Enterprise classification method, device, electronic device and readable medium

Abstract

The embodiment of the present invention provides a method, device, electronic device and readable medium for enterprise classification, which are applied to an annuity crossover enterprise classification platform, the method comprising: when an enterprise adds annuity business, identifying whether the enterprise is an existing user; if the enterprise is an existing user, determining that the enterprise is an annuity crossover enterprise, and obtaining the business type of the enterprise's existing business and the business type of the added business; using the business type of the enterprise's existing business and the business type of the added business, determining the coordinates of the enterprise in a window generated based on the business type and annuity enterprise classification; determining the annuity enterprise classification of the enterprise according to the coordinates of the enterprise. The annuity enterprise classification to which the enterprise is most likely to belong is efficiently and accurately determined through the window generated based on the business type and enterprise classification, thereby achieving accurate classification of the enterprise.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to an enterprise classification method, an enterprise classification device, an electronic device and a computer-readable medium. Background Art

[0002] Generally speaking, after an enterprise has handled a certain business, the enterprise can usually be managed by the business department corresponding to the business to provide the enterprise with the services it needs. However, after an enterprise that has already handled a business further handles a new business, in order to avoid possible management confusion, the enterprise can be managed by one of the multiple business departments.

[0003] In the prior art, when selecting a business department from multiple business departments that are connected to an enterprise to manage the enterprise, it is usually necessary to use manual classification or pre-set classification templates for matching to determine the business department corresponding to the enterprise. However, the manual classification method may not accurately determine the business department corresponding to the enterprise. If a new business appears, the operation and maintenance personnel need to adjust the template and add new rules, which makes the actual maintenance process more cumbersome. Summary of the invention

[0004] The embodiment of the present invention provides a method, device, electronic device and computer-readable storage medium for classifying enterprises, so as to improve the efficiency of classifying enterprises that handle multiple businesses, and the classification accuracy is high.

[0005] An embodiment of the present invention discloses a method for classifying enterprises, which is applied to an annuity cross-enterprise classification platform. The method includes:

[0006] When an enterprise adds annuity business, identify whether the enterprise is an existing user;

[0007] In the case where the enterprise is an existing user, determining that the enterprise is an annuity crossover enterprise, and obtaining the business type of the enterprise's existing business and the business type of the added business;

[0008] Using the business type of the existing business of the enterprise and the business type of the added business, determine the coordinates to which the enterprise belongs in a window generated based on the business type and the annuity enterprise classification;

[0009] According to the coordinates to which the enterprise belongs, the pension enterprise classification of the enterprise is determined.

[0010] Optionally, the step of using the business type of the existing business of the enterprise and the business type of the added business to determine the coordinates to which the enterprise belongs in a window generated based on the business type and the annuity enterprise classification includes:

[0011] Based on the business type of the business added by the enterprise, sliding an original window generated based on the business type of the existing business of the enterprise to update the prediction object in the original window to obtain an updated window;

[0012] In the update window, the coordinates to which the enterprise belongs are determined.

[0013] Optionally, the update window is divided into a plurality of cells, and the types of the cells include fully dominated, semi-dominated, and contour; and the type of the cell is determined based on a positional relationship of the predicted object.

[0014] Optionally, the step of determining the coordinates of the enterprise in the update window includes:

[0015] If the number of cells of the fully controlled type increases in the updated window compared to the original window, determining the annuity enterprise classification corresponding to the business type of the added business as the annuity enterprise classification of the enterprise;

[0016] If the number of cells of the fully-dominated type does not increase in the updated window compared to the original window, the coordinates to which the enterprise belongs are determined in the updated window.

[0017] Optionally, the step of determining the coordinates of the enterprise in the update window includes:

[0018] Based on a preset k value, determining a k-representative contour set corresponding to the window; the k-representative contour set includes at least one prediction object;

[0019] A target prediction object is determined in the k-representative profile set, and the coordinates of the target prediction object are used as the coordinates to which the enterprise belongs.

[0020] Optionally, the step of determining the k-representative contour set corresponding to the window based on a preset k value further includes:

[0021] Dividing the update window into a plurality of slices;

[0022] According to the segmentation result and the preset k value, at least one k-representative contour set corresponding to the future window is predicted.

[0023] Optionally, the step of determining a target prediction object in the k-representative profile set and using the coordinates of the target prediction object as the coordinates to which the enterprise belongs includes:

[0024] Determining a target k representative contour set in the k representative contour set according to the dominant area of ​​the cells corresponding to the k representative contour set;

[0025] A target prediction object is determined in the target k representative profile set, and the coordinates of the target prediction object are used as the coordinates to which the enterprise belongs.

[0026] The embodiment of the present invention further discloses an enterprise classification device, which is applied to an annuity cross-enterprise classification platform, and the device includes:

[0027] An identification module, used to identify whether the enterprise is an existing user when the enterprise adds annuity business;

[0028] A business determination module, used to determine that the enterprise is an annuity crossover enterprise when the enterprise is an existing user, and obtain the business type of the enterprise's existing business and the business type of the added business;

[0029] A coordinate determination module, for determining the coordinates of the enterprise in a window generated based on the business type and the classification of annuity enterprises, using the business type of the enterprise's existing business and the business type of the added business;

[0030] The classification module is used to determine the annuity enterprise classification of the enterprise according to the coordinates to which the enterprise belongs.

[0031] Optionally, the coordinate determination module includes:

[0032] A sliding submodule, configured to slide an original window generated based on the business type of the existing business of the enterprise based on the business type of the business added by the enterprise to update the prediction object in the original window to obtain an updated window;

[0033] The coordinate determination submodule is used to determine the coordinates to which the enterprise belongs in the update window.

[0034] Optionally, the update window is divided into a plurality of cells, and the types of the cells include fully dominated, semi-dominated, and contour; and the type of the cell is determined based on a positional relationship of the predicted object.

[0035] Optionally, the coordinate determination submodule includes:

[0036] a classification unit, configured to determine, if the number of cells of the fully dominated type increases in the updated window compared to the original window, the annuity enterprise classification corresponding to the business type of the added business as the annuity enterprise classification of the enterprise;

[0037] The first coordinate determining unit is configured to determine the coordinates to which the enterprise belongs in the updated window if the number of cells of the fully-dominated type does not increase in the updated window compared with the original window.

[0038] Optionally, the coordinate determination submodule includes:

[0039] A set determining unit, configured to determine, based on a preset k value, a k-representative contour set corresponding to the window; the k-representative contour set includes at least one prediction object;

[0040] The second coordinate determination unit is used to determine a target prediction object in the k-representative profile set, and use the coordinates of the target prediction object as the coordinates to which the enterprise belongs.

[0041] Optionally, the set determination unit includes:

[0042] A sharding subunit, used for dividing the update window into a plurality of shards;

[0043] The prediction subunit is used to predict at least one k-representative contour set corresponding to the future window according to the segmentation result and the preset k value.

[0044] Optionally, the second coordinate determining unit includes:

[0045] a target set determination module, configured to determine a target k representative contour set in the k representative contour set according to the dominant area of ​​the cells corresponding to the k representative contour set;

[0046] The coordinate determination module is used to determine the target prediction object in the target k representative contour set, and use the coordinates of the target prediction object as the coordinates to which the enterprise belongs.

[0047] The embodiment of the present invention further discloses an electronic device, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0048] The memory is used to store computer programs;

[0049] The processor is used to implement the method described in the embodiment of the present invention when executing the program stored in the memory.

[0050] The embodiment of the present invention further discloses one or more computer-readable media on which instructions are stored. When executed by one or more processors, the processors are enabled to execute the method described in the embodiment of the present invention.

[0051] The embodiments of the present invention include the following advantages:

[0052] Through the enterprise classification method of the embodiment of the present invention, when an enterprise adds annuity business, it is identified whether the enterprise is an existing user to determine whether the enterprise classification needs to be re-determined. In the case that the enterprise is an existing user, it is determined that the enterprise is an annuity crossover enterprise and needs to be re-determined. The business type of the enterprise's existing business and the business type of the added business are obtained, so as to determine the enterprise's annuity enterprise classification based on the business type of the enterprise. Thereafter, the business type of the enterprise's existing business and the business type of the added business are used to determine the coordinates of the enterprise in a window generated based on the business type and the annuity enterprise classification, so that the coordinates of the enterprise can be quickly determined in the window, and the annuity enterprise classification of the enterprise is determined according to the coordinates of the enterprise, so that the annuity enterprise classification of the enterprise is efficiently and accurately determined through the window generated based on the business type and the annuity enterprise classification, and the accurate classification of the enterprise is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 is a flowchart of the steps of an enterprise classification method provided in an embodiment of the present invention;

[0054] Figure 2 is a flowchart of another enterprise classification method provided in an embodiment of the present invention;

[0055] Figure 3 is a schematic diagram of a window according to an embodiment of the present invention;

[0056] Figure 4 is a structural block diagram of an enterprise classification device provided in an embodiment of the present invention;

[0057] Figure 5 is a block diagram of an electronic device provided in an embodiment of the present invention;

[0058] Figure 6 is a schematic diagram of a computer-readable medium provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0059] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0060] The embodiment of the present invention generates a window based on business type and annuity enterprise classification, and when the annuity enterprise classification needs to be determined, the coordinates that are most likely to match the enterprise are determined in the window generated based on the business type and annuity enterprise classification, and then the annuity enterprise classification of the enterprise is determined according to the coordinates to which the enterprise belongs, thereby efficiently determining the annuity enterprise classification of the enterprise with high accuracy.

[0061] Reference Figure 1, shows a flow chart of steps of an enterprise classification method provided in an embodiment of the present invention, which is applied to an annuity cross-enterprise classification platform. Specifically, the annuity cross-enterprise classification platform can be used to accurately classify enterprises when enterprises handle annuity businesses of multiple business types at the same time, avoid multiple business parts tracking the same enterprise customer at the same time, provide services to enterprises accurately, and avoid waste of human resources.

[0062] The method may specifically include the following steps:

[0063] Step 101, when an enterprise adds a service, identifying whether the enterprise is an existing user;

[0064] Specifically, if an enterprise is an existing user, the enterprise may have handled annuity business. In this case, if the enterprise adds annuity business, the enterprise may correspond to multiple business departments, thus having multiple possible annuity enterprise classifications. For example, an enterprise may handle a variety of annuity collection plans with different risk levels at the same time, or it may handle value-added business while handling annuity plans, resulting in the enterprise belonging to multiple annuity enterprise classifications at the same time. In this case, it is necessary to determine an annuity enterprise classification from multiple possible annuity enterprise classifications so that the same business department can subsequently connect with the enterprise.

[0065] Therefore, when an enterprise adds a business, it can be identified whether the enterprise is an existing user to determine whether the enterprise already has business and needs to redefine the category to which the enterprise belongs.

[0066] Step 102, when the enterprise is an existing user, determining that the enterprise is an annuity crossover enterprise, and obtaining the business type of the enterprise's existing business and the business type of the added business;

[0067] In the case where the enterprise is an existing user, the enterprise has added annuity business, which results in it having annuity business of multiple business types at the same time. At this time, the enterprise can be considered as an annuity crossover enterprise. In addition, since an enterprise can belong to multiple annuity enterprise classifications, when it is necessary to determine the corresponding annuity enterprise classification of the enterprise from multiple annuity enterprise classifications, the annuity enterprise classification of the enterprise can be changed from the original annuity enterprise classification to the new annuity enterprise classification, so it can be considered that the annuity enterprise classification currently corresponding to the enterprise may need to be changed. In this case, the business types of the enterprise's existing businesses and the business types of the added businesses can be further determined to re-determine the annuity enterprise classification to which the enterprise should belong based on all the business types currently corresponding to the enterprise.

[0068] Step 103, using the business type of the existing business of the enterprise and the business type of the added business, determine the coordinates of the enterprise in a window generated based on the business type and the annuity enterprise classification;

[0069] In order to determine the annuity enterprise classification to which the enterprise belongs, a window can be generated based on the business type and the annuity enterprise classification. The window can have several prediction objects, and each prediction object can represent a classification prediction result that may correspond to an enterprise. The coordinates of each prediction object in the window can be determined based on the business type of the enterprise's existing business and the business type of the added business. In the window, the target prediction object corresponding to the enterprise can be further determined, and the target prediction object coordinates are used as the coordinates to which the enterprise belongs.

[0070] In a specific implementation, the window may correspond to a two-dimensional coordinate system, in which the business type may be used as the horizontal coordinate and the annuity enterprise classification may be used as the vertical coordinate. The step lengths of the business type and the annuity enterprise classification are both 1. In the case where a new business type corresponding to the enterprise is added, a new prediction object may be generated based on the new business type of the enterprise. The original business type of the enterprise may also correspond to a prediction object. Afterwards, among all the prediction objects corresponding to the enterprise, based on the positional relationship between the prediction objects in the window, the prediction object that the enterprise is most likely to correspond to may be determined as the target prediction object, and the coordinates corresponding to the target prediction object may be used as the coordinates to which the enterprise belongs.

[0071] Step 104, determining the pension enterprise classification of the enterprise according to the coordinates to which the enterprise belongs.

[0072] After determining the coordinates to which the enterprise belongs, since the window is generated based on the business type and the annuity enterprise classification, the coordinates to which the enterprise belongs can point to the annuity enterprise classification to which the enterprise belongs. Thus, the enterprise can be classified through the window and the coordinates described by the enterprise.

[0073] Specifically, the window may use the business type as the horizontal axis and the annuity enterprise classification as the vertical axis. Thus, after determining the coordinates of the enterprise, the annuity enterprise classification pointed to by the vertical axis corresponding to the coordinates may be used as the annuity enterprise classification of the enterprise.

[0074] As an example of the present invention, if an enterprise's original business type corresponds to insurance business, its annuity enterprise classification is insurance business, and a new business type annuity business is added later. At this time, in the window generated based on the business type and annuity enterprise classification, there can be several prediction objects generated based on insurance business and annuity business. The prediction object that the enterprise is most likely to correspond to can be determined in the window as the target prediction object, and the coordinates corresponding to the target prediction object are used as the coordinates to which the enterprise belongs. Thereafter, according to the coordinates to which the enterprise belongs, the annuity enterprise classification of the enterprise is determined to be annuity classification.

[0075] Through the annuity enterprise classification method of the embodiment of the present invention, when an enterprise adds a business, it is identified whether the enterprise is an existing user to determine whether the annuity enterprise classification needs to be re-determined. In the case that the enterprise is an existing user, it is determined that the annuity enterprise classification needs to be re-determined, and the business type of the enterprise's existing business and the business type of the added business are determined, so as to determine the annuity enterprise classification of the enterprise based on the business type of the enterprise. Thereafter, the business type of the enterprise's existing business and the business type of the added business are used to determine the coordinates of the enterprise in a window generated based on the business type and the annuity enterprise classification, so that the coordinates of the enterprise can be quickly determined in the window, and the annuity enterprise classification of the enterprise is determined based on the coordinates of the enterprise, so that the annuity enterprise classification of the enterprise is efficiently and accurately determined through the window generated based on the business type and the annuity enterprise classification, so as to achieve accurate classification of the enterprise.

[0076] Reference Figure 2 , shows a flowchart of the steps of an annuity enterprise classification method provided in an embodiment of the present invention, which is applied to an annuity cross-enterprise classification platform. Specifically, the annuity cross-enterprise classification platform can be used to accurately classify enterprises when enterprises handle annuity businesses of multiple business types at the same time, avoid multiple business parts tracking the same enterprise customer at the same time, provide services to enterprises accurately, and avoid waste of human resources.

[0077] The method may specifically include the following steps:

[0078] Step 201, when an enterprise adds annuity business, identifying whether the enterprise is an existing user;

[0079] Specifically, if an enterprise is an existing user, the enterprise may have handled annuity business. In this case, if the enterprise adds annuity business, the enterprise may correspond to multiple business departments, and thus have multiple possible annuity enterprise classifications. In this case, it is necessary to determine an annuity enterprise classification from multiple possible annuity enterprise classifications so that the same business department can subsequently connect with the enterprise.

[0080] Therefore, when an enterprise adds a business, it can be identified whether the enterprise is an existing user to determine whether the enterprise already has business and needs to redefine the category to which the enterprise belongs.

[0081] Step 202, when the enterprise is an existing user, determining that the enterprise is an annuity crossover enterprise, and obtaining the business type of the enterprise's existing business and the business type of the added business;

[0082] In the case where the enterprise is an existing user, the enterprise has added annuity business, which results in it having annuity business of multiple business types at the same time. At this time, the enterprise can be considered as an annuity crossover enterprise. In addition, since an enterprise can belong to multiple annuity enterprise classifications, when it is necessary to determine the corresponding annuity enterprise classification of the enterprise from multiple annuity enterprise classifications, the annuity enterprise classification of the enterprise can be changed from the original annuity enterprise classification to the new annuity enterprise classification, so it can be considered that the annuity enterprise classification currently corresponding to the enterprise may need to be changed. In this case, the business types of the enterprise's existing businesses and the business types of the added businesses can be further determined to re-determine the annuity enterprise classification to which the enterprise should belong based on all the business types currently corresponding to the enterprise.

[0083] Step 203, based on the business type of the business added by the enterprise, slide the original window generated based on the business type of the existing business of the enterprise to update the prediction object in the original window to obtain an updated window;

[0084] Specifically, the window can continuously update the prediction object therein according to the inflow of data. If there is a business type of business added by the enterprise, it can be considered that there is a data inflow window at this time. At this time, it is necessary to slide the original window generated based on the business type of the existing business of the enterprise based on the business type of the business added by the enterprise, so that the prediction object contained in the original window can be updated based on the business type of the business added by the enterprise, and obtain an updated window.

[0085] After the enterprise adds business, based on the added business type, the prediction objects in the original window may change. Some prediction objects may disappear because the prediction results corresponding to the prediction objects no longer exist due to the added business type. At the same time, the enterprise may also increase the corresponding annuity enterprise classification due to the added business type, which may lead to the addition of new prediction objects. Therefore, based on the changes in the corresponding business types of the enterprise, new data streams can continuously flow in. The original window can slide based on the update of the data stream, and the prediction objects in the window can be updated to obtain an updated window.

[0086] Step 204, in the update window, determining the coordinates to which the enterprise belongs;

[0087] The window may correspond to a two-dimensional coordinate system, in which the business type may be used as the horizontal coordinate and the annuity enterprise classification may be used as the vertical coordinate. The step lengths of the business type and the annuity enterprise classification are both 1. Among all the prediction objects corresponding to the enterprise, based on the positional relationship between the prediction objects in the window, the prediction object that is most likely to correspond to the enterprise may be determined as the target prediction object, and the coordinates corresponding to the target prediction object may be used as the coordinates to which the enterprise belongs.

[0088] In an embodiment of the present invention, the update window is divided into a plurality of cells, and the types of the cells include fully dominated, semi-dominated, and contour; the type of the cell is determined based on the positional relationship of the predicted object.

[0089] In the present invention, the window can be divided into a plurality of cells based on the horizontal coordinate and the vertical coordinate of the window. Thus, the predicted object can be located in a certain cell according to its coordinates. The cells can be divided into three types: full domination, semi-domination, and outline according to the dominance relationship between the predicted objects contained in the cells and the predicted objects contained in other cells.

[0090] Specifically, for any cell, if there is no other cell with the predicted object located to the lower left of the cell, the cell can be a contour cell. If there is a cell located to the lower left of the cell and the contour cell contains the predicted object, the cell is a fully dominated cell. If there is a contour cell located to the lower left of the cell and the contour cell does not contain the predicted object, the cell is a semi-dominated cell.

[0091] As an example of the present invention, Figure 3 : is a window schematic diagram of an embodiment of the present invention. In the window schematic diagram, there is no predicted object located at the lower left of the cell {c0, c1, c2, c3, c5, c6, c7, c10, c11, c15}, so the cell {c0, c1, c2, c3, c5, c6, c7, c10, c11, c15} is a contour cell. There is an object located at the lower left of the cell {c9, c13, c14, c17, c18, c19, c21, c22, c23, c24}, so the lower left of the cell {c9, c13, c14, c17, c18, c19, c21, c22, c23, c24} is a fully dominated cell. The remaining cells are semi-dominated cells.

[0092] Additionally, cells in a window can have the following characteristics:

[0093] For an n×n grid G, the number of contour cells in G is recorded as |cS|, and the number of semi-dominated cells is recorded as |cSD|, then:

[0094] |cSD|+|cS|≤2n-1

[0095] Specifically, if cells c1 and c2, o1 and o2 are prediction objects in them respectively. Based on the dominance relationship between cells, o1 cannot be located at the lower left or upper right of o2, that is, c1 cannot be located at the lower left or upper right of c2. Therefore, it can be said that there is at most one cell containing a prediction object between any diagonal lines li in G, that is, i = 2n-1, and the prediction object only exists in contour cells and semi-dominated cells, so: |cSD|+|cS|≤2n-1.

[0096] In an embodiment of the present invention, the step of determining the coordinates to which the enterprise belongs in the update window includes:

[0097] S11, if the number of cells of the fully controlled type in the updated window increases compared with the original window, determining the annuity enterprise classification corresponding to the business type of the added business as the annuity enterprise classification of the enterprise;

[0098] If after the window is updated based on the newly added business type of the enterprise, the number of fully dominated cells in the updated window increases significantly compared with the original window, it can be considered that the dominance of the new business is stronger. At this time, there is no need to further compare different prediction objects, but the annuity enterprise classification corresponding to the newly added business type can be directly used as the annuity enterprise classification of the enterprise.

[0099] S12: If the number of cells of the fully-dominated type does not increase in the updated window compared to the original window, determine the coordinates to which the enterprise belongs in the updated window.

[0100] If the number of cells of the fully dominated type does not increase in the updated window compared to the original window, the dominance of the newly added business is not significantly stronger than the existing business of the enterprise. In this case, it is necessary to further determine the target prediction object corresponding to the enterprise in the updated window to determine the coordinates to which the enterprise belongs.

[0101] In an embodiment of the present invention, the step of determining the coordinates to which the enterprise belongs in the update window includes:

[0102] S21, based on a preset k value, determining a k-representative contour set corresponding to the window; the k-representative contour set includes at least one prediction object;

[0103] Specifically, based on a preset k value, a prediction object corresponding to at least one cell may be selected and added to the set, thereby obtaining a k-representative contour set corresponding to the current window.

[0104] Among them, the value of k can be determined according to actual needs. For example, the value of k is determined according to the PBA (Population Based Augmentation) algorithm, or the value of k is determined to be 1, 2, 3, 5, 10, etc. based on the actual classification effect. The present invention places no limitation thereon.

[0105] In specific implementation, an object can be arbitrarily selected from each of the contour cells, and on this basis, an approximate k-representative contour set is calculated according to the size relationship between the number of contour cells and the parameter k.

[0106] Specifically, for a contour set SK and its corresponding contour cell set cS, if the number of contour cells satisfies |c S |≥k, then a predicted object can be selected from each of the contour cells and added to the k-representative contour set RS. If |cS|<k, then first, predicted objects are selected from each of the contour cells and added to the k-representative contour set RS. Subsequently, k - |RS| predicted objects are arbitrarily selected from these contour cells and added to RS. Among them, predicted objects close to the cell edge can be preferentially selected to be added to the k-representative contour set RS. Thus, the k-representative contour set RS can be obtained.

[0107] S22. Determine a target predicted object in the k-representative contour set, and use the coordinates of the target predicted object as the coordinates to which the enterprise belongs.

[0108] Specifically, a target predicted object can be further determined in the k-representative contour set.

[0109] In specific implementation, based on the dominance relationship between predicted objects, a predicted object in the k-representative contour set that is not dominated by any other predicted object is determined as the target predicted object.

[0110] Specifically, the dominance relationship between predicted objects can be: for any non-fully dominated cell or semi-dominated cell, if there is a contour cell located at its lower left, then all objects in the non-fully dominated cell or semi-dominated cell are dominated by other objects.

[0111] In specific implementation, for a predicted object o in a cell c, it can have a dominance area, and its dominance area |DS(o)| can be the area of all fully dominated cells cFD with respect to the cell c, the area of the part dominated by the predicted object o in all semi-dominated cells cSD with respect to the cell c, and the area dominated by the predicted object o in the cell c.

[0112] Therefore, the prediction object with the largest dominant area in the k-representative profile set can be used as the target prediction object, and the coordinates of the target prediction object can be used as the coordinates of the enterprise to which it belongs.

[0113] The step of determining the k-representative contour set corresponding to the window based on the preset k value further includes:

[0114] S31, dividing the update window into multiple slices;

[0115] In order to further improve the efficiency of determining the target prediction object, the update window can be divided into multiple slices. Based on the slices, the area where the target prediction object is more likely to exist is determined, and for the area where the target prediction object is more likely to exist, the k-representative contour set of the future window in the area is predicted, so that the annuity enterprise classification can be determined quickly by predicting the future window without sliding the window in sequence.

[0116] Specifically, for the current window, the window may be first divided into slices {P0, P1, ..., Pu}. The number of slices may be determined according to actual needs, for example, 2 slices, 3 slices, 5 slices, etc., and the present invention does not limit this.

[0117] Specifically, whenever the window is updated and the predicted object is updated, the area where the new predicted object is located can be fragmented first, without waiting for all predicted objects to be updated before performing fragmentation.

[0118] S32, predicting at least one k-representative contour set corresponding to the future window according to the segmentation result and the preset k value.

[0119] After the window is segmented, the prediction objects can be screened according to the segmentation results, and the prediction objects that may appear in the future window can be predicted. Then, based on the preset k value, the possible prediction objects are added to the k-representative contour set, so that at least one k-representative contour set corresponding to the future window can be predicted.

[0120] Specifically, the following properties can be used to filter out invalid objects and predict “potential” prediction objects to become target prediction objects.

[0121] (1) Given any prediction object o in slice Pi, if prediction object o is dominated by prediction objects in slice Pj (j≥i), then prediction object o will not become the target prediction object during its life cycle.

[0122] (2) Given any prediction object o in slice Pi, if prediction object o is dominated by prediction objects in slice Pi, then prediction object o will not become the target prediction object before slice Pi becomes the first slice in the window.

[0123] According to the above properties, on the one hand, invalid data in the window (i.e., prediction objects that cannot become target prediction objects) can be deleted, and on the other hand, other "potential" prediction objects can be predicted to become target prediction objects.

[0124] For the prediction object o, if the prediction object o is not dominated by any of the objects, the prediction object o is called the prediction object of the future window Wi. The future window Wi contains the prediction objects in the slide SWi = {si, si+1,…, si+m-1} that change during the sliding window. Therefore, the slice SWi = {si, si+1,…, sm-1} is both a slice of the current window W0 and a slice of the future window Wi. Therefore, SWi can be used to predict the query results of the future window Wi. The prediction objects in the slice SWi constitute the k-representative contour set of the future window Wi, denoted as SKi. If o is the prediction object represented by k in SKi, the prediction object o is called the predicted k-representative contour object of the future window Wi. The prediction objects represented by k in SWi constitute the predicted k-representative contour set of the future window Wi, denoted as RSi.

[0125] For example, suppose that slice SW1 contains slides {s1,…,s31}, and the predicted objects in slice SW1 constitute the contour set SK1 of window W1. For any predicted object o in slice SW1, if the predicted object o is the k-representative contour object in SK1, then the predicted object o is called the predicted k-representative contour object of W1, denoted as RS1. Incremental maintenance can be performed using only the predicted k-representative prediction object set {RS1, RS2, RS4, RS8, RS16} in the current window W0 and the future windows {W1, W2, W4, W8, W16}. In other words, since the number of slices is log2m, at most O(klog2m+|SK1|) objects can be used to predict the target prediction object of the newly-incoming object. Among them, |SK1| represents the number of predicted contour objects in the future window W1.

[0126] Afterwards, the k-representative contour set can be further processed by using a reverse query method.

[0127] Specifically, reverse query can efficiently delete invalid prediction objects in the current window. Specifically, given the current prediction object o, if the current prediction object o is dominated by the prediction object that has been reversed, the current prediction object o will not become the query result. On the contrary, the current prediction object o is a potential query result. Therefore, reverse query can complete the deletion of invalid prediction objects based on the query results at an average computational cost of O(1).

[0128] The reverse query can predict the earliest time when the predicted object in the window W becomes the target query result. Specifically, given the current scan object o and the contour set SK in the reverse queried predicted object, if the predicted object o can dominate a certain predicted object o' in the contour set SK, the predicted object o' can become the target predicted object at the earliest when o flows out of the window.

[0129] Reverse query can characterize the transitivity of domination relations between objects. Specifically, given the prediction objects {o1, o2, o3}, if prediction object o1 dominates prediction object o2, and prediction object o2 dominates prediction object o3, then prediction object o1 must dominate prediction object o3. Assuming that prediction object o2(o1) is the latest to arrive among all prediction objects that can dominate o3(o2), the transitivity of domination relations between objects can be maintained during the reverse query process. That is, if it is found that o1 dominates o2, it can be inferred that o1 dominates o3. In order to maintain the transitivity of domination relations between prediction objects, a tag o.sky can be added to each prediction object o during the initialization process to record the latest arrival of the prediction object that can dominate o. In this way, when the prediction object is dominated by the newly added prediction object onew, all the prediction objects dominated by it can be deleted in batches. When the newly added prediction object onew arrives, the prediction objects dominated by the newly added prediction object onew in RS can be queried to delete all prediction objects that have become invalid due to the arrival of the newly added prediction object onew.

[0130] When a new set of prediction objects arrives, for a given prediction object o and the predicted k-representative contour sets RSi and RSj (j>i) of future windows Wi and Wj, if the prediction object o is not dominated by the objects in the predicted k-representative contour set RSi, then the prediction object o is also not dominated by the prediction objects in the predicted k-representative contour set RSj. At the current time point T0, for the prediction object o, if the prediction object o' that can dominate o and arrives the latest is queried, if the prediction object o'∈slice SWi, it means that the prediction object o becomes the target prediction object in the window Wi+1 at the earliest, and the prediction object o can be inserted into the bucket Bi+1 corresponding to the contour set SKi+1~SK2i-1. If there is no other prediction object dominating the prediction object o in the current window, it is inserted into the current window and regarded as a potential target prediction object of the future window W1. In particular, if there is a prediction object o' in the current window that is dominated by the prediction object o, it means that the prediction object o' becomes an invalid object and is deleted from the current window. Furthermore, based on the transitivity of the dominance relationship between prediction objects, when the prediction object o' is deleted, all objects in the window that are known to be dominated by the prediction object o' can be deleted.

[0131] As the window continues to slide, the first shard in the current window needs to be repartitioned. Therefore, the repartitioning can be achieved through the two-stage (merge-split) algorithm. The merge stage can be further divided into local merge and global merge.

[0132] Assuming that the slice Pi needs to be re-divided, in the local merging stage, the algorithm uses the initialization algorithm to access the bucket Bi corresponding to the slice Pi, and calculates the earliest time when the object in the bucket becomes a contour object. For any prediction object o in Bi, if it is judged to be an invalid object, it will be deleted. If it is dominated by the prediction object o' in Bi, and the prediction object o' flows into the window earlier than the prediction object o, the time when the prediction object o can become the target prediction object is calculated based on the time when the prediction object o' flows into the window, and it is placed in the bucket corresponding to other slices according to the calculation result.

[0133] After finishing scanning bucket Bi, global merging can be performed. At this time, the remaining objects in bucket Bi can be inserted into the R-Tree index Ti in sequence. For any predicted object o, if it can dominate the predicted objects in index Ti, these objects will be deleted. If the predicted object o is dominated by the predicted object o' in the R-Tree index Ti, the algorithm predicts the time when the predicted object o flows into the window based on the time when the predicted object o' flows into the window, and puts it into the corresponding bucket according to the predicted result.

[0134] Finally, splitting is performed. To achieve efficient splitting, the algorithm first constructs index TTi, which copies all non-leaf nodes in index Ti. Then, traverse index Ti to find the predicted object that can be added to the k-representative contour set in the future window, and transfer the predicted object from index Ti to index TTi. After the transfer, traverse index TTi and index Ti and delete all empty nodes. Next, repeat the above process to find all predicted objects that can be added to the k-representative contour set in the future window. Afterwards, the shards can be rebuilt based on index TTi to complete the redivision of the shards.

[0135] In one embodiment of the present invention, the step of determining a target prediction object in the k-representative profile set and using the coordinates of the target prediction object as the coordinates to which the enterprise belongs includes:

[0136] S41, determining a target k representative contour set in the k representative contour set according to the dominant area of ​​the cells corresponding to the k representative contour set;

[0137] Specifically, after determining the k-representative contour sets corresponding to the current window and multiple future windows, the dominance of each k-representative contour set can be further determined based on the dominance area of ​​the cells corresponding to the k-representative contour sets, and the target k-representative contour set is determined from the k-representative contour set with the strongest dominance, that is, the largest dominance area. Thus, the target k-representative contour set to which the target prediction object belongs is determined.

[0138] Specifically, for the contour set RS composed of any object in each contour cell, its dominant area satisfies:

[0139]

[0140] For k to represent a contour set, its dominant area satisfies:

[0141]

[0142] Here, |cFD| refers to the number of fully dominated cells, |cSD| refers to the number of semi-dominated cells, |cS| refers to the number of contour cells, n refers to the granularity, and ΔS refers to the difference between the area of ​​the k-representing contour set and the overall contour set.

[0143] Furthermore, for any predicted object o in cell c, its dominated area |DS(o)| includes the area of ​​all fully dominated cells cFD regarding cell c, the area of ​​the parts dominated by the predicted object o in all semi-dominated cells cSD regarding cell c, and the area dominated by the predicted object o in cell c.

[0144] For a k-representative contour set, its dominant area may be the union of the dominant areas of each object in the k-representative contour set.

[0145] Thus, the dominant area of ​​the k representative contour sets can be determined, and based on the dominant area, a target k representative contour set with the largest dominant area is further determined among the k representative contour sets.

[0146] S42, determining a target prediction object in the target k representative profile set, and using the coordinates of the target prediction object as the coordinates to which the enterprise belongs.

[0147] After determining the target k representative contour set, based on the dominance relationship between the prediction objects, a prediction object in the target k representative contour set that is not dominated by any other prediction object may be determined as a target prediction object.

[0148] Specifically, the dominance relationship between predicted objects can be: for any non-fully dominated cell or semi-dominated cell, if there is a contour cell located to its lower left, then all objects in the non-fully dominated cell or semi-dominated cell are dominated by other objects.

[0149] In a specific implementation, for the predicted object o in cell c, it can have a dominated area, and its dominated area |DS(o)| can be the area of ​​all fully dominated cells cFD regarding cell c, the area of ​​the part dominated by the predicted object o in all semi-dominated cells cSD regarding cell c, and the area dominated by the predicted object o in cell c.

[0150] Therefore, the prediction object with the largest dominant area in the k-representative profile set can be used as the target prediction object, and the coordinates of the target prediction object can be used as the coordinates of the enterprise to which it belongs.

[0151] Step 205, determining the pension enterprise classification of the enterprise according to the coordinates to which the enterprise belongs.

[0152] After determining the coordinates to which the enterprise belongs, since the window is generated based on the business type and the annuity enterprise classification, the coordinates to which the enterprise belongs can point to the annuity enterprise classification to which the enterprise belongs. Thus, the enterprise can be classified through the window and the coordinates described by the enterprise.

[0153] Specifically, the window may use the business type as the horizontal axis and the annuity enterprise classification as the vertical axis. Thus, after determining the coordinates of the enterprise, the annuity enterprise classification pointed to by the vertical axis corresponding to the coordinates may be used as the annuity enterprise classification of the enterprise.

[0154] As an example of the present invention, if an enterprise's original business type corresponds to insurance business, its annuity enterprise classification is insurance business, and a new business type annuity business is subsequently added. At this time, the original window generated based on the business type and the annuity enterprise classification can slide due to the inflow of data. After the window slides, it can be determined whether the number of cells of the updated window with full dominance type increases compared with the original window. If it increases, it can be considered that the added annuity business has a stronger dominance ability, and the annuity type corresponding to the annuity business can be directly determined as the annuity enterprise classification of the enterprise. If it does not increase, first determine the k-representative contour set corresponding to the current update window, and further slice the update window. And predict at least one k-representative contour set corresponding to the future window based on the slice result and the preset k value. Then, according to the dominance area of ​​the cell corresponding to the k-representative contour set, determine the target k-representative contour set in the k-representative contour set, and then determine the target prediction object in the target k-representative contour set, and use the coordinates of the target prediction object as the coordinates to which the enterprise belongs. Then, according to the coordinates to which the enterprise belongs, determine the annuity enterprise classification of the enterprise.

[0155] Through the enterprise classification method of the embodiment of the present invention, when an enterprise adds annuity business, it is identified whether the enterprise is an existing user to determine whether the annuity enterprise classification needs to be re-determined. In the case that the enterprise is an existing user, it is determined that the enterprise is an annuity crossover enterprise and needs to be re-determined. The business type of the enterprise's existing business and the business type of the added business are obtained, so as to determine the annuity enterprise classification of the enterprise based on the business type of the enterprise. Afterwards, the business type of the enterprise's existing business and the business type of the added business are used to determine the coordinates of the enterprise in the window generated based on the business type and the annuity enterprise classification, so that the coordinates of the enterprise can be quickly determined in the window, and the annuity enterprise classification of the enterprise is determined according to the coordinates of the enterprise, so that the annuity enterprise classification of the enterprise is efficiently and accurately determined through the window generated based on the business type and the annuity enterprise classification, and the accurate classification of the enterprise is achieved.

[0156] It should be noted that, for the sake of simplicity, the method embodiments are described as a series of action combinations, but those skilled in the art should be aware that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0157] Reference Figure 4 , shows a structural block diagram of an enterprise classification device provided in an embodiment of the present invention, which may specifically include the following modules:

[0158] Identification module 401, used to identify whether the enterprise is an existing user when the enterprise adds annuity business;

[0159] The business determination module 402 is used to determine that the enterprise is an annuity crossover enterprise when the enterprise is an existing user, and obtain the business type of the enterprise's existing business and the business type of the added business;

[0160] A coordinate determination module 403, for determining the coordinates of the enterprise in a window generated based on the business type and the annuity enterprise classification, using the business type of the enterprise's existing business and the business type of the added business;

[0161] The classification module 404 is used to determine the annuity enterprise classification of the enterprise according to the coordinates to which the enterprise belongs.

[0162] Optionally, the coordinate determination module includes:

[0163] A sliding submodule, configured to slide an original window generated based on the business type of the existing business of the enterprise based on the business type of the business added by the enterprise to update the prediction object in the original window to obtain an updated window;

[0164] The coordinate determination submodule is used to determine the coordinates to which the enterprise belongs in the update window.

[0165] Optionally, the update window is divided into a plurality of cells, and the types of the cells include fully dominated, semi-dominated, and contour; and the type of the cell is determined based on a positional relationship of the predicted object.

[0166] Optionally, the coordinate determination submodule includes:

[0167] a classification unit, configured to determine, if the number of cells of the fully dominated type increases in the updated window compared to the original window, the annuity enterprise classification corresponding to the business type of the added business as the annuity enterprise classification of the enterprise;

[0168] The first coordinate determining unit is configured to determine the coordinates to which the enterprise belongs in the updated window if the number of cells of the fully-dominated type does not increase in the updated window compared with the original window.

[0169] Optionally, the coordinate determination submodule includes:

[0170] A set determining unit, configured to determine, based on a preset k value, a k-representative contour set corresponding to the window; the k-representative contour set includes at least one prediction object;

[0171] The second coordinate determination unit is used to determine a target prediction object in the k-representative profile set, and use the coordinates of the target prediction object as the coordinates to which the enterprise belongs.

[0172] Optionally, the set determination unit includes:

[0173] A sharding subunit, used for dividing the update window into a plurality of shards;

[0174] The prediction subunit is used to predict at least one k-representative contour set corresponding to the future window according to the segmentation result and the preset k value.

[0175] Optionally, the second coordinate determining unit includes:

[0176] a target set determination module, configured to determine a target k representative contour set in the k representative contour set according to the dominant area of ​​the cells corresponding to the k representative contour set;

[0177] The coordinate determination module is used to determine the target prediction object in the target k representative contour set, and use the coordinates of the target prediction object as the coordinates to which the enterprise belongs.

[0178] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0179] In addition, an embodiment of the present invention further provides an electronic device, such as Figure 5 As shown, it includes a processor 501, a communication interface 502, a memory 503 and a communication bus 504, wherein the processor 501, the communication interface 502, and the memory 503 communicate with each other through the communication bus 504.

[0180] Memory 503, used for storing computer programs;

[0181] The processor 501 is used to execute the program stored in the memory 503, and implements the following steps:

[0182] When an enterprise adds annuity business, identify whether the enterprise is an existing user;

[0183] In the case where the enterprise is an existing user, determining that the enterprise is an annuity crossover enterprise, and obtaining the business type of the enterprise's existing business and the business type of the added business;

[0184] Using the business type of the existing business of the enterprise and the business type of the added business, determine the coordinates to which the enterprise belongs in a window generated based on the business type and the annuity enterprise classification;

[0185] According to the coordinates to which the enterprise belongs, the pension enterprise classification of the enterprise is determined.

[0186] Optionally, the step of using the business type of the existing business of the enterprise and the business type of the added business to determine the coordinates to which the enterprise belongs in a window generated based on the business type and the annuity enterprise classification includes:

[0187] Based on the business type of the business added by the enterprise, sliding an original window generated based on the business type of the existing business of the enterprise to update the prediction object in the original window to obtain an updated window;

[0188] In the update window, the coordinates to which the enterprise belongs are determined.

[0189] Optionally, the update window is divided into a plurality of cells, and the types of the cells include fully dominated, semi-dominated, and contour; and the type of the cell is determined based on a positional relationship of the predicted object.

[0190] Optionally, the step of determining the coordinates of the enterprise in the update window includes:

[0191] If the number of cells of the fully controlled type increases in the updated window compared to the original window, determining the annuity enterprise classification corresponding to the business type of the added business as the annuity enterprise classification of the enterprise;

[0192] If the number of cells of the fully-dominated type does not increase in the updated window compared to the original window, the coordinates to which the enterprise belongs are determined in the updated window.

[0193] Optionally, the step of determining the coordinates of the enterprise in the update window includes:

[0194] Based on a preset k value, determining a k-representative contour set corresponding to the window; the k-representative contour set includes at least one prediction object;

[0195] A target prediction object is determined in the k-representative profile set, and the coordinates of the target prediction object are used as the coordinates to which the enterprise belongs.

[0196] Optionally, the step of determining the k-representative contour set corresponding to the window based on a preset k value further includes:

[0197] Dividing the update window into a plurality of slices;

[0198] According to the segmentation result and the preset k value, at least one k-representative contour set corresponding to the future window is predicted.

[0199] Optionally, the step of determining a target prediction object in the k-representative profile set and using the coordinates of the target prediction object as the coordinates to which the enterprise belongs includes:

[0200] Determining a target k representative contour set in the k representative contour set according to the dominant area of ​​the cells corresponding to the k representative contour set;

[0201] A target prediction object is determined in the target k representative profile set, and the coordinates of the target prediction object are used as the coordinates to which the enterprise belongs.

[0202] The communication bus mentioned in the above terminal can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0203] The communication interface is used for communication between the above terminal and other devices.

[0204] The memory may include a random access memory (RAM) or a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0205] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0206] like Figure 6As shown, in another embodiment provided by the present invention, a computer-readable storage medium 601 is also provided, in which instructions are stored. When the computer-readable storage medium is run on a computer, the computer executes the enterprise classification method described in the above embodiment.

[0207] In another embodiment of the present invention, a computer program product including instructions is provided. When the computer program product is run on a computer, the computer executes the enterprise classification method described in the above embodiment.

[0208] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented by software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk Solid State Disk (SSD)), etc.

[0209] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0210] Each embodiment in this specification is described in a related manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0211] The above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention are included in the protection scope of the present invention.

Claims

1. A method for classifying enterprises, characterized in that: It is applied to an annuity cross-enterprise classification platform, and the method includes: When an enterprise adds annuity business, identify whether the enterprise is an existing user; In the case where the enterprise is an existing user, determining that the enterprise is an annuity crossover enterprise, and obtaining the business type of the enterprise's existing business and the business type of the added business; Using the business type of the existing business of the enterprise and the business type of the added business, determine the coordinates to which the enterprise belongs in a window generated based on the business type and the annuity enterprise classification; Determining the pension enterprise classification of the enterprise according to the coordinates to which the enterprise belongs; The step of using the business type of the existing business of the enterprise and the business type of the added business to determine the coordinates of the enterprise in a window generated based on the business type and the annuity enterprise classification includes: Based on the business type of the business added by the enterprise, sliding an original window generated based on the business type of the existing business of the enterprise to update the prediction object in the original window to obtain an updated window; In the update window, determining the coordinates to which the enterprise belongs; The update window is divided into a plurality of cells, and the types of the cells include fully dominated, semi-dominated, and contour; The type of the cell is determined based on the positional relationship of the prediction object; The type of the cell is determined based on the positional relationship of the prediction objects and the dominance relationship between the prediction objects contained in the cell and the prediction objects contained in other cells; The step of determining the type of the cell based on the positional relationship of the predicted objects and the dominance relationship between the predicted objects contained in the cell and the predicted objects contained in other cells includes: If there is no other cell where the predicted object is located to the lower left of the cell, the cell is a contour cell; If there is a cell located at the lower left of the cell, and the cell contains a prediction object, and all prediction objects in the cell are dominated by other prediction objects, then the cell is a fully dominated cell; If there is a contour cell located at the lower left of the cell, and the contour cell does not contain a predicted object, and the cell has a predicted object that is not dominated by other predicted objects, then the cell is a semi-dominated cell; Wherein, the step of determining the coordinates to which the enterprise belongs in the update window includes: If the number of cells of the fully controlled type increases in the updated window compared to the original window, determining the annuity enterprise classification corresponding to the business type of the added business as the annuity enterprise classification of the enterprise; If the number of cells of the fully controlled type does not increase in the updated window compared to the original window, in the updated window, the coordinates to which the enterprise belongs are determined; wherein the horizontal coordinate of the coordinates to which the enterprise belongs is the business type, and the vertical coordinate is the annuity enterprise classification.

2. The method according to claim 1, characterized in that The step of determining the coordinates of the enterprise in the update window includes: Based on a preset k value, determining a k-representative contour set corresponding to the window; the k-representative contour set includes at least one prediction object; A target prediction object is determined in the k-representative profile set, and the coordinates of the target prediction object are used as the coordinates to which the enterprise belongs.

3. The method according to claim 2, characterized in that The step of determining the k-representative contour set corresponding to the window based on the preset k value further includes: Dividing the update window into a plurality of slices; According to the segmentation result and the preset k value, at least one k-representative contour set corresponding to the future window is predicted.

4. The method according to claim 3, characterized in that The step of determining a target prediction object in the k-representative profile set and using the coordinates of the target prediction object as the coordinates to which the enterprise belongs includes: Determining a target k representative contour set in the k representative contour set according to the dominant area of ​​the cells corresponding to the k representative contour set; A target prediction object is determined in the target k representative profile set, and the coordinates of the target prediction object are used as the coordinates to which the enterprise belongs.

5. An enterprise classification device, characterized in that: It is applied to annuity cross-enterprise classification platform, and the device includes: An identification module, used to identify whether the enterprise is an existing user when the enterprise adds annuity business; A business determination module, used to determine that the enterprise is an annuity crossover enterprise when the enterprise is an existing user, and obtain the business type of the enterprise's existing business and the business type of the added business; A coordinate determination module, for determining the coordinates of the enterprise in a window generated based on the business type and the classification of annuity enterprises, using the business type of the enterprise's existing business and the business type of the added business; A classification module, used for determining the annuity enterprise classification of the enterprise according to the coordinates to which the enterprise belongs; Wherein, the coordinate determination module includes: A sliding submodule, configured to slide an original window generated based on the business type of the existing business of the enterprise based on the business type of the business added by the enterprise to update the prediction object in the original window to obtain an updated window; A coordinate determination submodule, used for determining the coordinates to which the enterprise belongs in the update window; The sliding submodule is further used to divide the update window into a plurality of cells, and the types of the cells include fully dominated, semi-dominated, and contour; The sliding submodule is further used to determine the type of the cell based on the positional relationship of the prediction object; The sliding submodule is specifically used to determine the type of the cell based on the position relationship of the prediction objects and the dominance relationship between the prediction objects contained in the cell and the prediction objects contained in other cells, including: If there is no other cell where the predicted object is located to the lower left of the cell, the cell is a contour cell; If there is a cell located at the lower left of the cell, and the cell contains a prediction object, and all prediction objects in the cell are dominated by other prediction objects, then the cell is a fully dominated cell; If there is a contour cell located at the lower left of the cell, and the contour cell does not contain a predicted object, and the cell has a predicted object that is not dominated by other predicted objects, then the cell is a semi-dominated cell; Wherein, the coordinate determination submodule includes: a classification unit, configured to determine, if the number of cells of the fully dominated type increases in the updated window compared to the original window, the annuity enterprise classification corresponding to the business type of the added business as the annuity enterprise classification of the enterprise; A first coordinate determination unit, configured to determine the coordinates to which the enterprise belongs in the updated window if the number of cells of the fully dominated type does not increase in the updated window compared to the original window; The first coordinate determination unit is further used to use the business type as the horizontal coordinate and the annuity enterprise classification as the vertical coordinate.

6. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; The memory is used to store computer programs; The processor is used to implement the method according to any one of claims 1 to 4 when executing the program stored in the memory.

7. One or more computer-readable media having instructions stored thereon, which, when executed by one or more processors, cause the processors to perform the method according to any one of claims 1 to 4.

Images