An Adaptive Processing Method, Device, and Medium for Abnormal Regions of a Clustering Map

Abstract

An embodiment of the present invention discloses an adaptive processing method, device, and medium for abnormal regions of a clustering map, which relates to the technical field of data visualization. Among them, the method includes: obtaining information of each class in a plurality of data classes in the clustering map, where the information includes in-class data points, normal regions, and abnormal regions; for an abnormal class with an abnormal region, determining at least one first region that is adjacent to the target normal region of the abnormal class and belongs to other classes; determining a second region that is the closest to the abnormal region of the abnormal class among the at least one first region; using the second region to merge the data points in the abnormal region into the target normal region. This embodiment is an adaptive processing method for clustering maps applicable to various data volumes and has very high robustness.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of data visualization, and in particular, to a method, device, and medium for adaptively processing abnormal regions of a clustering map. Background Art

[0002] With the continuous improvement of computer processing power and the expansion of application fields, the demand for data visualization has also been increasing day by day. As an important branch of data visualization, clustering visualization has also developed rapidly. With the rapid development of Internet technology and the advent of the big data era, clustering visualization is facing new challenges and opportunities. The characteristics of big data such as scale, diversity, and speed have put forward new requirements for visualization technology.

[0003] In the field of clustering map visualization, each category of the clustering map is unique. When the visualization area of a certain category is not unique, it means that there are abnormal regions in this category, and there can be one or more such abnormal regions.

[0004] For the processing of abnormal regions in clustering maps, the traditional method is manual editing, which is inefficient; and for maps with different data volumes, there are different editing methods. In addition to the traditional method, some automated processing of abnormal regions can also be carried out, such as point movement, clustering algorithms, or anomaly detection, etc., all of which involve the setting of multiple thresholds, and in the case of different data volumes, the thresholds need to be modified according to the situation. Therefore, these methods are only simple automation and do not have robustness and adaptability.

[0005] In view of this, the present invention provides a method for adaptively processing abnormal regions of a clustering map applicable to various data volumes. Summary of the Invention

[0006] The embodiments of the present invention provide a method, device, and medium for adaptively processing abnormal regions of a clustering map to solve the above technical problems.

[0007] In a first aspect, the embodiments of the present invention provide a method for adaptively processing abnormal regions of a clustering map, including:

[0008] S10. Obtain the information of each class in a plurality of data classes in the clustering map, where the information includes in-class data points, normal regions, and abnormal regions;

[0009] S20. For an abnormal class with an abnormal region, determine at least one first region that is adjacent to the target normal region of the abnormal class and belongs to other classes;

[0010] S30. Determine a second region that is the closest to the abnormal region of the abnormal class among the at least one first region;

[0011] S40. Use the second region to merge the data points in the abnormal region into the target normal region.

[0012] In a second aspect, an embodiment of the present invention provides an electronic device, which includes:

[0013] One or more processors;

[0014] A memory for storing one or more programs,

[0015] When the one or more programs are executed by the one or more processors, the one or more processors implement the method for adaptively processing abnormal regions of a clustering map according to any embodiment.

[0016] In a third aspect, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the method for adaptively processing abnormal regions of a clustering map according to any embodiment.

[0017] An embodiment of the present invention provides a method for adaptively processing abnormal regions of a clustering map. Since the abnormal region and the normal region essentially belong to the same class, the abnormal region will be near the normal region. Accordingly, in this embodiment, the adjacent region of the normal region (i.e., the first region) is first obtained, then the region closest to the abnormal region in the adjacent regions is determined according to the relationship between polygons (i.e., the second region), and finally, the abnormal region is processed according to different situations with the help of the second region. This method does not require setting specific parameters and thresholds, nor is it limited to the position of points, but processes through the regional boundary of abnormal points, and can be applied to different data volumes and one or more abnormal region problems, with high robustness. At the same time, starting from the overall visualization effect, this embodiment gives priority to filling the normal region with a convex polygon. After completion, the closest distance between the point and the region is calculated. In this way, the boundary of the region will not be very abrupt, optimizing the visualization effect; using data point replacement instead of movement can minimize the change of the visualization boundary and avoid uncontrollable adjustment situations. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0019] Figure 1 It is a flowchart of a method for adaptively processing abnormal regions of a clustering map provided by an embodiment of the present invention;

[0020] Figure 2 It is a schematic diagram of an area including a relationship provided by an embodiment of the present invention;

[0021] Figure 3 It is a schematic diagram of an area with an intersection relationship provided by an embodiment of the present invention;

[0022] Figure 4 It is a schematic diagram of an area sharing a boundary line provided by an embodiment of the present invention;

[0023] Figure 5 It is a schematic diagram of the outermost area of a map provided by an embodiment of the present invention;

[0024] Figure 6 It is a schematic diagram of a convex polygon of a normal area provided by an embodiment of the present invention;

[0025] Figure 7 It is a schematic diagram of processing an abnormal area provided by an embodiment of the present invention;

[0026] Figure 8 It is another schematic diagram of processing an abnormal area provided by an embodiment of the present invention;

[0027] Figure 9 It is a comparison diagram of before and after processing an abnormal class using the method of this embodiment provided by an embodiment of the present invention;

[0028] Figure 10 It is another comparison diagram of before and after processing an abnormal class using the method of this embodiment provided by an embodiment of the present invention;

[0029] Figure 11 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. Detailed implementation manners

[0030] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without any creative effort belong to the scope protected by the present invention.

[0031] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. In addition, the terms "first", "second", "third" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance.

[0032] In the description of the present invention, it should also be noted that unless otherwise clearly defined and limited, the terms "installation", "connection", "connection" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0033] As described in the background art, a clustering map can display multiple data classes as multiple regions. Ideally, each data class corresponds to each region one by one, and the data points in each region represent one data in the corresponding class. However, in actual applications, in the clustering map directly generated by visualization software, one data class may correspond to multiple regions, including a normal region and at least one abnormal region. Among them, the definitions of the normal region and the abnormal region are as follows: If 95% or more of the data points in a data class are in one region, then this region is called the normal region of the data class; the regions where the remaining data points less than or equal to 5% are located are all abnormal regions of the data class. This embodiment will process these abnormal regions so that each class in the processed clustering map only corresponds to one normal region.

[0034] Figure 1 is a flowchart of a method for adaptively processing abnormal regions of a clustering map provided by an embodiment of the present invention. This method is applicable to the situation where there are abnormal regions in a data clustering map and is executed by an electronic device. As Figure 1 shown, this method specifically includes:

[0035] S110. Obtain the information of each class in the clustering map for multiple data classes, where the information includes in-class data points, normal regions, and abnormal regions.

[0036] Among them, the multiple data classes refer to multiple classes of data that have been pre-clustered, and the specific content of the data is not limited. Exemplarily, in the field of network security, the data can be an ASN number, and each data point in the clustering map of the data class represents an ASN number. In this step, first, obtain the data point coordinates, normal area boundary coordinates, and abnormal area boundary coordinates of each data class in the clustering map as the data source for the entire method.

[0037] In a specific embodiment, all normal area boundary coordinates can be denoted as normal_boundary, all abnormal area classes can be denoted as abnormal_boundary, and the data point coordinates of all classes can be denoted as cluster_postion. All three are stored in the form of a dictionary, where the key is the id of the class and the value is the area boundary coordinates. In normal_boundary, its value corresponds to only one normal area boundary coordinate and is a one-dimensional list; while in abnormal_boundary, its value contains the boundary coordinates of multiple abnormal areas and is a two-dimensional list; in cluster_postion, its value is the two-dimensional coordinate points of multiple points and is also a two-dimensional list.

[0038] S120. Optionally select a data class with an abnormal area, and determine at least one area that is adjacent to the normal area of the data class and belongs to other classes.

[0039] In this embodiment, the data class with an abnormal area is called an abnormal class. In this step, arbitrarily select one from all abnormal classes as the object to be processed currently, and the abnormal class will be called the current abnormal class in subsequent operations. Optionally, when using the above data body abnormal_boundary, the abnormal class ranked first in abnormal_boundary can be selected as the current abnormal class.

[0040] For the same abnormal class, although the abnormal area is separated from the normal area, it must exist near the normal area, that is, it must be in contact with a certain adjacent area of the normal area. This adjacent area may be the normal area of other classes or the abnormal area of other classes. Therefore, in this embodiment, for the current abnormal class, first determine the areas of other classes adjacent to the normal area of the abnormal class as the auxiliary areas for processing the abnormal area.

[0041] To distinguish it from the normal regions in other classes, the normal region in the current abnormal class in this embodiment is called the target normal region. In this step, all regions of other classes can be traversed, and it is sequentially determined whether there is an adjacent relationship between this region and the target normal region. Optionally, the region boundary can be converted into a Polygon object, and the intersects function can be used to determine whether two regions are adjacent. Exemplarily, for two objects polygon1 and polygon2, if the result of polygon1.intersects(polygon2) is True, it indicates that there is an adjacent relationship between the polygon1 and polygon2 regions; otherwise, there is no adjacent relationship between the polygon1 and polygon2 regions.

[0042] S130. Select an abnormal region from the abnormal class, and determine the region closest to the abnormal region among the at least one region.

[0043] The abnormal regions in the current abnormal class may be one or multiple. In this step, an abnormal region is randomly selected from them as the current object to be processed, and this abnormal region will be called the current abnormal region in subsequent operations. Optionally, in the case of using the above data body abnormal_boundary, the first abnormal region of the current abnormal class in abnormal_boundary can be selected as the current abnormal region.

[0044] For the current abnormal region, in this step, an adjacent region closest to the abnormal region is selected from the at least one adjacent region determined in S120 as the auxiliary region for processing the abnormal region. For the convenience of distinction and description, each adjacent region determined in S120 in this embodiment is called the first region; the first region determined to be the closest to the current abnormal region among all the first regions is called the second region.

[0045] In a specific embodiment, the first region containing the abnormal region can be first found as the second region. As Figure 2 shown, the inclusion relationship means that all data points of the abnormal region are inside the second region. If there is no Figure 2 such situation, then continue to find the first region that intersects the abnormal region as the second region, as Figure 3 shown. If there is no Figure 3 such situation, then continue to find the first region that has a common boundary with the abnormal region as the second region, as Figure 4 shown. There are two second regions satisfying this situation in the figure, and any one of them can be selected to participate in subsequent operations.

[0046] Optionally, the above three cases can also be judged according to the Polygon object. First, find the first region that intersects with the abnormal region and all the data points in the abnormal region belong to this intersection. This region is the second region that contains the abnormal region. If the second region that contains the abnormal region is not found, continue to find the first region that intersects with the abnormal region and there is no common boundary with the second region in the intersection. This region is the second region that intersects with the abnormal region. If the second region that intersects with the abnormal region is not found, continue to find the first region that intersects with the abnormal region, and the common boundary with the second region in this intersection is the first region. This region is the last type of second region.

[0047] It is worth mentioning that Figure 3 The intersection situation shown cannot be directly seen in the visualized clustering map because only one color is filled for the overlapping regions in the clustering map, so the intersection situation will not be seen. However, this situation can be judged from the data points and boundary information of the data class behind the clustering map. For this embodiment, there will be no region overlapping situation in the clustering map finally processed by the method of this embodiment, but region overlapping may occur in the data points and region boundaries updated after a certain loop. Therefore, this situation also needs to be considered.

[0048] Of course, there is another situation: none of the above three second regions exist. For example, when the normal region of the current abnormal class is the outermost region of the map, as Figure 5 shown. At this time, there is no any other region between the abnormal region and the target normal region, so there is no need to judge the second region closest to the abnormal region anymore. At this time, the second region is empty. Optionally, the second region can be denoted as nearest_region, which is also in the form of a dictionary. The key represents the id of the class, and the value represents the boundary of the nearest adjacent class region. When there is no second region, nearest_region is empty.

[0049] S140. Use the second region to merge the data points in the abnormal region into the target normal region.

[0050] In this step, with the help of the second region, the data points in the current abnormal region are replaced to the position adjacent to the target normal region, and together with the target normal region, they form a new normal region, thus eliminating the abnormal region. Optionally, according to whether the second region is empty, the following several optional implementation manners are provided in this embodiment:

[0051] The first optional implementation manner, if the second region is empty, randomly generate data points in the target normal region to replace the data points in the abnormal region. Optionally, the number of data points in this abnormal region can be counted, and the same number of points can be randomly generated into the target normal region according to the counted number.

[0052] The second optional implementation. If the second region is not empty, first, extract the region where the convex polygon of the second region intersects with the target normal region. The points in this region will be preferentially used to replace the points in the current target region. Optionally, a convex polygon function can be used to generate the convex polygon of the target normal region. As Figure 6 shown, the convex polygon ABCDED is the convex polygon of the target normal region. This convex polygon is not the real boundary in the clustering map and can be understood as a virtual region. The region where this virtual region intersects with the second region can fill the target normal region into a convex polygon. Optionally, this virtual convex polygon can be denoted as current_convex, and the boundary of the target normal region can be denoted as current_boundary. Count the number and coordinates of all the coordinate points in the second region nearest_region in the convex polygon current_convex. Denote the number as replace_num, and assign the set of its coordinate points to the set of data points replace_point that can be replaced with the abnormal region.

[0053] Then, select at least one data point closest to the target normal region from the intersecting region, and replace the data points in the at least one data point and the abnormal region with each other. Optionally, according to the number of data points in the intersecting region and the abnormal region, this replacement process can include the following situations:

[0054] The first situation is that the number of data points replace_num in the intersecting region is equal to the number of data points in the abnormal region. In this case, replace all the data points in the intersecting region and the abnormal region with each other. Optionally, pre-statistics the number and coordinates of the abnormal points in the abnormal region, denoted as current_num and current_point respectively. In this case, replace_num = current_num, then replace replace_point and current_point completely, and replace the coordinates of the abnormal region into the convex polygon of the normal region.

[0055] The second situation is that the number of data points replace_num in the intersecting region is greater than the number of data points current_num in the abnormal region. Then select the current_num data points closest to the target normal region, and replace all the data points in the current_num data points and the abnormal region with each other. Optionally, combined with Figure 7 , first calculate replace_point( Figure 7The point closest to the boundary current_boundary of the target normal region (shown by the red and green dots) in the intersection region is denoted as near_point; then, calculate the distances between the closest point near_point and all points in replace_point, and sort the points in the intersection region in ascending order according to these distances; take the coordinates of the first current_num points in the ordered sequence as the new replace_point; then replace the coordinates of the new replace_point with the coordinates of current_point, and replace the coordinates of the abnormal region into the convex polygon of the normal region.

[0056] In the third case, the number of data points replace_num in the intersection region is less than the number of data points current_num in the abnormal region. At this time, first replace all the data points in the intersection region with replace_num data points in the abnormal region. For the remaining un-replaced data points in the abnormal region, first determine another data point in the target normal region that is closest to the un-replaced data point, then determine another data point in the second region that is closest to the other data point, and finally replace the other data point with the un-replaced data point. For the sake of easy distinction and description, in this embodiment, the un-replaced data point is called the first data point, the other data point is called the second data point, and the other data point is called the third data point. Optionally, combined with Figure 8 , denote the number of un-replaced data points and the region formed by these data points as rest_num and rest_point respectively; calculate the coordinate point in the coordinates of rest_point that is closest to the normal boundary region current_boundary, denoted as normal_point; then, calculate the distances between normal_point and all coordinate points in the second region nearest_region, and sort these coordinate points in ascending order according to the distances, take the first rest_num data as the new replace_point, and replace the new replace_point with rest_point, so as to complete the processing of an abnormal region.

[0057] After the replacement is completed according to the above three cases, it is necessary to update the information of each region and each category in the clustering map. Optionally, the updated content includes: expanding the boundary of the target normal region and shrinking the boundary of the second region according to at least one data point closest to the target normal region; deleting the abnormal region; updating the data points of the categories where the target normal region and the second region are located. In a specific embodiment, the boundary of the abnormal region can be set to be empty, and the external polygon boundary created by the replace_point (now the point coordinates in the abnormal region) is merged into the target normal region current_boundary to update the boundary of the normal region of the current category; at the same time, the difference set of the external polygon created by the second region and the replace_point is taken, and the circumscribed polygon created by the current_point (i.e., the abnormal region before replacement) is merged into the boundary of the second region to update the boundary of the second region.

[0058] Thus, the originally existing abnormal region is integrated with the target normal region, and the processing of the current abnormal region in the current abnormal category is completed.

[0059] S150. According to the new information of each category in the clustering map after replacement, re-execute S130 - S150 until all the latest abnormal regions of the abnormal category are processed.

[0060] In this step, according to the updated data points, normal regions and abnormal regions of each category, return to S130 again, select an abnormal region from the latest abnormal regions of the current abnormal category as the new current abnormal region; and execute S140 for the new current abnormal region to complete the processing of this abnormal region; then execute S150 to select another abnormal region from the updated abnormal regions for processing,.... Repeat this process until there is no abnormal region in the current abnormal category, and at this time, this category is no longer an abnormal category.

[0061] S160. According to the information updated in the last S150, re-execute S120 - S160 until all the latest abnormal categories are processed.

[0062] In this step, according to the data points, normal regions and abnormal regions of each category updated in the last S150, return to S120 again, select an abnormal category from the latest data categories as the new current abnormal category; and execute S130 - S150 for the new current abnormal category to make the new current abnormal category become a normal category; then execute S160 again to select another abnormal category from the updated data categories for processing,.... Repeat this process until all the abnormal categories in the clustering map are processed.

[0063] Figure 9 and Figure 10 respectively show the comparison diagrams before and after processing two abnormal classes by using the method of this embodiment. The ellipses in the figures are used to mark the abnormal areas of the abnormal classes, and the rectangles are used to mark the normal areas of the abnormal classes. Among them, Figure 9 the abnormal class in Figure 10 includes one abnormal area, and the abnormal class in

[0064] includes 3 abnormal areas. After processing, the abnormal areas and normal areas in the two figures are integrated into one.

[0064] In summary, this embodiment provides a method for adaptively processing abnormal areas of a clustering map. Since the abnormal areas and normal areas essentially belong to the same class, the abnormal areas will be near the normal areas. Accordingly, this embodiment first obtains the adjacent areas of the normal areas (i.e., the first areas), then determines the area closest to the abnormal area among the adjacent areas according to the relationship between polygons (i.e., the second areas), and finally processes the abnormal areas according to different situations with the help of the second areas. This method does not require setting specific parameters and thresholds, nor is it limited to the positions of points, but processes through the regional boundaries of abnormal points, and can be applied to different data magnitudes and one or more abnormal area problems, with high robustness. At the same time, starting from the overall visualization effect, this embodiment gives priority to filling convex polygons for the normal areas, and then calculates the closest distance between the points and the areas after completion, so that the boundaries of the areas will not be very abrupt, optimizing the visualization effect; using the replacement of data points instead of moving can minimize the change of the visualization boundary and avoid uncontrollable adjustment situations.

[0065] Figure 11 is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. As Figure 11 shown, the device includes a processor 60, a memory 61, an input device 62, and an output device 63; the number of processors 60 in the device can be one or more, Figure 11 and one processor 60 is taken as an example in Figure 11 ; the processor 60, memory 61, input device 62, and output device 63 in the device can be connected by a bus or other means, Figure 11 and taking the connection by bus as an example in

[0066] The memory 61, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the method for adaptively processing abnormal areas of a clustering map in an embodiment of the present invention. The processor 60 executes various functional applications and data processing of the device by running the software programs, instructions, and modules stored in the memory 61, that is, implements the above-mentioned method for adaptively processing abnormal areas of a clustering map.

[0067] The memory 61 may mainly include a program storage area and a data storage area. Among them, the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory 61 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state storage devices. In some instances, the memory 61 may further include a memory remotely provided with respect to the processor 60, and these remote memories may be connected to the device through a network. Examples of the above-mentioned network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0068] The input device 62 can be used to receive input digital or character information, and generate key signal inputs related to user settings and function controls of the device. The output device 63 may include a display device such as a display screen.

[0069] An embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the abnormal area adaptive processing method of the clustering map in any embodiment.

[0070] The computer storage medium of the embodiment of the present invention may adopt any combination of one or more computer-readable media. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (non-exhaustive list) of the computer-readable storage medium include: an electrical connection having 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 this document, the computer-readable storage medium may be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system, apparatus, or device.

[0071] The computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which computer-readable program code is carried. Such a propagated data signal may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium may also be any computer-readable medium other than the computer-readable storage medium, and this computer-readable medium can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device.

[0072] The program code contained on a computer-readable medium can be transmitted using any appropriate medium, including but not limited to wireless, wire, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0073] The computer program code for performing the operations of the present invention can be written in one or more programming languages or combinations thereof. The programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the C language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, executed as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can 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 it can be connected to an external computer (e.g., by connecting through the Internet using an Internet service provider).

[0074] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features. These modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the technical solutions of the embodiments of the present invention.

Claims

1. A method for adaptively processing abnormal areas of a cluster map, characterized in that: include: S10, obtaining information of each class in a cluster map of a plurality of data classes, wherein the information includes data points within the class, normal areas, and abnormal areas; S20, for an abnormal class with an abnormal area, determining at least one first area that is adjacent to a target normal area of ​​the abnormal class and belongs to another class; S30, determining a second area in the at least one first area that is closest to the abnormal area of ​​the abnormal class; S40, if the second area is empty, randomly generate data points in the target normal area to replace the data points in the abnormal area; if the second area is not empty, determine the area where the second area intersects with the convex polygon of the target normal area; If the number of data points a in the intersecting area is equal to the number of data points b in the abnormal area, all the data points in the intersecting area are replaced with the data points in the abnormal area; if the number of data points a in the intersecting area is greater than the number of data points b in the abnormal area, b data points closest to the target normal area are selected from the intersecting area, and the b data points are replaced with all the data points in the abnormal area; if the number of data points a in the intersecting area is less than the number of data points b in the abnormal area, all the data points in the intersecting area are replaced with a data points in the abnormal area.

2. The method according to claim 1, characterized in that S30 includes: in the at least one first region: Determine a first region including an abnormal region of the abnormal class as a second region closest to the abnormal region; If there is no first region including the abnormal region, determining the first region intersecting with the abnormal region as the second region closest to the abnormal region; If there is no first region intersecting with the abnormal region, the first region having a common border with the abnormal region is determined to be the second region closest to the abnormal region.

3. The method according to claim 1, after the operation of the second area being non-empty in S40 is completed, further comprising: If after all the data points in the intersecting area are replaced, there are still first data points in the abnormal area that have not been replaced, determine the second data point in the target normal area that is closest to the first data point, and replace the first data point with the data point in the second area that is closest to the second data point.

4. The method according to claim 1, characterized in that: In the case where there are multiple abnormal regions in the abnormal class, S30 includes: selecting an abnormal region from the abnormal class, and determining a second region in the at least one first region that is closest to the abnormal region; Correspondingly, after S40, it also includes: S50, re-execute S30-S50 according to the new information of each category in the cluster map after replacement, until the latest abnormal areas in the abnormal category are processed.

5. The method according to claim 4, characterized in that Prior to S50, it also included: According to the at least one data point closest to the target normal area, expanding the boundary of the target normal area and shrinking the boundary of the second area; Deleting the abnormal area; Update the data points of the class where the target normal area and the second area are located.

6. The method according to claim 1, characterized in that In the case where there are multiple abnormal classes, S20 includes: selecting an abnormal class with an abnormal area, and determining at least one first area that is adjacent to a target normal area of ​​the abnormal class and belongs to another class; Correspondingly, after S40, it also includes: S60, re-execute S20-S60 according to the new information of each class in the cluster map after replacement, until the latest abnormal classes are processed.

7. An electronic device, characterized in that: include: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the abnormal area adaptive processing method of the cluster map according to any one of claims 1-6.

8. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the program is executed by a processor, the abnormal area adaptive processing method of the cluster map described in any one of claims 1-6 is implemented.

Images