A method and system for extracting user transaction data features based on dynamic graph

Abstract

The present invention discloses a method and system for extracting user transaction data features based on dynamic graphs. The method divides the dynamic graph into multiple sub-dynamic graphs, collects user transaction data fed back by the sub-dynamic graphs, forms an initial feature set, processes the data in the initial feature set, determines the main features of the sub-dynamic graphs, and determines the number of occurrences of the main features of the sub-dynamic graphs by presetting a first quantity threshold, thereby improving the accuracy of extracting user transaction data features from the dynamic graphs.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing, and more specifically, to a method and system for extracting user transaction data features based on dynamic graphs. Background Art

[0002] With the development of networking, the proportion of online transactions has gradually increased, and electronic archives of user transaction data are becoming easier to obtain or generate. Accurate analysis of user transaction data is helpful for buyers to buy and sellers to sell. Traditional user transaction data feature extraction methods mainly involve fixed user transaction data feature extraction, which cannot support incremental updates and cannot timely and accurately extract the main features of the latest user transaction data.

[0003] Therefore, the prior art has defects and needs to be improved urgently. Summary of the invention

[0004] In view of the above problems, an object of the present invention is to provide a method and system for extracting user transaction data features based on dynamic graphs, which can improve the accuracy of extracting user transaction data features from dynamic graphs.

[0005] A first aspect of the present invention provides a method for extracting user transaction data features based on a dynamic graph, comprising:

[0006] Get user transaction data information of dynamic graph;

[0007] Based on a preset time period, user transaction data of the dynamic graph is segmented and processed to obtain user transaction data information of sub-dynamic graphs within multiple time periods;

[0008] Extracting an initial feature set of user transaction data of a sub-dynamic graph within a time period;

[0009] According to the initial feature set of user transaction data of the sub-dynamic graph within the time period, the main features of the corresponding sub-dynamic graph are obtained;

[0010] Based on n consecutive time periods, if the main features of the sub-dynamic images are the same, the number of occurrences of the corresponding same main features is accumulated to obtain the number of main features of the sub-dynamic images in the n consecutive time periods;

[0011] If the number of times of the main feature of the sub-dynamic graph in n consecutive time periods is greater than a preset first quantity threshold, the main feature of the sub-dynamic graph in n consecutive time periods is set as the main feature of the user transaction data of the dynamic graph.

[0012] In this solution, after extracting the initial feature set of user transaction data of the sub-dynamic graph within the time period, it also includes:

[0013] Extract features in the initial feature set and the corresponding initial feature value set;

[0014] Extract the initial eigenvalues ​​from the initial eigenvalue set. If the initial eigenvalues ​​are the same, merge the corresponding initial eigenvalues ​​to obtain the merged initial eigenvalues.

[0015] Determine whether the initial feature value is less than the preset first feature threshold of the corresponding feature. If so, delete the corresponding initial feature value; if not, store the corresponding initial feature value.

[0016] In this solution, if the initial eigenvalues ​​are the same, the step of fusing the corresponding initial eigenvalues ​​specifically includes:

[0017] Get the quantity values ​​corresponding to the same initial eigenvalues;

[0018] Obtain the transaction price difference corresponding to the same initial characteristic value, and set it as the first transaction price difference; obtain the transaction price difference within the time period corresponding to the same initial characteristic value, and set it as the second transaction price difference;

[0019] Calculate the average value of the first transaction price difference to obtain the first transaction price average value;

[0020] Divide the average value of the first transaction price by the difference of the second transaction price to obtain the fusion coefficient corresponding to the initial eigenvalue;

[0021] Multiply the same initial eigenvalue by the quantity value corresponding to the same initial eigenvalue by the fusion coefficient to obtain the fusion value corresponding to the same initial eigenvalue;

[0022] The fusion value corresponding to the same initial eigenvalue and the fusion value corresponding to the same initial eigenvalue will be the initial eigenvalue after fusion.

[0023] In this solution, the step of obtaining the transaction price difference corresponding to the same initial characteristic value and the step of obtaining the transaction price difference within the time period corresponding to the same initial characteristic value specifically include:

[0024] Obtain the transaction mark price and actual transaction price at the corresponding transaction time node for the same initial feature value;

[0025] Calculate the difference between the marked transaction price and the actual transaction price, and take the absolute value to obtain the corresponding first transaction price difference;

[0026] Obtain the user transaction price set within the time period corresponding to the same initial characteristic value;

[0027] Extracting the lowest price and the highest price in the user transaction price set;

[0028] Subtract the lowest price from the highest price to get the transaction difference corresponding to the second price.

[0029] In this solution, the step of obtaining the main features of the corresponding sub-dynamic graph according to the initial feature set of user transaction data of the sub-dynamic graph within the time period specifically includes:

[0030] Extracting initial eigenvalues ​​from the initial eigenvalue set;

[0031] According to the preset eigenvalue interval into which the initial eigenvalue in the initial feature set falls, an influence weight coefficient corresponding to the initial eigenvalue is obtained;

[0032] Multiply the initial eigenvalue by the influence weight coefficient corresponding to the initial eigenvalue to obtain the influence value corresponding to the initial eigenvalue;

[0033] Accumulate the influence values ​​corresponding to the initial eigenvalues ​​in the initial feature set to obtain the total influence value of the initial eigenvalues ​​in the corresponding time period;

[0034] If the total impact value of the initial feature value within the corresponding time period is greater than or equal to the preset impact threshold, the feature corresponding to the initial feature value is set as the main feature.

[0035] This plan also includes:

[0036] Get the format information of the dynamic image;

[0037] Determine whether the format of the dynamic image is the same as the preset format. If so, store the corresponding dynamic image; if not, convert the format of the corresponding dynamic image to the same as the preset format based on a preset format conversion algorithm.

[0038] This plan also includes:

[0039] Based on a preset time period, extract m nodes on the sub-dynamic graph, where m is an odd number greater than 1;

[0040] Divide the nodes into intermediate nodes and edge nodes according to their positions on the sub-dynamic graph;

[0041] Extract the value of the node on the sub-dynamic graph;

[0042] Compare and analyze the values ​​of the nodes on the sub-dynamic graph. If the value of the intermediate node on the sub-dynamic graph is greater than the values ​​of all the edge nodes on the sub-dynamic graph or less than the values ​​of all the edge nodes on the sub-dynamic graph, calculate the characteristic angle of the corresponding sub-dynamic graph;

[0043] If the value of the intermediate node on the sub-dynamic graph is greater than or less than the value of the edge node on the sub-dynamic graph, then calculate the slope of the connecting line of the corresponding edge node;

[0044] If the characteristic angle of the corresponding sub-dynamic graph is less than or equal to the preset characteristic angle threshold or the slope of the connecting line of the corresponding edge node is greater than the preset slope threshold, it is obtained that the main characteristic information exists in the corresponding time period.

[0045] In this solution, the step of calculating the characteristic angle of the corresponding sub-dynamic graph specifically includes:

[0046] Extract the minimum value or the maximum value of the values ​​of the intermediate nodes on the sub-dynamic graph, and set them as the first minimum value or the first maximum value respectively;

[0047] When it is the first maximum value, taking the middle node corresponding to the first maximum value as the reference, respectively extract the nodes corresponding to the second minimum values ​​on both sides of the middle node corresponding to the first maximum value, and connect them to obtain the characteristic angle of the corresponding sub-dynamic graph;

[0048] When it is the first minimum value, taking the middle node corresponding to the first minimum value as the reference, respectively extract the nodes corresponding to the second maximum values ​​on both sides of the middle node corresponding to the first minimum value, and connect them to obtain the characteristic angle of the corresponding sub-dynamic graph;

[0049] The second minimum value is the minimum value of the values ​​of all nodes in front of the intermediate node corresponding to the first maximum value on the sub-dynamic graph and the minimum value of the values ​​of all nodes behind the intermediate node corresponding to the first maximum value on the sub-dynamic graph; the second maximum value is the maximum value of the values ​​of all nodes in front of the intermediate node corresponding to the first minimum value on the sub-dynamic graph and the maximum value of the values ​​of all nodes behind the intermediate node corresponding to the first minimum value on the sub-dynamic graph.

[0050] This plan also includes:

[0051] If the total impact value of the initial feature value in the corresponding time period is less than the preset impact threshold, and the feature angle of the corresponding sub-dynamic graph is less than or equal to the preset feature angle threshold, the preset impact threshold revision information in the corresponding time period is generated;

[0052] Based on the preset impact threshold modification information within the corresponding time period, the preset impact base value is subtracted from the corresponding preset impact threshold to obtain the revised preset impact value within the corresponding time period;

[0053] If the total impact value of the initial feature value in the corresponding time period is greater than or equal to the revised preset impact threshold, the feature corresponding to the initial feature value is set as the main feature;

[0054] If the total impact value of the initial characteristic value within the corresponding time period is less than the revised preset impact threshold, the revised preset impact threshold is further revised according to the preset impact base value until the main features of the corresponding sub-dynamic graph are determined.

[0055] A second aspect of the present invention provides a user transaction data feature extraction system based on a dynamic graph, comprising a memory and a processor, wherein the memory stores a user transaction data feature extraction method program based on a dynamic graph as described in any one of the above.

[0056] The present invention discloses a method and system for extracting user transaction data features based on dynamic graphs. The method divides the dynamic graph into multiple sub-dynamic graphs, collects user transaction data fed back by the sub-dynamic graphs, forms an initial feature set, processes the data in the initial feature set, determines the main features of the sub-dynamic graphs, and determines the number of occurrences of the main features of the sub-dynamic graphs by presetting a first quantity threshold, thereby improving the accuracy of extracting user transaction data features from the dynamic graphs. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 A flow chart of a method for extracting user transaction data features based on dynamic graphs according to the present invention is shown;

[0058] Figure 2 A schematic diagram showing the slope of the connection line corresponding to the edge node in the present invention;

[0059] Figure 3 A schematic diagram showing the characteristic angle of the neutron dynamic diagram of the present invention is shown;

[0060] Figure 4 A block diagram of a user transaction data feature extraction method system based on dynamic graphs of the present invention is shown. DETAILED DESCRIPTION

[0061] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0062] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0063] Figure 1 A flow chart of a method for extracting user transaction data features based on dynamic graphs according to the present invention is shown;

[0064] like Figure 1 As shown, the present invention discloses a method for extracting user transaction data features based on a dynamic graph, comprising:

[0065] S101, obtaining user transaction data information of the dynamic graph;

[0066] S102, based on a preset time period, segmenting the user transaction data of the dynamic graph to obtain user transaction data information of sub-dynamic graphs within multiple time periods;

[0067] S103, extracting an initial feature set of user transaction data of the sub-dynamic graph within a time period;

[0068] S104, obtaining main features of the corresponding sub-dynamic graph according to the initial feature set of user transaction data of the sub-dynamic graph within the time period;

[0069] S105, based on n consecutive time periods, if the main features of the sub-dynamic images are the same, then the number of occurrences of the corresponding same main features is accumulated to obtain the number of main features of the sub-dynamic images in the n consecutive time periods;

[0070] S106, if the number of times of the main feature of the sub-dynamic graph in n consecutive time periods is greater than a preset first quantity threshold, the main feature of the sub-dynamic graph in n consecutive time periods is set as the main feature of the user transaction data of the dynamic graph.

[0071] According to an embodiment of the present invention, for example, if the preset time period is 30 seconds, the dynamic image is segmented at intervals of 30 seconds to obtain sub-dynamic images with a duration of 30 seconds. The dynamic image is segmented, and then the main features of each sub-dynamic image are collected, and then a judgment is made based on a preset first quantity threshold. For example, if n consecutive time periods are 10, the judgment is made based on the number values ​​of the main features of the sub-dynamic images in 10 consecutive time periods and the preset first quantity threshold. For example, the preset first quantity threshold is 6. When the number values ​​of the main features of the sub-dynamic images in n consecutive time periods are greater than 6, the main features of the sub-dynamic images in n consecutive time periods are set as the main features of the user transaction data of the dynamic image.

[0072] According to an embodiment of the present invention, after extracting the initial feature set of user transaction data of the sub-dynamic graph within the time period, the method further includes:

[0073] Extract features in the initial feature set and the corresponding initial feature value set;

[0074] Extract the initial eigenvalues ​​from the initial eigenvalue set. If the initial eigenvalues ​​are the same, merge the corresponding initial eigenvalues ​​to obtain the merged initial eigenvalues.

[0075] Determine whether the initial feature value is less than the preset first feature threshold of the corresponding feature. If so, delete the corresponding initial feature value; if not, store the corresponding initial feature value.

[0076] It should be noted that the initial feature values ​​are pre-processed by fusion, deletion, etc. to reduce the calculation of user transaction data of the sub-dynamic graph. The features in the initial feature set include transaction quantity, transaction buyer, transaction seller, transaction number, etc. If the feature is transaction quantity, then the corresponding initial feature value in the initial feature value set is the transaction quantity value; if the feature is transaction buyer, then the corresponding initial feature value in the initial feature value set is the transaction buyer consumption value, etc.

[0077] According to an embodiment of the present invention, if the initial eigenvalues ​​are the same, the step of fusing the corresponding initial eigenvalues ​​specifically includes:

[0078] Get the quantity values ​​corresponding to the same initial eigenvalues;

[0079] Obtain the transaction price difference corresponding to the same initial characteristic value, and set it as the first transaction price difference; obtain the transaction price difference within the time period corresponding to the same initial characteristic value, and set it as the second transaction price difference;

[0080] Calculate the average value of the first transaction price difference to obtain the first transaction price average value;

[0081] Divide the average value of the first transaction price by the difference of the second transaction price to obtain the fusion coefficient corresponding to the initial eigenvalue;

[0082] Multiply the same initial eigenvalue by the quantity value corresponding to the same initial eigenvalue by the fusion coefficient to obtain the fusion value corresponding to the same initial eigenvalue;

[0083] The fusion value corresponding to the same initial eigenvalue and the fusion value corresponding to the same initial eigenvalue will be the initial eigenvalue after fusion.

[0084] It should be noted that the fusion coefficient of the initial eigenvalue is determined by the transaction price. If the transaction price fluctuates greatly, it means that the current user transaction fluctuates greatly. The more obvious the main characteristics of the corresponding user transaction data are, the greater the fusion coefficient of the corresponding initial eigenvalue is.

[0085] According to an embodiment of the present invention, the step of obtaining the transaction price difference corresponding to the same initial characteristic value and the step of obtaining the transaction price difference within a time period corresponding to the same initial characteristic value specifically include:

[0086] Obtain the transaction mark price and actual transaction price at the corresponding transaction time node for the same initial feature value;

[0087] Calculate the difference between the marked transaction price and the actual transaction price, and take the absolute value to obtain the corresponding first transaction price difference;

[0088] Obtain the user transaction price set within the time period corresponding to the same initial characteristic value;

[0089] Extracting the lowest price and the highest price in the user transaction price set;

[0090] Subtract the lowest price from the highest price to get the transaction difference corresponding to the second price.

[0091] It should be noted that the first transaction price difference is the transaction price difference at the time node corresponding to the same initial characteristic value, and the second transaction price difference is the transaction price difference within the time period corresponding to the same initial characteristic value and is the maximum value.

[0092] According to an embodiment of the present invention, the step of obtaining the main features of the corresponding sub-dynamic graph according to the initial feature set of user transaction data of the sub-dynamic graph within the time period specifically includes:

[0093] Extracting initial eigenvalues ​​from the initial eigenvalue set;

[0094] According to the preset eigenvalue interval into which the initial eigenvalue in the initial feature set falls, an influence weight coefficient corresponding to the initial eigenvalue is obtained;

[0095] Multiply the initial eigenvalue by the influence weight coefficient corresponding to the initial eigenvalue to obtain the influence value corresponding to the initial eigenvalue;

[0096] Accumulate the influence values ​​corresponding to the initial eigenvalues ​​in the initial feature set to obtain the total influence value of the initial eigenvalues ​​in the corresponding time period;

[0097] If the total impact value of the initial feature value within the corresponding time period is greater than or equal to the preset impact threshold, the feature corresponding to the initial feature value is set as the main feature.

[0098] It should be noted that the interval cardinality of the preset eigenvalue interval corresponding to the initial eigenvalue is determined according to the characteristics. For example, if the corresponding interval cardinality is 20, 20 is taken as a preset eigenvalue interval. For example, the preset eigenvalue intervals are (0, 20], (20, 40], (40, 60]... and so on, wherein each preset eigenvalue interval corresponds to an influence weight coefficient of an initial eigenvalue, and the larger the value of the preset eigenvalue interval, the larger the influence weight coefficient of the corresponding initial eigenvalue.

[0099] According to an embodiment of the present invention, it also includes:

[0100] Get the format information of the dynamic image;

[0101] Determine whether the format of the dynamic image is the same as the preset format. If so, store the corresponding dynamic image; if not, convert the format of the corresponding dynamic image to the same as the preset format based on a preset format conversion algorithm.

[0102] It should be noted that the format of the dynamic graph is standardized by a preset format, so as to improve the accuracy of extracting user transaction data in the dynamic graph.

[0103] Figure 2 A schematic diagram showing the slope of the connection line corresponding to the edge node in the present invention;

[0104] like Figure 2 As shown, according to an embodiment of the present invention, it also includes:

[0105] Based on a preset time period, extract m nodes on the sub-dynamic graph, where m is an odd number greater than 1;

[0106] Divide the nodes into intermediate nodes and edge nodes according to their positions on the sub-dynamic graph;

[0107] Extract the value of the node on the sub-dynamic graph;

[0108] Compare and analyze the values ​​of the nodes on the sub-dynamic graph. If the value of the intermediate node on the sub-dynamic graph is greater than the values ​​of all the edge nodes on the sub-dynamic graph or less than the values ​​of all the edge nodes on the sub-dynamic graph, calculate the characteristic angle of the corresponding sub-dynamic graph;

[0109] If the value of the intermediate node on the sub-dynamic graph is greater than or less than the value of the edge node on the sub-dynamic graph, then calculate the slope of the connecting line of the corresponding edge node;

[0110] If the characteristic angle of the corresponding sub-dynamic graph is less than or equal to the preset characteristic angle threshold or the slope of the connecting line of the corresponding edge node is greater than the preset slope threshold, it is obtained that the main characteristic information exists in the corresponding time period.

[0111] It should be noted that if Figure 2 As shown in , m in the schematic diagram is set to 5, the X-axis corresponds to time, and the Y-axis corresponds to the initial eigenvalue, where the time corresponding to node A to node E is just a preset time period, the edge nodes are A and E, and the intermediate nodes are C, D and B, where the values ​​of the intermediate nodes B, C and E on the sub-dynamic graph do not exist, which are greater than or less than the values ​​of the edge nodes A and B on the sub-dynamic graph, that is, the values ​​of the intermediate nodes B, C and E on the sub-dynamic graph are greater than the value of one edge node on the sub-dynamic graph and less than the value of another edge node on the sub-dynamic graph; therefore, the slope of the connecting line corresponding to the edge node, that is, the slope of the dotted line AE, is calculated, and the node A (x 1 ,y 1 ) and node E(x 2 ,y 2 ), the slope of the dotted line AE

[0112] Figure 3 A schematic diagram showing the characteristic angle of the neutron dynamic diagram of the present invention is shown,

[0113] like Figure 3 As shown, according to an embodiment of the present invention, the step of calculating the characteristic angle of the corresponding sub-dynamic graph specifically includes:

[0114] Extract the minimum value or the maximum value of the values ​​of the intermediate nodes on the sub-dynamic graph, and set them as the first minimum value or the first maximum value respectively;

[0115] When it is the first maximum value, taking the middle node corresponding to the first maximum value as the reference, respectively extract the nodes corresponding to the second minimum values ​​on both sides of the middle node corresponding to the first maximum value, and connect them to obtain the characteristic angle of the corresponding sub-dynamic graph;

[0116] When it is the first minimum value, taking the middle node corresponding to the first minimum value as the reference, respectively extract the nodes corresponding to the second maximum values ​​on both sides of the middle node corresponding to the first minimum value, and connect them to obtain the characteristic angle of the corresponding sub-dynamic graph;

[0117] The second minimum value is the minimum value of the values ​​of all nodes in front of the intermediate node corresponding to the first maximum value on the sub-dynamic graph and the minimum value of the values ​​of all nodes behind the intermediate node corresponding to the first maximum value on the sub-dynamic graph; the second maximum value is the maximum value of the values ​​of all nodes in front of the intermediate node corresponding to the first minimum value on the sub-dynamic graph and the maximum value of the values ​​of all nodes behind the intermediate node corresponding to the first minimum value on the sub-dynamic graph.

[0118] If necessary, it is stated that Figure 3 As shown in , m in the schematic diagram is set to 5, the X-axis corresponds to time, and the Y-axis corresponds to the initial eigenvalue, wherein the time corresponding to node A to node E is just a preset time period, the edge nodes are A and E, and the intermediate nodes are C, D and B, wherein the value of the intermediate node C on the sub-dynamic graph is the largest, the first maximum value, and the minimum value of the value of the intermediate node on the sub-dynamic graph are both larger than the values ​​of the edge nodes A and E on the sub-dynamic graph, therefore, there is no first minimum value, and the intermediate node C corresponding to the first maximum value is taken as the benchmark, and there are two nodes in front, namely A and B, respectively, and it is obvious that the value of node A on the sub-dynamic graph is the smallest, and there are two nodes behind the intermediate node C, namely D and E, respectively, and it is obvious that the value of node E on the sub-dynamic graph is the smallest, therefore, the nodes ACE are connected to form an angle θ, and the angle θ is the characteristic angle of the corresponding sub-dynamic graph, and then the coordinate points of the positions of nodes A, C and E are extracted, the lengths of AC, AE and CE are determined, and the angle values ​​of the three angles of the corresponding triangle ACE are determined by combining trigonometric functions and inverse trigonometric functions, that is, the angle θ is obtained.

[0119] According to an embodiment of the present invention, it also includes:

[0120] If the total impact value of the initial feature value in the corresponding time period is less than the preset impact threshold, and the feature angle of the corresponding sub-dynamic graph is less than or equal to the preset feature angle threshold, the preset impact threshold revision information in the corresponding time period is generated;

[0121] Based on the preset impact threshold modification information within the corresponding time period, the preset impact base value is subtracted from the corresponding preset impact threshold to obtain the revised preset impact value within the corresponding time period;

[0122] If the total impact value of the initial feature value in the corresponding time period is greater than or equal to the revised preset impact threshold, the feature corresponding to the initial feature value is set as the main feature;

[0123] If the total impact value of the initial characteristic value within the corresponding time period is less than the revised preset impact threshold, the revised preset impact threshold is further revised according to the preset impact base value until the main features of the corresponding sub-dynamic graph are determined.

[0124] It should be noted that the smaller the feature angle, the greater the change in user transaction data within the corresponding time period, and the more obvious the characteristics of the user transaction data within the corresponding time period. When the feature angle of the corresponding sub-dynamic graph is less than or equal to the preset feature angle threshold and the total impact value of the initial feature value within the corresponding time period is less than the preset impact threshold, the preset impact threshold is revised by the preset impact base value. For example, if the preset impact base is 10, each time the preset impact threshold revision information for the corresponding time period is generated, the preset impact threshold for the corresponding time period is subtracted from the preset impact base to obtain the revised preset impact threshold.

[0125] Figure 4 A block diagram of a user transaction data feature extraction method system based on dynamic graphs of the present invention is shown.

[0126] like Figure 4 As shown, the second aspect of the present invention provides a user transaction data feature extraction system 4 based on dynamic graphs, including a memory 41 and a processor 42, wherein the memory stores a user transaction data feature extraction method program based on dynamic graphs as described in any one of the above.

[0127] The present invention discloses a method and system for extracting user transaction data features based on dynamic graphs. The method divides the dynamic graph into multiple sub-dynamic graphs, collects user transaction data fed back by the sub-dynamic graphs, forms an initial feature set, processes the data in the initial feature set, determines the main features of the sub-dynamic graphs, and determines the number of occurrences of the main features of the sub-dynamic graphs by presetting a first quantity threshold, thereby improving the accuracy of extracting user transaction data features from the dynamic graphs.

[0128] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0129] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0130] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0131] A person of ordinary skill in the art can understand that: all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: mobile storage devices, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks or optical disks, and other media that can store program codes.

[0132] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

Claims

1. A method for extracting user transaction data features based on dynamic graphs, characterized in that: include: Get user transaction data information of dynamic graph; Based on a preset time period, user transaction data of the dynamic graph is segmented and processed to obtain user transaction data information of sub-dynamic graphs within multiple time periods; Extracting an initial feature set of user transaction data of a sub-dynamic graph within a time period; According to the initial feature set of user transaction data of the sub-dynamic graph within the time period, the main features of the corresponding sub-dynamic graph are obtained; Based on n consecutive time periods, if the main features of the sub-dynamic images are the same, the number of occurrences of the corresponding same main features is accumulated to obtain the number of main features of the sub-dynamic images in the n consecutive time periods; If the number of times of the main feature of the sub-dynamic graph in n consecutive time periods is greater than a preset first quantity threshold, the main feature of the sub-dynamic graph in the n consecutive time periods is set as the main feature of the user transaction data of the dynamic graph; The step of obtaining the main features of the corresponding sub-dynamic graph according to the initial feature set of user transaction data of the sub-dynamic graph within the time period specifically includes: Extracting initial eigenvalues ​​from the initial eigenvalue set; According to the preset eigenvalue interval into which the initial eigenvalue in the initial feature set falls, an influence weight coefficient corresponding to the initial eigenvalue is obtained; Multiply the initial eigenvalue by the influence weight coefficient corresponding to the initial eigenvalue to obtain the influence value corresponding to the initial eigenvalue; Accumulate the influence values ​​corresponding to the initial eigenvalues ​​in the initial feature set to obtain the total influence value of the initial eigenvalues ​​in the corresponding time period; If the total impact value of the initial feature value within the corresponding time period is greater than or equal to the preset impact threshold, the feature corresponding to the initial feature value is set as the main feature.

2. The method for extracting user transaction data features based on dynamic graph according to claim 1, characterized in that: After extracting the initial feature set of user transaction data of the sub-dynamic graph within the time period, the method further includes: Extract features in the initial feature set and the corresponding initial feature value set; Extract the initial eigenvalues ​​from the initial eigenvalue set. If the initial eigenvalues ​​are the same, merge the corresponding initial eigenvalues ​​to obtain the merged initial eigenvalues. Determine whether the initial feature value is less than the preset first feature threshold of the corresponding feature. If so, delete the corresponding initial feature value; if not, store the corresponding initial feature value.

3. The method for extracting user transaction data features based on dynamic graph according to claim 2, characterized in that: If the initial eigenvalues ​​are the same, the step of fusing the corresponding initial eigenvalues ​​specifically includes: Get the quantity values ​​corresponding to the same initial eigenvalues; Obtain the transaction price difference corresponding to the same initial characteristic value, and set it as the first transaction price difference; obtain the transaction price difference within the time period corresponding to the same initial characteristic value, and set it as the second transaction price difference; Calculate the average value of the first transaction price difference to obtain the first transaction price average value; Divide the average value of the first transaction price by the difference of the second transaction price to obtain the fusion coefficient corresponding to the initial eigenvalue; Multiply the same initial eigenvalue by the quantity value corresponding to the same initial eigenvalue by the fusion coefficient to obtain the fusion value corresponding to the same initial eigenvalue; The fused values ​​corresponding to the same initial eigenvalue and the same initial eigenvalue are accumulated to obtain the fused initial eigenvalue.

4. The method for extracting user transaction data features based on dynamic graph according to claim 3, characterized in that: The step of obtaining the transaction price difference corresponding to the same initial characteristic value and the step of obtaining the transaction price difference within the time period corresponding to the same initial characteristic value specifically include: Obtain the transaction mark price and actual transaction price at the corresponding transaction time node for the same initial feature value; Calculate the difference between the marked transaction price and the actual transaction price, and take the absolute value to obtain the corresponding first transaction price difference; Obtain the user transaction price set within the time period corresponding to the same initial characteristic value; Extracting the lowest price and the highest price in the user transaction price set; Subtract the lowest price from the highest price to get the transaction difference corresponding to the second price.

5. The method for extracting user transaction data features based on dynamic graph according to claim 1, further comprising: Get the format information of the dynamic image; Determine whether the format of the dynamic image is the same as the preset format, and if so, store the corresponding dynamic image; If not, based on a preset format conversion algorithm, the format of the corresponding dynamic image is converted to the same as the preset format.

6. The method for extracting user transaction data features based on dynamic graph according to claim 1, further comprising: Based on a preset time period, extract m nodes on the sub-dynamic graph, where m is an odd number greater than 1; Divide the nodes into intermediate nodes and edge nodes according to their positions on the sub-dynamic graph; Extract the value of the node on the sub-dynamic graph; Compare and analyze the values ​​of the nodes on the sub-dynamic graph. If the value of the intermediate node on the sub-dynamic graph is greater than the values ​​of all the edge nodes on the sub-dynamic graph or less than the values ​​of all the edge nodes on the sub-dynamic graph, calculate the characteristic angle of the corresponding sub-dynamic graph; If the value of the intermediate node on the sub-dynamic graph is greater than or less than the value of the edge node on the sub-dynamic graph, then calculate the slope of the connecting line of the corresponding edge node; If the characteristic angle of the corresponding sub-dynamic graph is less than or equal to the preset characteristic angle threshold or the slope of the connecting line of the corresponding edge node is greater than the preset slope threshold, it is obtained that the main characteristic information exists in the corresponding time period.

7. According to the method for extracting user transaction data features based on dynamic graphs in claim 6, the step of calculating the feature angle of the corresponding sub-dynamic graph specifically comprises: Extract the minimum value or the maximum value of the values ​​of the intermediate nodes on the sub-dynamic graph, and set them as the first minimum value or the first maximum value respectively; When the value of the middle node on the sub-dynamic graph is the first maximum value, the middle node corresponding to the first maximum value is used as a reference, and the nodes corresponding to the second minimum values ​​on both sides of the middle node corresponding to the first maximum value are respectively extracted and connected to obtain the characteristic angle of the corresponding sub-dynamic graph; When the value of the middle node on the sub-dynamic graph is the first minimum value, the middle node corresponding to the first minimum value is used as a reference, and the nodes corresponding to the second maximum values ​​on both sides of the middle node corresponding to the first minimum value are respectively extracted and connected to obtain the characteristic angle of the corresponding sub-dynamic graph; The second minimum value is the minimum value of the values ​​of all nodes in front of the intermediate node corresponding to the first maximum value on the sub-dynamic graph and the minimum value of the values ​​of all nodes behind the intermediate node corresponding to the first maximum value on the sub-dynamic graph; the second maximum value is the maximum value of the values ​​of all nodes in front of the intermediate node corresponding to the first minimum value on the sub-dynamic graph and the maximum value of the values ​​of all nodes behind the intermediate node corresponding to the first minimum value on the sub-dynamic graph.

8. The method for extracting user transaction data features based on dynamic graph according to claim 7, further comprising: If the total impact value of the initial feature value in the corresponding time period is less than the preset impact threshold, and the feature angle of the corresponding sub-dynamic graph is less than or equal to the preset feature angle threshold, the preset impact threshold revision information in the corresponding time period is generated; Based on the preset impact threshold revision information within the corresponding time period, subtract the preset impact base value from the corresponding preset impact threshold to obtain the revised preset impact threshold within the corresponding time period; If the total impact value of the initial feature value in the corresponding time period is greater than or equal to the revised preset impact threshold, the feature corresponding to the initial feature value is set as the main feature; If the total impact value of the initial characteristic value within the corresponding time period is less than the revised preset impact threshold, the revised preset impact threshold is further revised according to the preset impact base value until the main features of the corresponding sub-dynamic graph are determined.

9. A user transaction data feature extraction system based on dynamic graph, characterized in that: It comprises a memory and a processor, wherein the memory stores a user transaction data feature extraction method program based on a dynamic graph as claimed in any one of claims 1 to 8.

Images