and generates an online ride order. When the ride-hailing vehicle delivers the passenger to the order destination, the passenger pays the platform for the order, which then remits the fare to the driver. This order-now-pay-later model has led many passengers to exploit loopholes and avoid paying. The platform typically advances the fare to the driver and then reminds passengers to pay promptly via text messages, phone calls, and app notifications. However, a significant number of unpaid orders remain.

[0042] In the existing technology, the accounts of passengers who run away without paying are usually monitored. However, passengers who run away without paying usually only use the ride-hailing platform's service once and then discard the account. Since passengers can successfully register on the online ride-hailing platform only through their mobile phone number, the platform fails to obtain the passengers' personal information, making it difficult to effectively monitor passengers' behavior of running away without paying.

[0043] In related technologies, Uber provides a hexagonal hierarchical grid system H3. H3 is a grid-based spatial index that uses hexagons as the basic unit of grid indexing. It covers the entire earth with hexagonal grids, encodes longitude and latitude according to different hexagonal areas, and converts them into codes with different digits according to different grid precisions. The same code is used in the same area. By collecting the codes containing longitude and latitude of the grid where the unpaid orders are located, statistical analysis is performed on these data to identify the time and space points of the unpaid orders. However, the width of each hexagonal grid is pre-set. When the grid size needs to be adjusted, the entire grid index system needs to be rebuilt. Moreover, when the width of the hexagonal grid is too large, the accuracy is low, and it is difficult to accurately identify the longitude and latitude characteristics of the unpaid orders. If the width of the hexagonal grid is too small, discontinuous blank areas are easily generated between adjacent grids, and unpaid orders in the blank areas are easily missed.

[0044] Example 1

[0045] like Figure 1 As shown, in a first aspect, the present invention provides a method for identifying abnormal spatiotemporal points, comprising the following steps:

[0046] Receive a monitoring request and select a monitoring time and space area;

[0047] Generating a space-time grid corresponding to the space-time region;

[0048] Traversing the grid nodes of the spatiotemporal grid in a preset spatiotemporal kernel density estimation model using a grid search method and outputting an estimated value; the preset spatiotemporal kernel density estimation model includes a paid order kernel density estimation model and an unpaid order kernel density estimation model; the grid nodes include paid order spatiotemporal data and unpaid spatiotemporal data; the paid order spatiotemporal data and unpaid spatiotemporal data include latitude and longitude data and time data;

[0049] The estimated value is compared with a preset abnormal time-space threshold to identify abnormal time-space points.

[0050] In practice, it is found that most passengers who run away without paying are repeat offenders. They have repeatedly exploited loopholes in this payment model to run away without paying. The places where these passengers run away without paying usually occur in specific time periods and specific areas, and are clustered to a certain extent in time and space. The platform can identify these areas with clustered unpaid orders and unpaid time periods, and mark these areas as abnormal time and space points. By analyzing these abnormal time and space points, it can effectively identify passengers who run away without paying.

[0051] In this embodiment, the kernel density estimation model for paid orders outputs an estimated value of paid orders, which reflects the probability of occurrence of paid orders at the grid node. The kernel density estimation model for unpaid orders outputs an estimated value of unpaid orders, which reflects the probability of occurrence of unpaid orders at the grid node. By introducing time data and establishing proximity constraints in the time dimension, a spatiotemporal kernel density estimation model is established to calculate the spatiotemporal distribution density of unpaid and paid order behaviors, thereby analyzing the frequency of occurrence of unpaid orders and paid orders, effectively identifying abnormal spatiotemporal points that are clustered in specific time periods and specific time regions, and making up for the defects of the spatial and temporal split analysis methods, and effectively identifying the behavior of passengers escaping orders.

[0052] Step 1: Receive a monitoring request and select the spatiotemporal area to be monitored.

[0053] Specifically, before selecting the space-time area, a spatial kernel density estimation model is established to analyze and find spatial areas that are in a clustered state, and a data clock is established to find the clustered time period. Monitoring requests are received, and time areas with clustered spatial areas and clustered time periods are selected as monitoring areas. The specific latitude and longitude areas of these areas and the specific time period with the most unpaid orders are identified to effectively identify the behavior of running away from the order.

[0054] Step 2: Generate a space-time grid corresponding to the space-time region.

[0055] Specifically, squares are used as the basic units of grid indexes, and the entire space-time area is covered with square grids. The longitude and latitude are encoded according to different square areas, and converted into codes with different digits according to different grid precisions. The codes are the same in the same area. By collecting the codes containing the longitude and latitude of the grids where unpaid orders and paid orders are located, if the first five digits of the two coordinates are the same, it indicates that the two coordinates are located in the same space-time grid in the fifth-level precision. If the first five digits of the two coordinates are the same, but the sixth digits are different, it indicates that the two coordinates are located in the same space-time grid in the fifth-level precision, but in different space-time grids in the sixth-level precision.

[0056] In another embodiment, equilateral triangles or hexagons are used as the basic unit of grid indexing, and the entire spatiotemporal region is covered with equilateral triangle or hexagonal grids. Because hexagonal grids have a relatively low perimeter, sample bias caused by boundary effects of the grid shape can be reduced. Furthermore, the distance between the centroids of hexagonal grids and the surrounding grids is equal. When the hexagonal grids cover the entire spatiotemporal region, there are no gaps. This makes it easier and faster to find all adjacent spatiotemporal grids when searching for spatiotemporal grids in adjacent areas.

[0057] Step 3: Traverse the grid nodes of the spatiotemporal grid in a preset spatiotemporal kernel density estimation model through a grid search method and output an estimated value; the preset spatiotemporal kernel density estimation model includes a paid order kernel density estimation model and an unpaid order kernel density estimation model; the grid nodes include paid order spatiotemporal data and unpaid spatiotemporal data; the paid order spatiotemporal data and unpaid spatiotemporal data include latitude and longitude data and time data.

[0058] Specifically, the grid search method trains the unpaid data and paid order data in each grid node through the spatiotemporal kernel density estimation model according to the coding order, wherein the unpaid order spatiotemporal data is trained through the unpaid order kernel density estimation model to obtain the kernel density estimation value KDE of the unpaid order of the grid node up The paid order spatiotemporal data is trained through the paid order kernel density estimation model to obtain the kernel density estimation value KDE of the paid order of the grid node p ,The grid search method can avoid missing the difficult-to-detect and relatively hidden grid nodes in the ,space-time region and avoid missing the identification of abnormal space-time points.

[0059] Specifically, the preset spatiotemporal kernel density estimation model is obtained by the following method, including:

[0060] Step 301: Obtain the spatiotemporal data of paid or unpaid orders on the passenger side and establish a spatiotemporal dataset;

[0061] Step 302: Perform kernel density estimation on each paid order spatiotemporal data or unpaid spatiotemporal data in the spatiotemporal dataset using a preset distribution function, spatial bandwidth, and temporal bandwidth to generate a kernel density estimation function corresponding to the spatiotemporal dataset.

[0062] The preset distribution function is a Gaussian kernel function, and the expression of the kernel density estimation function is:

[0063]

[0064] Where KDE(x, y, t) is the kernel density estimate, n is the number of paid or unpaid order spatiotemporal data in the spatiotemporal dataset, h1 is the spatial bandwidth, h1∈(0.05, 2), h2 is the temporal bandwidth, h2∈(0.05, 2), (x i ,y i ) is the longitude and latitude data of the i-th paid order spatiotemporal data or the unpaid spatiotemporal data, t i is the time data in the spatiotemporal data of the i-th paid order or the spatiotemporal data of the unpaid order.

[0065] Specifically, since discretely distributed latitude and longitude data are difficult to capture relatively subtle continuous changes, the preset distribution function uses a Gaussian kernel function. The Gaussian kernel function can cross the similarity between different longitude and latitude data, allowing the longitude and latitude data to be better clustered together within a certain area, making these starting point location data linearly separable, thereby improving the accuracy of the kernel density estimation function. At the same time, based on the longitude and latitude data, by introducing time data and establishing constraints on the proximity of the time dimension, a spatiotemporal kernel density estimation model is established to calculate the spatiotemporal distribution density of unpaid and paid orders, and then analyze the frequency of unpaid orders and the frequency of paid orders, effectively identifying abnormal spatiotemporal points that are clustered in specific time periods and specific time regions, making up for the shortcomings of the spatial and temporal separation analysis method, and effectively identifying passengers' behavior of escaping orders.

[0066] The spatial and temporal bandwidths are adaptively adjusted. When they are large, the kernel function curve is smoother, contains less detail, and results in larger errors. When they are small, the kernel function curve is bumpier and steeper, contains more noise, and is not conducive to analyzing and finding anomalous spatiotemporal points. The spatial bandwidth can be adjusted based on the number of unpaid and paid orders within the spatiotemporal grid, making it easier to clearly identify whether a spatiotemporal node is an anomalous point.

[0067] Step 4: Compare the estimated value with the preset abnormal time-space threshold to identify abnormal time-space points.

[0068] Specifically, the estimated value of the kernel density estimation model for paid orders and the estimated value of the kernel density estimation model for unpaid orders are compared through a preset formula, and a real value with monotonic meaning is output; if the real value is greater than the preset abnormal spatiotemporal threshold, the grid node is an abnormal spatiotemporal point.

[0069] The expression of the preset formula is:

[0070]

[0071] Where K(x,y,t) is a real value, KDE up is the estimated value of the kernel density estimation model for unpaid orders, KDE p is the estimated value of the kernel density estimation model for paid orders, KDE up The input is KDE up (x,y,t), KDE p The input is KDE p (x,y,t).

[0072] For example, K(x,y,t) indicates the difference between the number of unpaid orders and paid orders at this grid node. The larger the value of K(x,y,t), the higher the proportion of unpaid orders at this grid node. This grid node is very likely to be an abnormal time and space point, and it is necessary to focus on monitoring and prevent orders within this time period and at the longitude and latitude of this location.

[0073] Preferably, after generating the space-time grid corresponding to the space-time region, the method further includes:

[0074] Step 5: Receive a resolution selection command, and re-divide the spatiotemporal grid of the spatiotemporal region according to the resolution selection command.

[0075] If the initial space-time grid is of low precision, meaning there are fewer space-time grids within the space-time region and each grid covers a larger area, it can be difficult to accurately locate anomalous space-time points at specific latitudes, longitudes, and times. If the initial space-time grid is of high precision, meaning there are more space-time grids within the space-time region and each grid covers a smaller area, analysis requires more data and maintenance time. In this case, resetting the resolution and selecting an appropriate number of space-time grids can facilitate quick and accurate locating of anomalous space-time points within the space-time grid.

[0076] In summary, when the method of the present invention is executed, on the one hand, the present invention, based on the longitude and latitude data, introduces time data to establish the constraints of the proximity of the time dimension, and then establishes a spatiotemporal kernel density estimation model to calculate the spatiotemporal distribution density of the behaviors of unpaid orders and paid orders, and then analyzes the frequency of occurrence of unpaid orders and the frequency of occurrence of paid orders, effectively identifying abnormal spatiotemporal points that are clustered in a specific time period and a specific time area, making up for the defects of the spatial and temporal split analysis method, and effectively identifying the passenger's behavior of running away from the order. On the other hand, by resetting the resolution and selecting the appropriate number of spatiotemporal grids, it is convenient to quickly and accurately find abnormal spatiotemporal points in the spatiotemporal grid. In addition, the grid search method can avoid missing the grid nodes that are difficult to detect and relatively hidden in the spatiotemporal area, and avoid missing the identification of abnormal spatiotemporal points.

[0077] The other steps of the method for identifying abnormal time-space points of the present invention refer to the prior art.

[0078] Example 2

[0079] like Figure 2 As shown, in a second aspect, the present invention discloses a device for identifying abnormal spatiotemporal points, including a spatiotemporal region selection module M1, a spatiotemporal grid generation module M2, a traversal module M3 and an identification module M4.

[0080] The spatiotemporal region selection module M1 is used to receive a monitoring request and select a spatiotemporal region to be monitored;

[0081] The space-time grid generation module M2 is used to generate a space-time grid corresponding to the space-time region;

[0082] The traversal module M3 is used to traverse the grid nodes of the spatiotemporal grid in a preset spatiotemporal kernel density estimation model through a grid search method and output an estimated value; the preset spatiotemporal kernel density estimation model includes a paid order kernel density estimation model and an unpaid order kernel density estimation model; the grid nodes include paid order spatiotemporal data and unpaid spatiotemporal data; the paid order spatiotemporal data and unpaid spatiotemporal data include latitude and longitude data and time data;

[0083] The identification module M4 is used to compare the estimated value with a preset abnormal time-space threshold to identify abnormal time-space points.

[0084] With respect to the second aspect, the first preferred implementation is also included, which further includes a module M5 for re-dividing the spatiotemporal grid.

[0085] The spatiotemporal grid re-dividing module M5 is used to receive a resolution selection command and re-divide the spatiotemporal grid of the spatiotemporal region according to the resolution selection command.

[0086] In summary, when the device described in this embodiment is in operation, it can implement all the steps of the method for identifying abnormal time-space points described in Example 1 to achieve the technical effect achieved in Example 1.

[0087] For other structures of the device for identifying abnormal time-space points described in this embodiment, refer to the prior art.

[0088] Example 3

[0089] The present invention further discloses an electronic device, comprising 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. When the at least one processor executes the instructions, the following steps are specifically implemented:

[0090] Receive a monitoring request and select a monitoring time and space area;

[0091] Generating a space-time grid corresponding to the space-time region;

[0092] Traversing the grid nodes of the spatiotemporal grid in a preset spatiotemporal kernel density estimation model using a grid search method and outputting an estimated value; the preset spatiotemporal kernel density estimation model includes a paid order kernel density estimation model and an unpaid order kernel density estimation model; the grid nodes include paid order spatiotemporal data and unpaid spatiotemporal data; the paid order spatiotemporal data and unpaid spatiotemporal data include latitude and longitude data and time data;

[0093] The estimated value is compared with a preset abnormal time-space threshold to identify abnormal time-space points.

[0094] Example 4

[0095] The present invention further discloses a storage medium storing a computer program. When the computer program is executed by a processor, the following steps are specifically implemented:

[0096] Receive a monitoring request and select a monitoring time and space area;

[0097] Generating a space-time grid corresponding to the space-time region;

[0098] Traversing the grid nodes of the spatiotemporal grid in a preset spatiotemporal kernel density estimation model using a grid search method and outputting an estimated value; the preset spatiotemporal kernel density estimation model includes a paid order kernel density estimation model and an unpaid order kernel density estimation model; the grid nodes include paid order spatiotemporal data and unpaid spatiotemporal data; the paid order spatiotemporal data and unpaid spatiotemporal data include latitude and longitude data and time data;

[0099] The estimated value is compared with a preset abnormal time-space threshold to identify abnormal time-space points.

[0100] The present disclosure may be a method, an apparatus, a system and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing various aspects of the present disclosure.

[0101] A computer-readable storage medium can be a tangible device that can hold and store instructions for use by an instruction execution device. A computer-readable storage medium can be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: 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), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punch card or a raised structure in a groove on which instructions are stored, and any suitable combination thereof. As used herein, a computer-readable storage medium is not to be construed as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse through a fiber optic cable), or an electrical signal transmitted through an electrical wire.

[0102] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions to be stored in the computer-readable storage medium in each computing / processing device.

[0103] The computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, Java, and conventional procedural programming languages ​​such as "C" language or similar programming languages. Computer-readable program instructions may be executed entirely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via 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., utilizing an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be personalized by utilizing the state information of the computer-readable program instructions. The electronic circuit may execute the computer-readable program instructions, thereby realizing various aspects of the present disclosure.

[0104] The embodiments of the present disclosure have been described above. The above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for identifying abnormal spatiotemporal points, characterized in that: The following steps are involved: Receive a monitoring request and select a monitoring time and space area; Generating a space-time grid corresponding to the space-time region; Traversing the grid nodes of the spatiotemporal grid in a preset spatiotemporal kernel density estimation model by a grid search method, and outputting an estimated value; The preset spatiotemporal kernel density estimation model includes a paid order kernel density estimation model and an unpaid order kernel density estimation model; the grid nodes include paid order spatiotemporal data and unpaid spatiotemporal data; the paid order spatiotemporal data and unpaid spatiotemporal data include latitude and longitude data and time data; Comparing the estimated value with a preset abnormal time-space threshold to identify abnormal time-space points; The comparison of the estimated value with a preset abnormal spatiotemporal threshold to identify abnormal spatiotemporal points is specifically as follows: The estimated value of the kernel density estimation model for paid orders and the estimated value of the kernel density estimation model for unpaid orders are compared through a preset formula, and a real value with monotonic meaning is output; if the real value is greater than the preset abnormal spatiotemporal threshold, the grid node is an abnormal spatiotemporal point.

2. The method for identifying abnormal spatiotemporal points according to claim 1, characterized in that: After generating the space-time grid corresponding to the space-time region, the method further includes: A resolution selection command is received, and the space-time grid of the space-time region is re-divided according to the resolution selection command.

3. The method for identifying abnormal spatiotemporal points according to claim 2, characterized in that: The expression of the preset formula is: Where K(x,y,t) is a real value, KDE up is the estimated value of the kernel density estimation model for unpaid orders, KDE p is the estimated value of the kernel density estimation model for paid orders, KDE up The input is KDE up (x,y,t), KDE p The input is KDE p (x,y,t).

4. The method for identifying abnormal spatiotemporal points according to claim 1, wherein: The preset spatiotemporal kernel density estimation model is obtained by the following method, including: Obtain the spatiotemporal data of paid or unpaid orders from the passenger side and build a spatiotemporal dataset; By presetting the distribution function, spatial bandwidth and temporal bandwidth, kernel density estimation is performed on each paid order spatiotemporal data or unpaid spatiotemporal data in the spatiotemporal dataset to generate a kernel density estimation function corresponding to the spatiotemporal dataset.

5. The method for identifying abnormal spatiotemporal points according to claim 4, characterized in that: The preset distribution function is a Gaussian kernel function, and the expression of the kernel density estimation function is: Where KDE(x, y, t) is the kernel density estimate, n is the number of paid or unpaid order spatiotemporal data in the spatiotemporal dataset, h1 is the spatial bandwidth, h2 is the temporal bandwidth, (x i ,y i ) is the longitude and latitude data of the i-th paid order spatiotemporal data or the unpaid spatiotemporal data, t i is the time data in the spatiotemporal data of the i-th paid order or the spatiotemporal data of the unpaid order.

6. A device for identifying abnormal time and space points, characterized in that: include: A spatiotemporal region selection module, configured to receive a monitoring request and select a spatiotemporal region to be monitored; A space-time grid generation module, configured to generate a space-time grid corresponding to the space-time region; A traversal module, configured to traverse the grid nodes of the spatiotemporal grid in a preset spatiotemporal kernel density estimation model by a grid search method and output an estimated value; The preset spatiotemporal kernel density estimation model includes a paid order kernel density estimation model and an unpaid order kernel density estimation model; the grid nodes include paid order spatiotemporal data and unpaid spatiotemporal data; the paid order spatiotemporal data and unpaid spatiotemporal data include latitude and longitude data and time data; an identification module, the identification module being configured to compare the estimated value with a preset abnormal spatiotemporal threshold and identify abnormal spatiotemporal points; The comparison of the estimated value with a preset abnormal spatiotemporal threshold to identify abnormal spatiotemporal points is specifically as follows: The estimated value of the kernel density estimation model for paid orders and the estimated value of the kernel density estimation model for unpaid orders are compared through a preset formula, and a real value with monotonic meaning is output; if the real value is greater than the preset abnormal spatiotemporal threshold, the grid node is an abnormal spatiotemporal point.

7. The device for identifying abnormal space-time points according to claim 6, characterized in that: Also includes: The spatiotemporal grid re-dividing module is used to receive a resolution selection command and re-divide the spatiotemporal grid of the spatiotemporal area according to the resolution selection command.

8. An electronic device comprising at least one processor and a memory communicatively connected to the at least one processor, wherein: 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, characterized in that when the at least one processor executes the instructions, it specifically executes a method for identifying abnormal space-time points according to any one of claims 1 to 6.

9. A storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for identifying abnormal space-time points according to any one of claims 1 to 5 is specifically implemented.

Images