A method, device and computer-readable storage medium for detecting flow rate

Abstract

The embodiment of the present application discloses a method, device and computer-readable storage medium for main flow detection. The embodiment of the present application obtains multiple target push information click records of the target traffic main identifier; generates target heat map information corresponding to each target push information position identifier based on the multiple target push information click records; calculates the first entropy value of the target click horizontal coordinate set corresponding to the horizontal axis distribution of each target push information position identifier and the second entropy value of the target click vertical coordinate set corresponding to the vertical axis distribution based on the target heat map information; counts the first target entropy value of the horizontal axis distribution corresponding to each target push information position identifier and the second target entropy value of the vertical axis distribution of multiple traffic main identifiers; determines the detection result of the target traffic main identifier based on the first relationship between each first entropy value and the corresponding first target entropy value and the second relationship between each second entropy value and the corresponding second target entropy value. The efficiency of main flow detection is greatly improved.

Description

Technical Field

[0001] The present application relates to the field of information detection technology, and in particular to a flow detection method, device, and computer-readable storage medium. Background Art

[0002] With the development of the internet and the widespread use of computers, the online push market for information has expanded rapidly. Paid push notification providers pay publishers (push platforms) to distribute their push notifications through web pages, search engines, social media platforms, browsers, or other online platforms, thereby promoting their products. Currently, one of the mainstream methods for charging for push notifications is the cost-per-click (CPC) model.

[0003] In the existing technology, under the CPC push mode, the push host only needs to pay for the user's behavior of clicking on the push information, and does not need to pay for the exposure of the push information, thereby avoiding the risk of exposure without clicks. Since the push host needs to pay the publisher once every time a user clicks on the push information, the push host hopes that each paid push information click is a valid click by a real user rather than a cheating click (also known as a "malicious click"). In addition, traffic hosts who provide carriers of user traffic, such as media hosts, website hosts, software hosts, or public accounts with a certain number of fans, can participate in the profit commission of push information. Under the same push information exposure, the higher the click-through rate, the higher the profit shared by the traffic host. Therefore, the traffic host has a strong motivation to cheat to increase the click-through rate of push information.

[0004] During the research and practice of the existing technology, the inventors of the present invention found that although there are anti-cheating methods in the existing technology to determine whether a user is cheating, whether it is to judge cheating through some result-based indicators or to analyze the underlying code to determine whether the computer device has been implanted with malicious code and thus maliciously manipulated to cheat, it requires a lot of manual costs to detect whether the traffic owner is cheating, resulting in very low efficiency of traffic owner detection. Summary of the Invention

[0005] The embodiments of the present application provide a method, device, and computer-readable storage medium for traffic master detection, which can improve the efficiency of traffic master detection.

[0006] To solve the above technical problems, the embodiments of the present application provide the following technical solutions:

[0007] A main flow detection method, comprising:

[0008] Acquire multiple target push information click records of the target traffic main identifier, wherein the target push information click record includes at least a target push information position identifier, a click horizontal coordinate, and a click vertical coordinate;

[0009] Generate target heat map information corresponding to each target push information position identifier based on multiple target push information click records, wherein the target heat map information includes at least the target push information position identifier, the target click horizontal coordinate set, and the target click vertical coordinate set;

[0010] Based on the target heat map information, calculate the first entropy value of the target click abscissa set corresponding to the horizontal axis distribution of each target push information position identifier, and the second entropy value of the target click ordinate set corresponding to the vertical axis distribution;

[0011] Counting the first target entropy value of the horizontal axis corresponding to each target push information bit identifier of multiple traffic main identifiers and the second target entropy value of the vertical axis distribution;

[0012] The detection result of the target traffic primary identifier is determined according to a first relationship between each first entropy value and the corresponding first target entropy value and a second relationship between each second entropy value and the corresponding second target entropy value.

[0013] A flow detection device, comprising:

[0014] an acquiring unit, configured to acquire a plurality of target push information click records of a target traffic primary identifier, wherein the target push information click record comprises at least a target push information position identifier, a click horizontal coordinate, and a click vertical coordinate;

[0015] A generating unit, configured to generate target heat map information corresponding to each target push information position identifier based on a plurality of target push information click records, wherein the target heat map information includes at least the target push information position identifier, a target click abscissa set, and a target click ordinate set;

[0016] A calculation unit, configured to calculate, based on the target heat map information, a first entropy value of the target click abscissa set corresponding to the horizontal axis distribution of each target push information position identifier, and a second entropy value of the target click ordinate set corresponding to the vertical axis distribution;

[0017] A statistical unit, configured to count a first target entropy value of each target push information bit identifier of a plurality of traffic main identifiers distributed on the horizontal axis, and a second target entropy value of the vertical axis;

[0018] A determination unit is used to determine the detection result of the target traffic main identifier based on a first relationship between each first entropy value and the corresponding first target entropy value and a second relationship between each second entropy value and the corresponding second target entropy value.

[0019] In some embodiments, the determining unit includes:

[0020] The first calculation subunit is used to calculate the ratio of the first entropy value under each target push information bit identifier and the corresponding first target entropy value to obtain multiple first entropy value coefficients;

[0021] The second calculation subunit is used to calculate the ratio of the second entropy value under each target push information bit identifier and the corresponding second target entropy value to obtain multiple second entropy value coefficients;

[0022] A determination subunit is used to determine the detection result of the target traffic main identifier based on the multiple first entropy coefficients and the multiple second entropy coefficients.

[0023] In some embodiments, the determining subunit is configured to:

[0024] Determine the entropy coefficient with the smaller coefficient value between the first entropy coefficient and the second entropy coefficient under each target push information bit identifier as the first target entropy coefficient;

[0025] Count the click rate corresponding to each target push information position identifier;

[0026] Each first target entropy coefficient is weighted and summed according to the corresponding click rate to obtain the corresponding second target entropy coefficient;

[0027] When it is detected that the second target entropy coefficient is less than the first preset entropy coefficient, determining that the detection result of the target flow main identifier is an abnormal state;

[0028] When it is detected that the second target entropy coefficient is not less than the first preset entropy coefficient, it is determined that the detection result of the target flow main identifier is a non-abnormal state.

[0029] In some embodiments, the determining subunit is further configured to:

[0030] Counting a first number of first entropy coefficients smaller than a second preset entropy coefficient and a second number of second entropy coefficients smaller than the second preset entropy coefficient;

[0031] summing the first quantity and the second quantity to obtain a target quantity;

[0032] When it is detected that the target quantity is greater than a preset quantity threshold, determining that the detection result of the target flow main identifier is an abnormal state;

[0033] When it is detected that the target quantity is not greater than a preset quantity threshold, it is determined that the detection result of the target flow main identifier is a non-abnormal state.

[0034] In some embodiments, the acquiring unit is configured to:

[0035] Obtain multiple click records of the target traffic master corresponding to the target traffic master identifier within a preset time period;

[0036] The target push information bit identifier, the click horizontal coordinate, and the click vertical coordinate in each click record are extracted to generate multiple target push information click records.

[0037] In some embodiments, the target push information click record further includes the push information width and the push information height, and the generating unit is configured to:

[0038] Normalize each click horizontal coordinate and the corresponding push information width to obtain multiple target click horizontal coordinates;

[0039] Normalize each click vertical coordinate and the corresponding push information height to obtain multiple target click vertical coordinates;

[0040] Determine the target click abscissa set and the target click ordinate set corresponding to each target push information bit identifier;

[0041] A plurality of target heat map information is generated according to each target push information position identifier, the corresponding target click horizontal coordinate set and the target click vertical coordinate set.

[0042] In some embodiments, the computing unit is configured to:

[0043] Obtain first distribution information of the target click abscissa set corresponding to each target push information bit identifier;

[0044] Calculate the first entropy value corresponding to each first distribution information by using the information entropy formula;

[0045] Obtain second distribution information of the target click vertical coordinate set corresponding to each target push information bit identifier;

[0046] The second entropy value corresponding to each second distribution information is calculated using the information entropy formula.

[0047] In some embodiments, the statistics unit is configured to:

[0048] Acquire multiple push information click records of multiple traffic master identifiers, wherein the push information click records include at least a push information position identifier, a click horizontal coordinate, and a click vertical coordinate, and the push information position identifier includes a target push information position identifier;

[0049] Generate statistical heat map information corresponding to each target push information position identifier based on multiple push information click records, wherein the statistical heat map information includes at least the target push information position identifier, the statistical click horizontal coordinate, and the statistical click vertical coordinate;

[0050] Based on the statistical heat map information, a first target entropy value of the horizontal axis distribution corresponding to the statistical click horizontal coordinate set of each target push information position identifier and a second target entropy value of the vertical axis distribution corresponding to the statistical click vertical coordinate set are calculated.

[0051] A computer-readable storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor to execute the steps in the above-mentioned main flow detection method.

[0052] A computer device includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor implements the steps in the above-mentioned main flow detection method when executing the computer program.

[0053] A computer program product or computer program includes computer instructions stored in a storage medium. A processor of a computer device reads the computer instructions from the storage medium and executes the computer instructions, causing the computer to perform the steps of the above-mentioned flow detection method.

[0054] The embodiment of the present application obtains multiple target push information click records of the target traffic master identifier; generates target heat map information corresponding to each target push information position identifier based on the multiple target push information click records; calculates the first entropy value of the target click horizontal coordinate set corresponding to the horizontal axis distribution of each target push information position identifier, and the second entropy value of the target click vertical coordinate set corresponding to the vertical axis distribution based on the target heat map information; counts the first target entropy value of the horizontal axis distribution corresponding to each target push information position identifier, and the second target entropy value of the vertical axis distribution of multiple traffic master identifiers; determines the detection result of the target traffic master identifier based on the first relationship between each first entropy value and the corresponding first target entropy value and the second relationship between each second entropy value and the corresponding second target entropy value. In this way, in an unsupervised mode, the detection result of the target traffic master is quickly determined by the relationship between the entropy value of the target traffic master and the entropy values ​​of multiple traffic masters of the overall market. Compared with the solution that requires manual detection of whether the traffic master is cheating, the efficiency of traffic master detection is greatly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0056] Figure 1 This is a schematic diagram of a scenario of a main flow detection system provided by an embodiment of the present application;

[0057] Figure 2a Schematic diagram of the flow chart of the main flow detection method provided in the embodiment of the present application;

[0058] Figure 2b This is a scenario diagram of the main flow detection method provided in an embodiment of the present application;

[0059] Figure 2c This is another scenario diagram of the main flow detection method provided by an embodiment of the present application;

[0060] Figure 2d This is a heat map diagram of the main flow detection method provided by the embodiment of the present application;

[0061] Figure 2e This is another heat map diagram of the main flow detection method provided by the embodiment of the present application;

[0062] Figure 2f This is another heat map diagram of the main flow detection method provided by the embodiment of the present application;

[0063] Figure 3a This is another flow chart of the main flow detection method provided by the embodiment of the present application;

[0064] Figure 3b Another flow chart of the main flow detection method provided in an embodiment of the present application;

[0065] Figure 4 It is a structural diagram of the main flow detection device provided in an embodiment of the present application;

[0066] Figure 5 It is a structural diagram of the server provided in an embodiment of the present application. DETAILED DESCRIPTION

[0067] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0068] Embodiments of the present application provide a method, device, and computer-readable storage medium for detecting traffic flow.

[0069] See also Figure 1 , Figure 1The scenario diagram of the traffic master detection system provided in the embodiment of the present application includes: a user client, an access layer server, a database and a real-time computing server (the traffic master detection system may include multiple user clients, and the specific number of user clients is not limited here). The user client and the access layer server, as well as the access layer server and the real-time computing server, can be connected through a communication network. The communication network may include a wireless network and a wired network, wherein the wireless network includes a combination of one or more of a wireless wide area network, a wireless local area network, a wireless metropolitan area network, and a wireless personal area network. The network includes network entities such as routers and gateways, which are not shown in the figure.

[0070] The database can be stored in an access layer server or a real-time computing server, or stored separately in a specific server, which is not specifically limited here. In one embodiment, the access layer server, database, and real-time computing server can be integrated into a single server. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms.

[0071] The user client can be installed in a terminal, which can be a smart phone, tablet computer, laptop computer, desktop computer, smart speaker, smart watch, car terminal, smart TV, etc. The user client can be a media client, browser client or instant messaging client, etc. The user client can exchange information with the server through the communication network. For example, when the user client detects that a user clicks on push information (such as an advertisement), the user client will upload the current click behavior data including but not limited to (current time, user ID, scene, user IP, traffic main ID, click coordinates, target push information bit identifier) ​​to the access layer server.

[0072] The access layer server will store the uploaded click behavior data (i.e., click records) in the database, and then request the real-time computing server to perform cheating detection.

[0073] The real-time computing server can obtain multiple target push information click records of the target traffic main identifier, and the target push information click records include at least the target push information position identifier, the click horizontal coordinate and the click vertical coordinate; generate target heat map information corresponding to each target push information position identifier based on the multiple target push information click records, and the target heat map information includes at least the target push information position identifier, the target click horizontal coordinate set and the target click vertical coordinate set; based on the target heat map information, calculate the first entropy value of the target click horizontal coordinate set corresponding to the horizontal axis distribution of each target push information position identifier, and the second entropy value of the target click vertical coordinate set corresponding to the vertical axis distribution; count the first target entropy value of the horizontal axis distribution corresponding to each target push information position identifier of multiple traffic main identifiers, and the second target entropy value of the vertical axis distribution; determine the detection result of the target traffic main identifier based on the first relationship between each first entropy value and the corresponding first target entropy value and the second relationship between each second entropy value and the corresponding second target entropy value.

[0074] It should be noted that Figure 1 The scenario diagram of the main flow detection system shown is only an example. The main flow detection system and scenario described in the embodiment of the present application are intended to more clearly illustrate the technical solution of the embodiment of the present application, and do not constitute a limitation on the technical solution provided by the embodiment of the present application. Ordinary technicians in this field can know that with the evolution of the main flow detection system and the emergence of new business scenarios, the technical solution provided by the embodiment of the present application is also applicable to similar technical problems.

[0075] The following are detailed descriptions of each.

[0076] An embodiment of the present application provides a method for detecting a primary flow rate. The embodiment of the present application is described by taking the execution of the primary flow rate detection method by a server as an example.

[0077] See also Figure 2a , Figure 2a : is a flow chart of a method for detecting a primary flow rate provided in an embodiment of the present application. The method for detecting a primary flow rate includes:

[0078] In step 101, multiple target push information click records of the target traffic main identifier are obtained.

[0079] Targeted push information refers to information about a product promoted by the pusher, paid by the push platform, through web pages, search engines, browsers, online media, official accounts, or mini-programs. This information can be advertisements, articles, or game links. The pusher can also be commonly understood as the advertiser, who is the user or service provider who pays for the placement of advertisements. Advertisers hope that every click on their paid ads is a valid click from a real user, not a fraudulent one.

[0080] For example, see Figure 2b As shown, Figure 2b A scenario diagram of the traffic master detection method provided in an embodiment of the present application, where the instant messaging client A may include a public account B and a mini-program C. The public account B and the mini-program C may apply to open a push information position, which can be understood as opening an advertising position to become a traffic master. When a user clicks on the push information exposed on the public account B and the mini-program C, the traffic master can obtain a share of the income.

[0081] The push information bit can include multiple types, each of which has a unique push information bit identifier (id), see Figure 2c As shown, Figure 2c This is another scenario diagram of the main flow detection method provided in an embodiment of the present application. Figure 2c It includes interface 11, interface 12 and interface 13. Interface 11 displays push information position 111, interface 12 displays push information position 121, and interface 13 displays push information position 131. The display position and push information content of each push information position are different, and each push information position has a different push information position identifier.

[0082] Because in the CPC model, the more clicks the push information exposed in the push information position receives, the more revenue the traffic owner will get. Therefore, some traffic owners will cheat to create invalid click exposures to increase advertising revenue, but the advertising effect is very poor, which will damage the interests of advertisers (advertisers). Cheating seriously affects the effect of advertising, causing advertisers to stop or abandon advertising, and even cause public relations risks.

[0083] In order to avoid the above problems, the embodiment of the present application can obtain multiple target push information click records of the target traffic master corresponding to the target traffic master identifier (that is, the specific traffic master that needs to be detected) within a period of time. The target push information click record is the data extracted from the click record generated by any user client clicking on the target push information exposed on the push information position on the carrier that provides traffic for browsing the target traffic master.

[0084] The target push information click record includes at least the target push information position identifier, the click horizontal coordinate and the click vertical coordinate. The click horizontal coordinate is the horizontal coordinate on the screen when the user clicks the push information, and the click vertical coordinate is the vertical coordinate on the screen when the user clicks the push information.

[0085] It is understandable that in the embodiments of the present application, when it comes to data related to target push information click records, when the above embodiments of the present application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions.

[0086] The click patterns of the horizontal and vertical axes are key factors in determining whether the target traffic owner is cheating. To better illustrate the embodiments of this application, we first introduce the concept of automated cheating ads. This refers to the use of automated scripts or software to manipulate dozens or even hundreds of mobile phones through one or more computers, thereby manipulating non-normal users to click on ads.

[0087] Normal users' clicks are not controlled by rules. The click behaviors of the two are different as follows:

[0088] (1) The horizontal and vertical coordinates of normal user clicks are relatively scattered. Since the display position of the push information under a traffic master is randomly generated, the coordinates of normal user clicks must be relatively scattered.

[0089] (2) Normal users are more likely to have deep conversions (i.e., subsequent downloads and payment behaviors): The motivation for script / machine cheating comes from the small amount of income from pushing information. Although the income is small, it is easy to copy it to multiple traffic owners of the group at zero cost. Therefore, the script developer has no motivation to let each machine account complete the high-cost deep conversion such as payment after completing the cheating click. This will not be worth the loss. That is, after the script / machine cheats and clicks on the push information, there will be no next step of behavior, such as clicking on the promotional link in the push information, paying, etc.

[0090] In short, script / machine-type ad clicks cannot predict the random push of information materials, so the click position cannot be changed. Clicks are often made according to the settings of the script maker, and the heat map information of clicks will show obvious patterns. Since all machine cheating users share a cheating script, they are likely to present a click pattern such as a straight line, a rectangle, or several points.

[0091] For example, please also see Figure 2d 、 2e As shown in 2f, Figure 2d This is a heat map diagram of the main flow detection method provided in the embodiment of the present application. Figure 2e Another heat map diagram of the main flow detection method provided in the embodiment of the present application is shown. Figure 2f Another heat map diagram of the main traffic detection method provided in an embodiment of the present application. In this heat map information, the depth of color represents the frequency of clicks. The higher the click frequency, the darker the color, and the lower the click frequency, the lighter the color. The heat map information can reflect the pattern of clicks.

[0092] like Figure 2d As shown, in the point-type cheating script method, the heat map information D is displayed as a point shape, indicating that the pattern of cheating clicks is repeated clicks within a point range.

[0093] like Figure 2e As shown, in the straight line cheating script method, the heat map information E is displayed as a straight line shape, indicating that the pattern of cheating clicks is repeated clicks within a straight line range.

[0094] like Figure 2f As shown, in the rectangular cheating script method, the heat map information D is displayed as a rectangular shape, indicating that the pattern of cheating clicks is repeated clicks within a rectangular range.

[0095] From the above, we can see that the coordinates of cheating clicks show a certain regularity.

[0096] In some implementations, the step of obtaining multiple target push information click records of the target traffic primary identifier includes:

[0097] (1.1) Obtain multiple click records of the target traffic master corresponding to the target traffic master identifier within a preset time period;

[0098] (1.2) Extract the target push information bit identifier, click horizontal coordinate, and click vertical coordinate from each click record to generate multiple target push information click records.

[0099] Among them, multiple click records of the target traffic master corresponding to the target traffic master identifier within seven days or thirty days can be extracted. The click record is generated when the client clicks on the push information at the push information position exposed by the target traffic master.

[0100] In one embodiment, the click record does not include a preset conversion click event, which may be a deep conversion click event. Since cheating clicks do not include preset conversion click events, click records that do not include preset conversion click events can more accurately judge cheating clicks.

[0101] Furthermore, when the user client clicks on the push information on the push information identification position exposed on the carrier of the target traffic main identifier provided by the target traffic main identifier, the current time of the click, user identifier (ID), scene, user IP, traffic main ID, click coordinates and target push information position identifier and other data corresponding to the click will be formed into a click record and uploaded to the server. Therefore, the target push information position identifier, click horizontal coordinate and click vertical coordinate in each click record can be directly extracted to generate multiple target push information click records. Each target push information click record contains at least the target push information position identifier, click horizontal coordinate and click vertical coordinate. The display unit of the click horizontal coordinate and click vertical coordinate can be pixels (PIXEL). For example, the target push information click record can be:

[0102] (Traffic main ID, target push information bit identifier 1, click on the horizontal axis X1, click on the vertical axis Y1).

[0103] (Traffic main ID, target push information bit identifier 2, click on the horizontal axis X2, click on the vertical axis Y2).

[0104] (Traffic master ID, target push information bit identifier n, click on the horizontal axis Xn, click on the vertical axis Yn).

[0105] The examples given here are not intended to be a specific limitation. In one embodiment, the target push information may further include data such as push information width, push information height, and number of clicks.

[0106] In step 102, target heat map information corresponding to each target push information bit identifier is generated based on multiple target push information click records.

[0107] Among them, in order to better analyze the rules of the target push information click records, it is necessary to convert the target push information click records of the target traffic main identifier into heat map information. In actual application scenarios, due to the different push information position identifiers corresponding to the push information positions (i.e., advertising positions) display methods and positions are different. For example, the push information position corresponding to the push information position identifier J is displayed at the top of the official account, and the display size is 10 pixels square. The push information position corresponding to the push information position identifier K is displayed at the bottom of the official account, and the display size is 15 pixels square.

[0108] Therefore, the click patterns on the push information positions corresponding to different push information position identifiers may be different, and it is necessary to generate target heat map information corresponding to the click coordinates of each target push information position identifier and perform click pattern analysis respectively.

[0109] In one embodiment, based on multiple target push information click records, the target push information position identifier is used as a limiting dimension, and the click horizontal coordinate and click vertical coordinate under each target push information position identifier are aggregated to obtain the target click horizontal coordinate set and the target click vertical coordinate set under each target push information position identifier. In this way, target heat map information corresponding to each target push information position identifier is generated. The target heat map information includes at least the target push information position identifier, the target click horizontal coordinate set, and the target click vertical coordinate set. For example, the target heat map information can be expressed as:

[0110] (target push information bit identifier k, target click horizontal coordinate set Xi, target click vertical coordinate Yi).

[0111] In another embodiment, the target heat map information may also include information such as the click ratio PVi of each click behavior.

[0112] In some embodiments, the target heat map information corresponding to each target push information bit identifier is generated based on multiple target push information click records, including:

[0113] (1) Normalize each click horizontal coordinate and the corresponding push information width to obtain multiple target horizontal coordinates;

[0114] (2) normalizing each click vertical coordinate and the corresponding push information height to obtain multiple target vertical coordinates;

[0115] (3) determining the target horizontal coordinate set and target vertical coordinate set corresponding to each target push information bit identifier;

[0116] (4) Generate multiple target heat map information based on each target push information bit identifier, the corresponding target horizontal coordinate and the target vertical coordinate.

[0117] Among them, since the sizes of display screens of different terminals are different, the push information at the same push information position provided by the same traffic owner is displayed in different sizes on the display screens of different terminals. Therefore, even if the same script / machine performs cheating clicks in the same place, the horizontal and vertical coordinates of the clicks formed on different terminals are different.

[0118] However, the size ratio of the push information in the same push information position is the same. Therefore, in order to implement the subsequent click pattern analysis, it is necessary to normalize each click horizontal coordinate and the corresponding push information width to obtain multiple target click horizontal coordinates. The push information width is the width of the push information displayed on the display screen of the corresponding terminal, and the unit can be pixels. The specific normalization method can refer to the following formula:

[0119] X=X_raw / W*100

[0120] X_raw is the horizontal coordinate of the click, W is the width of the push information, and X is the horizontal coordinate of the target click after conversion. Through the above formula, each horizontal coordinate of the click and the corresponding push information width can be normalized to obtain multiple target horizontal coordinates of the click, and the target horizontal coordinates of the click can be normalized to between 0 and 100.

[0121] Each click vertical coordinate and the corresponding push information height need to be normalized to obtain multiple target click vertical coordinates. The push information height is the height of the push information displayed on the display screen of the corresponding terminal. The unit can be pixels. The specific normalization method can refer to the following formula:

[0122] Y=Y_raw / H*100

[0123] Y_raw is the vertical coordinate of the click, H is the height of the push information, and Y is the vertical coordinate of the target click after conversion. Through the above formula, each vertical coordinate of the click and the corresponding push information height can be normalized to obtain multiple target click vertical coordinates, and the target click vertical coordinates can be normalized to between 0 and 100.

[0124] By normalizing the horizontal and vertical coordinates of clicks on the display screens of different terminals through the above operations, the regularity of clicks can be better analyzed. In this way, the target horizontal coordinate set and target vertical coordinate set corresponding to each target push information position identifier can be determined. Furthermore, based on the dimension of the target push information position identifier, target heat map information of each target push information position identifier, the corresponding target horizontal coordinate set and the target vertical coordinate set are generated, and multiple target heat map information is obtained, namely:

[0125] (target push information bit identifier k, target click horizontal coordinate set Xi, target click vertical coordinate Yi).

[0126] In step 103, based on the target heat map information, the first entropy value of the target click horizontal coordinate set corresponding to the horizontal axis distribution of each target push information position identifier and the second entropy value of the target click vertical coordinate set corresponding to the vertical axis distribution are calculated.

[0127] It should be noted that in information theory, entropy is a measure of uncertainty. The more concentrated the information, the smaller the uncertainty, and the smaller the entropy; the greater the uncertainty, the more chaotic the information, and the larger the entropy. Corresponding to the embodiment of the present application, when a normal user clicks on the pushed information, the click position is uncertain, the information is chaotic, and the entropy is larger.

[0128] When a script / machine cheats, the click location is controlled by the script internally and has a certain regularity. Therefore, the smaller the uncertainty, the smaller the entropy value.

[0129] Based on the above theory, the first entropy value corresponding to the coordinate distribution on the horizontal axis of the target click horizontal coordinate set corresponding to each target push information position identifier can be calculated based on all target heat map information. The calculation method of the first entropy value can be based on the information entropy formula, which is as follows:

[0130]

[0131] The b is the base used for logarithms, which can be 2, the natural constant e or 10. In the embodiment of the present application, the b is 2, the H(X) is the information entropy, the unit is bit, and the P(x i ) represents a random click coordinate x is x iThe probability of each target push information bit identifier corresponding to the target horizontal coordinate set can be obtained by this method, that is, the distribution information of the number of times each target horizontal coordinate appears. For example, the target horizontal coordinate set of a certain target push information bit identifier contains target horizontal coordinates x1, x2, and x3, x1 appears 500 times, x2 appears 250 times, and x3 appears 250 times. Then the first distribution information corresponding to the target horizontal coordinate set is (500 times, 250 times, 250 times). Substituting this first distribution information into the above information entropy formula, it can be calculated that the information entropy is 1.5. Similarly, the first entropy value of the target click horizontal coordinate set corresponding to each push information bit identifier can be calculated in sequence.

[0132] Similarly, we can also obtain the second distribution information of the target click vertical coordinate set corresponding to each target push information bit identifier, and calculate the second entropy value corresponding to each second distribution information using the information entropy formula to obtain the second entropy value of the target click vertical coordinate set corresponding to each target push information bit identifier. The first entropy value can reflect the click pattern of the click horizontal coordinate. The magnitude of the second entropy value can reflect the click pattern of the click vertical coordinate.

[0133] In step 104, the first target entropy value of the horizontal axis corresponding to each target push information bit identifier of the plurality of traffic main identifiers and the second target entropy value of the vertical axis are counted.

[0134] Among them, after obtaining the first entropy value and the second entropy value of each target push information bit identifier, since the first entropy value and the second entropy value can only represent the click pattern of each target push information bit identifier under the single target traffic master.

[0135] Since there is no reference entropy value, at this time, it is not known whether the first entropy value and the second entropy value are reasonable values. Therefore, it is necessary to obtain the entire market, that is, the click patterns of a large number of traffic masters for comparison, that is, the click patterns of traffic masters with multiple traffic master identifiers are required for comparison. The multiple traffic master identifiers are a large number of traffic master identifiers stored in the server. The target traffic master identifier can be included in the multiple traffic master identifiers. Each traffic master identifier contains multiple push information click records, so the click patterns formed by multiple traffic masters are very scattered, the uncertainty of the click position will be very large, and the entropy value will be correspondingly large.

[0136] In this way, multiple push information click records of multiple traffic owners can be obtained. The number of push information click records is much larger than the number of target push information click records. The push information click record at least includes the push information position identifier, the click horizontal coordinate and the click vertical coordinate. It should be noted that the data volume of the push information click position identifier is much larger than the data volume of the target push information position identifier, and the push information click position identifier includes the target push information position identifier.

[0137] Furthermore, based on multiple push information click records, a first target entropy value corresponding to the coordinate distribution of all traffic master identifiers on the horizontal axis corresponding to each target push information position identifier, and a second target entropy value corresponding to the coordinate distribution on the vertical axis can be calculated. The first target entropy value can reflect the click pattern of all traffic masters clicking on the horizontal coordinate under each target push information position identifier, and the second target entropy value can reflect the click pattern of all traffic masters clicking on the vertical coordinate under each target push information position identifier.

[0138] In some embodiments, the statistics of the first target entropy value distributed on the horizontal axis corresponding to each target push information bit identifier of the plurality of traffic main identifiers and the second target entropy value distributed on the vertical axis include:

[0139] (1) Obtain multiple push information click records of multiple traffic master identifiers;

[0140] (2) generating statistical heat map information corresponding to each target push information position identifier based on the plurality of push information click records, wherein the statistical heat map information includes at least the target push information position identifier, the statistical click horizontal coordinate, and the statistical click vertical coordinate;

[0141] (3) Based on the statistical heat map information, calculate the first target entropy value of the horizontal axis distribution corresponding to the statistical click horizontal coordinate set of each target push information position identifier and the second target entropy value of the vertical axis distribution corresponding to the statistical click vertical coordinate set.

[0142] Among them, multiple push information click records of multiple traffic main identifiers in the server can be obtained first. The push information click record includes at least the push information position identifier, the click horizontal coordinate and the click vertical coordinate. The data volume of the push information click position identifier is much larger than the data volume of the target push information position identifier, that is, the push information click position identifier contains the target push information position identifier.

[0143] Furthermore, since what needs to be analyzed is the click pattern corresponding to the target push information position corresponding to the target push information position identifier, based on multiple push information click records, with the target push information position identifier as the limiting dimension, the click horizontal coordinates and click vertical coordinates of each target push information position identifier under multiple traffic main identifiers can be aggregated to obtain the statistical click horizontal coordinate set and statistical click vertical coordinate set of multiple traffic main identifiers under each target push information position identifier. Generate a statistical heat map corresponding to each target push information position identifier under multiple traffic main identifiers. The statistical heat map information includes at least the target push information position identifier, the statistical click horizontal coordinate and the statistical click vertical coordinate. For example, the target heat map information can be expressed as:

[0144] (Target push information bit identifier k, statistical click horizontal coordinate set Xm, target click vertical coordinate Ym).

[0145] In one embodiment, in order to achieve subsequent regularity comparison, the coordinates in the statistical click horizontal coordinate set and the statistical click vertical coordinate set can also be normalized according to the normalization processing method in the above embodiment, which is not described in detail here.

[0146] Based on all statistical heat map information, the first target entropy value of the coordinate distribution on the horizontal axis corresponding to the statistical click horizontal coordinate set of each target push information position identifier under multiple traffic main identifiers can be calculated respectively. The calculation method of the first target entropy value can be based on the information entropy formula.

[0147] Similarly, the second target entropy value of the coordinate distribution on the vertical axis corresponding to the statistical vertical coordinate set of each target push information bit identifier under multiple traffic main identifiers can also be calculated respectively. The calculation method of the second target entropy value can be based on the information entropy formula.

[0148] In step 105, a detection result of the target traffic primary identifier is determined based on a first relationship between each first entropy value and a corresponding first target entropy value and a second relationship between each second entropy value and a corresponding second target entropy value.

[0149] Among them, since the size of the entropy value reflects the click regularity, the smaller the entropy value, the smaller the uncertainty, the more regular the click position, the larger the entropy value, the greater the uncertainty, the more irregular the click position, therefore, it is possible to determine the first relationship between the first entropy value under each target push information bit identifier and the corresponding first target entropy value, and to determine the second relationship between the second entropy value under each target push information bit identifier and the corresponding second target entropy value, and this relationship can be the ratio between the two.

[0150] Since the first entropy value is the entropy value corresponding to the horizontal axis click pattern of a single target traffic main identifier under each target push information position identifier, and the first target entropy value is the entropy value corresponding to the horizontal axis click pattern of multiple traffic main identifiers under each target push information position identifier. The first target entropy value reflects the click pattern of the entire background on the horizontal axis. Therefore, the closer the first entropy value is to the first target entropy value, the more dispersed the click pattern on the horizontal axis reflected by the first entropy value is, and the greater the probability of being a normal traffic main. On the contrary, the less the first entropy value is to the first target entropy value, the more concentrated the click pattern on the horizontal axis reflected by the first entropy value is, and the greater the probability of script cheating.

[0151] Similarly, the second entropy value is the entropy value corresponding to the vertical click pattern of a single target traffic main identifier under each target push information position identifier, and the second target entropy value is the entropy value corresponding to the vertical click pattern of multiple traffic main identifiers under each target push information position identifier. The second target entropy value reflects the click pattern of the entire background on the vertical axis. Therefore, the closer the second entropy value is to the second target entropy value, the more dispersed the click pattern on the vertical axis reflected by the second entropy value is, and the greater the probability of being a normal traffic main. On the contrary, the less the second entropy value is to the second target entropy value, the more concentrated the click pattern on the vertical axis reflected by the second entropy value is, and the greater the probability of script cheating.

[0152] Based on this, the smaller the ratio between the first entropy value and the corresponding first target entropy value, the more concentrated the click pattern on the horizontal axis of the target push information position mark is, and the greater the probability of cheating clicks. Conversely, the larger the ratio between the first entropy value and the corresponding first target entropy value, the more dispersed the click pattern on the horizontal axis of the target push information position mark is, and the smaller the probability of cheating clicks is.

[0153] The smaller the ratio between the second entropy value and the corresponding second target entropy value, the more concentrated the click pattern on the vertical axis of the target push information position mark is, and the greater the probability of cheating clicks. On the contrary, the larger the ratio between the second entropy value and the corresponding second target entropy value, the more dispersed the click pattern on the vertical axis of the target push information position mark is, and the smaller the probability of cheating clicks is.

[0154] In this way, a preset threshold can be set, which serves as the critical value for distinguishing between cheating clicks and normal clicks. When the ratio is greater than or equal to the preset threshold, it is determined to be a normal click. Conversely, when the ratio is less than the preset threshold, it is determined to be a cheating click. In this way, the number of cheating clicks and the number of normal clicks are recorded. If the number of normal clicks is greater than or equal to the number of cheating clicks, the detection result of the target traffic main identifier is determined to be normal. If the number of normal clicks is less than the number of cheating clicks, the detection result of the target traffic main identifier is determined to be cheating.

[0155] In some embodiments, determining the detection result of the target traffic primary identifier according to a first relationship between each first entropy value and the corresponding first target entropy value and a second relationship between each second entropy value and the corresponding second target entropy value includes:

[0156] (1) calculating the ratio of the first entropy value under each target push information bit identifier to the corresponding first target entropy value to obtain multiple first entropy value coefficients;

[0157] (2) calculating the ratio of the second entropy value under each target push information bit identifier to the corresponding second target entropy value to obtain multiple second entropy value coefficients;

[0158] (3) Determine a detection result of the target traffic main identifier based on the multiple first entropy coefficients and the multiple second entropy coefficients.

[0159] Among them, the ratio of the first entropy value and the corresponding first target entropy value under each target push information bit identifier can be calculated, and the first entropy coefficient under each target push information bit identifier can be determined to obtain multiple first entropy coefficients. Each first entropy coefficient reflects the pattern of push information clicks in the horizontal axis direction of the target push information bit corresponding to the corresponding target push information bit identifier. The smaller the first entropy coefficient, the more concentrated the click pattern on the horizontal axis of the target push information bit identifier, and the greater the probability of cheating clicks. On the contrary, the larger the ratio between the first entropy value and the corresponding first target entropy value, the more dispersed the click pattern on the horizontal axis of the target push information bit identifier, and the smaller the probability of cheating clicks.

[0160] Similarly, the ratio of the second entropy value under each target push information bit identifier and the corresponding second target entropy value can be calculated to determine the second entropy coefficient under each target push information bit identifier, and obtain multiple second entropy coefficients. Each second entropy coefficient reflects the pattern of push information clicks on the target push information bit corresponding to the corresponding target push information bit identifier in the vertical axis direction. The smaller the second entropy coefficient, the more concentrated the click pattern on the vertical axis of the target push information bit identifier, and the greater the probability of cheating clicks. On the contrary, the larger the ratio between the second entropy value and the corresponding second target entropy value, the more dispersed the click pattern on the vertical axis of the target push information bit identifier, and the smaller the probability of cheating clicks.

[0161] In this way, a preset entropy coefficient can be set, which is the critical value for defining a cheating click or a normal click. When the first entropy coefficient is detected to be greater than or equal to the preset entropy coefficient, it is determined to be a normal click. On the contrary, when the first entropy coefficient is detected to be less than the preset entropy coefficient, it is determined to be a cheating click.

[0162] When it is detected that the second entropy coefficient is greater than or equal to the preset entropy coefficient, it is determined to be a normal click. On the contrary, when it is detected that the second entropy coefficient is less than the preset entropy coefficient, it is determined to be a cheating click.

[0163] Furthermore, the number of clicks determined to be cheating and the number of normal clicks can be recorded. If the number of normal clicks is greater than or equal to the number of cheating clicks, the detection result of the target traffic main identifier is determined to be normal. If the number of normal clicks is less than the number of cheating clicks, the detection result of the target traffic main identifier is determined to be cheating.

[0164] In some embodiments, the detection result of the target traffic primary identifier is determined based on the multiple first entropy coefficients and the multiple second entropy coefficients.

[0165] (1.1) counting a first number of first entropy coefficients smaller than a second preset entropy coefficient and a second number of second entropy coefficients smaller than the second preset entropy coefficient;

[0166] (1.2) summing the first quantity and the second quantity to obtain a target quantity;

[0167] (1.3) When it is detected that the target quantity is greater than a preset quantity threshold, determining that the detection result of the target flow main identifier is an abnormal state;

[0168] (1.4) When it is detected that the target quantity is not greater than the preset quantity threshold, it is determined that the detection result of the target flow main identifier is a non-abnormal state.

[0169] The second preset entropy coefficient is a critical value for defining a click as cheating or normal. When the first entropy coefficient is greater than or equal to the preset entropy coefficient, it is determined to be a normal click. Conversely, when the first entropy coefficient is less than the preset entropy coefficient, it is determined to be a cheating click. When the second entropy coefficient is greater than or equal to the preset entropy coefficient, it is determined to be a normal click. Conversely, when the second entropy coefficient is less than the preset entropy coefficient, it is determined to be a cheating click.

[0170] Based on this, the first number of first entropy coefficients that are smaller than the second preset entropy coefficient can be counted, for example, 8 times, and the second number of second entropy coefficients that are smaller than the second preset entropy coefficient can be counted, for example, 6 times, indicating that 8 cheating clicks have occurred on the horizontal coordinate of the click identified by the target push information position, and 6 cheating clicks have occurred on the vertical coordinate of the click.

[0171] The first number and the second number are summed to obtain a total target number of 14 times, which means that 14 cheating clicks occurred.

[0172] Furthermore, the preset quantity threshold is a critical value for determining whether the target traffic master corresponding to the target traffic master identifier is normal or abnormal. The preset quantity threshold can be half of the total number of coordinates corresponding to the target push information bit identifier or can be manually set. For example, the target push information bit identifier can contain 10 quantities, then the corresponding number of coordinates is 10 times 2, that is, 20, and half is 10. In other words, under this mechanism, when the target quantity detected is greater than 10, that is, more than half of the coordinates are abnormal, it can be determined that the detection result of the target traffic master corresponding to the target traffic master identifier is in an abnormal state. When the target quantity detected is not greater than the preset quantity threshold, that is, less than half of the coordinates are not abnormal, it can be determined that the detection result of the target traffic master corresponding to the target traffic master identifier is in a non-abnormal state.

[0173] In this way, the above embodiment can quickly determine the cheating status of the target traffic host corresponding to the target traffic host identifier, and appropriately control the target traffic host determined to be abnormal.

[0174] As can be seen from the above, the embodiment of the present application obtains multiple target push information click records of the target traffic master identifier; generates target heat map information corresponding to each target push information position identifier based on multiple target push information click records; calculates the first entropy value of the target click horizontal coordinate set corresponding to the horizontal axis distribution of each target push information position identifier, and the second entropy value of the target click vertical coordinate set corresponding to the vertical axis distribution based on the target heat map information; counts the first target entropy value of the horizontal axis distribution corresponding to each target push information position identifier, and the second target entropy value of the vertical axis distribution of multiple traffic master identifiers; determines the detection result of the target traffic master identifier based on the first relationship between each first entropy value and the corresponding first target entropy value and the second relationship between each second entropy value and the corresponding second target entropy value. In this way, in an unsupervised mode, the detection result of the target traffic master is quickly determined by the relationship between the entropy value of the target traffic master and the entropy values ​​of multiple traffic masters of the overall market. Compared with the solution that requires manual detection of whether the traffic master is cheating, the efficiency of traffic master detection is greatly improved.

[0175] In this embodiment, the main flow detection device will be integrated into a server as an example for description, and specific reference is made to the following description.

[0176] See also Figure 3a , Figure 3a This is another flow chart of the main flow detection method provided in an embodiment of the present application. The method flow may include:

[0177] In step 201, the server obtains multiple click records of the target traffic master corresponding to the target traffic master identifier within a preset time period.

[0178] In order to better illustrate the embodiment of the present application, the push information in the embodiment of the present application is described by taking the subordinate concept of advertisement as an example.

[0179] When a user client clicks on an advertisement on an advertisement space exposed on a carrier provided by a target traffic master identified by a target traffic master, the user client will upload the data corresponding to the click, such as the current time, user ID, scene, user IP, traffic master ID, click coordinates and target advertisement space identification, to the server as a click record. Each advertisement space corresponds to a unique advertisement space identification. For example, please continue to refer to Figure 2c As shown, ad slot 111, ad slot 121, and ad slot 131 have different ad slot identifiers. Therefore, the server can extract all click records of the target traffic master corresponding to the specific target traffic master identifier within fourteen days.

[0180] Furthermore, since cheating clicks will not include preset conversion click events, that is, deep conversion click events, such as payment or downloading after clicking on an advertisement, the click records including preset conversion click events are filtered from all click records to obtain multiple filtered click records, which include but are not limited to (current time, user identifier (ID), scene, user IP, traffic master ID, click coordinates, target ad position identifier, ad width, ad height), and the click coordinates include the horizontal coordinate of the click and the vertical coordinate of the click.

[0181] In step 202, the server extracts the target push information bit identifier, the click horizontal coordinate, and the click vertical coordinate in each click record to generate multiple target push information click records.

[0182] Among them, since the click pattern of the click horizontal coordinate and click vertical coordinate is the key factor for analyzing whether the target traffic owner is cheating. Therefore, the server can directly extract the target ad position identifier, the corresponding click horizontal coordinate, and the click vertical coordinate from each filtered click record to generate multiple target ad click records. Each target ad click record contains at least the target ad position identifier, click horizontal coordinate, and click vertical coordinate, and may also include the ad width, ad height, and number of clicks. The display unit of the click horizontal coordinate and click vertical coordinate can be pixels. For example, the target ad click record can be:

[0183] (Traffic master ID, target ad slot identifier 1, click horizontal coordinate X1, click vertical coordinate Y1, ad width W1, ad height H1, 1 click).

[0184] (Traffic master ID, target ad slot identifier 2, click on horizontal coordinate X2, click on vertical coordinate Y2, ad width W2, ad height H2, 1 click).

[0185] (Traffic master id, target ad slot identifier n, click horizontal coordinate Xn, click vertical coordinate Yn, ad width W2, ad height H2, 1 click).

[0186] In one embodiment, in order to achieve better subsequent statistical calculations, the target advertising click records with the same attributes can be aggregated to obtain (traffic master ID, target advertising position identifier k, click horizontal coordinate X-raw, click vertical coordinate Y-raw, advertising width W, advertising height H, N clicks).

[0187] In step 203, the server normalizes the horizontal coordinate of each click and the corresponding push information width to obtain multiple target click horizontal coordinates, normalizes each click vertical coordinate and the corresponding push information height to obtain multiple target click vertical coordinates, determines the target click horizontal coordinate set and target click vertical coordinate set corresponding to each target push information position identifier, and generates multiple target heat map information according to each target push information position identifier, the corresponding target click horizontal coordinate set and the target click vertical coordinate set.

[0188] Among them, since the sizes of the display screens of user clients on different terminals are different, the sizes of advertisements on the same advertising position provided by the same traffic owner are different on the display screens of different terminals. Therefore, even if the same script / machine performs cheating clicks in the same place, the horizontal and vertical coordinates of the clicks formed on different terminals are different.

[0189] However, the size ratios of ads in the same ad slot are the same. Therefore, in order to perform subsequent click pattern analysis, it is necessary to normalize each click horizontal coordinate and the corresponding ad width to obtain multiple target click horizontal coordinates. The ad width is the width of the ad displayed on the display screen of the corresponding terminal, and the unit can be pixels. The specific normalization method can refer to the following formula:

[0190] X=X_raw / W*100

[0191] X_raw is the horizontal coordinate of the click, W is the ad width, and X is the target click horizontal coordinate after conversion. Through the above formula, each click horizontal coordinate and the corresponding ad width can be normalized to obtain multiple target click horizontal coordinates, and the target click horizontal coordinates can be normalized to between 0 and 100.

[0192] Each click ordinate and the corresponding ad height need to be normalized to obtain multiple target click ordinates. The ad height is the height of the ad displayed on the display screen of the corresponding terminal, and the unit can be pixels. The specific normalization method can refer to the following formula:

[0193] Y=Y_raw / H*100

[0194] Y_raw is the vertical coordinate of the click, H is the ad height, and Y is the target click vertical coordinate after conversion. Through the above formula, each click vertical coordinate and the corresponding ad height can be normalized to obtain multiple target click vertical coordinates, and the target click vertical coordinates can be normalized to between 0 and 100.

[0195] By normalizing the horizontal and vertical coordinates of clicks on the display screens of different terminals through the above operations, the regularity of clicks can be better analyzed. In this way, the target horizontal and vertical coordinate sets of clicks corresponding to each target ad slot identifier can be determined. Furthermore, based on the dimension of the target ad slot identifier, target heat map information is generated for each target ad slot identifier, the corresponding target horizontal and vertical coordinate sets of clicks, and multiple target heat map information is obtained. The target heat map information can be expressed as:

[0196] (target ad position identifier k, target click abscissa set Xi, target click ordinate set Yi, click ratio PVi).

[0197] The k is the number of target ad slot identifiers, and the i is the number of coordinates.

[0198] The click ratio PVi is the click rate of each coordinate in the total number of clicks. For example, if the number of clicks for the coordinate (2, 3) is 3 and the total number of clicks is 30, then the click ratio PVi for the coordinate (2, 3) is 0.1.

[0199] In step 204, the server obtains the first distribution information of the target click horizontal coordinate set corresponding to each target push information bit identifier, calculates the first entropy value corresponding to each first distribution information through the information entropy formula, obtains the second distribution information of the target click vertical coordinate set corresponding to each target push information bit identifier, and calculates the second entropy value corresponding to each second distribution information through the information entropy formula.

[0200] It should be noted that when a normal user clicks on an ad, the click location is uncertain, the information is confusing, and the entropy value is greater.

[0201] When a script / machine cheats, the click location is controlled by the script internally and has a certain regularity. Therefore, the smaller the uncertainty, the smaller the entropy value.

[0202] Based on the above theory, the first entropy value corresponding to the coordinate distribution on the horizontal axis corresponding to the target click horizontal coordinate set of each target ad position can be calculated based on all target heat map information. The calculation method of the first entropy value can be based on the information entropy formula, which is as follows:

[0203]

[0204] The b is the base of the logarithm, which can be 2. The H(X) is the information entropy, in bits. The P(x i ) represents a random click coordinate x is x iThe probability of each target ad position identifier can be obtained by this method. The first distribution information of the target horizontal coordinate set corresponding to each target ad position identifier can be obtained. The first distribution information is the distribution information of the number of times each target horizontal coordinate appears. For example, the target horizontal coordinate set of a target ad position identifier contains target horizontal coordinates x1, x2, and x3 in sequence. The number of occurrences of x1 is 1000 times, the number of occurrences of x2 is 500 times, and the number of occurrences of x3 is 500 times. Then the first distribution information corresponding to the target horizontal coordinate set is (1000 times, 500 times, 500 times). Substituting the first distribution information into the above information entropy formula, we can get 0.5*1+0.25*2+0.25*2, and the calculated information entropy is 1.5. Based on this, the first entropy value of the target click horizontal coordinate set corresponding to each ad position identifier can be calculated in turn.

[0205] Similarly, the second distribution information of the target click ordinate set corresponding to each target ad slot identifier can be obtained, and the second entropy value corresponding to each second distribution information can be calculated using the information entropy formula to obtain the second entropy value of the target click ordinate set corresponding to each target ad slot identifier. The first entropy value can reflect the click pattern of the click abscissa. The magnitude of the second entropy value can reflect the click pattern of the click ordinate.

[0206] In step 205, the server obtains multiple push information click records of multiple traffic master identifiers.

[0207] Among them, after obtaining the first entropy value and the second entropy value of each target advertising position identification, since the first entropy value and the second entropy value can only represent the click pattern of the single target traffic owner on each advertising position identification.

[0208] Since there is no reference entropy value, at this time, it is unknown whether the first entropy value and the second entropy value are reasonable values. Therefore, it is also necessary to obtain the entire server background, that is, the click patterns of a large number of traffic masters on advertisements for comparison, that is, the click patterns of traffic masters with multiple traffic master identifiers on advertisements are required to be compared. The multiple traffic master identifiers are a large number of traffic master identifiers stored in the server. The target traffic master identifier can be included in the multiple traffic master identifiers, and each traffic master identifier contains corresponding multiple advertisement click records. Therefore, the click patterns formed by multiple traffic masters are very scattered, the uncertainty of the click position will be very large, and the entropy value will also be correspondingly large.

[0209] In this way, multiple ad click records for multiple traffic owners can be obtained. The number of ad click records is much larger than the number of target ad click records. The ad click records at least include the ad slot identifier, the horizontal coordinate of the click, and the vertical coordinate of the click. It should be noted that the data volume of the ad click slot identifier is much larger than the data volume of the target ad slot identifier, and the ad click slot identifier includes the target ad slot identifier. For example, the ad slot identifiers can include ad slot identifier 1, ad slot identifier 2, ad slot identifier 3, ad slot identifier 4, ad slot identifier 5, and ad slot identifier 6. The target ad slot identifiers can include ad slot identifier 1, ad slot identifier 2, and ad slot identifier 3. Therefore, it is obvious that the ad slot click slot identifier includes the target ad slot identifier.

[0210] In step 206, the server generates statistical heat map information corresponding to each target push information position identifier based on the multiple push information click records.

[0211] Among them, because the server needs to analyze the click pattern corresponding to the target ad position corresponding to the target ad position identifier, it can use the target ad position identifier as a limiting dimension based on multiple ad click records, and collect the click horizontal coordinates and click vertical coordinates of each target ad position identifier under multiple traffic main identifiers to obtain the statistical click horizontal coordinate set and statistical click vertical coordinate set of multiple traffic main identifiers under each target ad position identifier. Generate a statistical heat map corresponding to each target ad position identifier under multiple traffic main identifiers. The statistical heat map information includes at least the target ad position identifier, the statistical click horizontal coordinate and the statistical click vertical coordinate. The coordinates in the statistical click horizontal coordinate set and the statistical click vertical coordinate set are the coordinates after normalization according to the normalization processing method in the above steps. No specific details are given here. For example, the target heat map information can be expressed as:

[0212] (Target ad position identifier k, statistical click horizontal coordinate set Xm, target click vertical coordinate Ym).

[0213] The k is the number of target ad slot identifiers, and the m is the number of statistical click horizontal coordinates.

[0214] In step 207, the server calculates the first target entropy value of the horizontal axis distribution corresponding to the statistical click horizontal coordinate set of each target push information position identifier and the second target entropy value of the vertical axis distribution corresponding to the statistical click vertical coordinate set based on the statistical heat map information.

[0215] Among them, the server can calculate the first target entropy value of the coordinate distribution on the horizontal axis corresponding to the statistical click horizontal coordinate set of each target advertising position identifier under multiple traffic main identifiers based on all statistical heat map information. The calculation method of the first target entropy value can be based on the information entropy formula provided in the above steps.

[0216] Similarly, the second target entropy value of the coordinate distribution on the vertical axis corresponding to the statistical vertical coordinate set of each target advertising position identifier under multiple traffic main identifiers can also be calculated. The calculation method of the second target entropy value can be based on the information entropy formula provided in the above steps.

[0217] The first target entropy value can reflect the click pattern of all traffic owners on the horizontal axis of each target advertising position identification, and the second target entropy value can reflect the click pattern of all traffic owners on the vertical axis of each target advertising position identification.

[0218] In step 208, the server base calculates the ratio of the first entropy value under each target push information bit identifier and the corresponding first target entropy value to obtain multiple first entropy coefficients, and calculates the ratio of the second entropy value under each target push information bit identifier and the corresponding second target entropy value to obtain multiple second entropy coefficients.

[0219] Among them, the size of the entropy value reflects the click pattern. The smaller the entropy value, the smaller the uncertainty, and the more regular the click position. The larger the entropy value, the greater the uncertainty, and the more irregular the click position. Therefore, the ratio between the first entropy value under each target ad position identifier and the corresponding first target entropy value can be determined to obtain multiple first entropy coefficients. And the ratio between the second entropy value under each target ad position identifier and the corresponding second target entropy value can be determined to obtain multiple second entropy coefficients. For example, the first entropy coefficient and the second entropy coefficient can be determined by the following formula:

[0220] h_score(X)_k=H(X)_k / H_dp(X)_k

[0221] h_score(Y)_k=H(Y)_k / H_dp(Y)_k

[0222] h_score(X)_k is the first entropy coefficient of target ad slot identifier k, H(X)_k is the first entropy value of target ad slot identifier k, and H_dp(X)_k is the first target entropy value of target ad slot identifier k. h_score(Y)_k is the second entropy coefficient of target ad slot identifier k, H(Y)_k is the second entropy value of target ad slot identifier k, and H_dp(Y)_k is the second target entropy value of target ad slot identifier k.

[0223] Since the first entropy value is the entropy value corresponding to the horizontal click pattern of a single target traffic master identifier under each target ad slot identifier, and the first target entropy value is the entropy value corresponding to the horizontal click pattern of multiple traffic master identifiers under each target ad slot identifier, each first entropy coefficient reflects the pattern of ad clicks on the horizontal axis of the target ad slot corresponding to the corresponding target ad slot identifier. The smaller the first entropy coefficient, the more concentrated the click pattern on the horizontal axis of the target ad slot identifier, and the greater the probability of cheating clicks. Conversely, the larger the first entropy coefficient, the more dispersed the click pattern on the horizontal axis of the target ad slot identifier, and the lower the probability of cheating clicks.

[0224] Correspondingly, each second entropy coefficient reflects the pattern of advertisement clicks on the vertical axis of the target advertisement position corresponding to the corresponding target advertisement position identifier. The smaller the second entropy coefficient is, the more concentrated the click pattern on the vertical axis of the target advertisement position identifier is, and the greater the probability of cheating clicks is. On the contrary, the larger the ratio between the second entropy value and the corresponding second target entropy value is, the more dispersed the click pattern on the vertical axis of the target advertisement position identifier is, and the smaller the probability of cheating clicks is.

[0225] In step 209, the server determines the entropy coefficient with a smaller coefficient value between the first entropy coefficient and the second entropy coefficient under each target push information bit identifier as the first target entropy coefficient, counts the click rate corresponding to each target push information bit identifier, and weights and sums each first target entropy coefficient according to the corresponding click rate to obtain the corresponding second target entropy coefficient.

[0226] Among them, under the same target advertising position identification, there are a first entropy coefficient and a second entropy coefficient, which respectively represent the click pattern on the horizontal axis and the click pattern on the vertical axis under the target advertising position identification. The smaller the entropy coefficient, the greater the probability of cheating, and the larger the entropy coefficient, the lower the probability of cheating.

[0227] Based on this, the smaller the entropy coefficient value under each target ad position identifier, the greater the probability of cheating. For more accurate and efficient calculations, the server can compare the first entropy coefficient and the second entropy coefficient under each target ad position identifier, and determine the smaller entropy coefficient between the first entropy coefficient and the second entropy coefficient under the same target ad position identifier as the first target entropy coefficient, that is, the entropy coefficient with a greater probability of cheating under the same target ad position identifier is determined as the first target entropy coefficient. To better illustrate the embodiment of the present application, the following formula can be referred to:

[0228] hscore(k)=min(h_score(X)_k, h_score(Y)_k)

[0229] min() is a minimum function, h_score(X)_k is the first entropy coefficient for the kth target ad slot, and h_score(Y)_k is the second entropy coefficient for the kth target ad slot. The above formula is used to determine the first target entropy coefficient hscore(k), whichever has the smaller entropy coefficient of the first and second entropy coefficients for each target ad slot.

[0230] Furthermore, different target ad slot identifiers have different click volumes. Therefore, the click rate corresponding to each target ad slot identifier can be counted, that is, the total click volume can be calculated, and then the ratio of the click volume of each target ad slot identifier to the total click volume can be determined as the click rate. For example, the click volume of target ad slot identifier 1 is 20, the click volume of target ad slot identifier 2 is 30, and the click volume of target ad slot identifier 3 is 50. Then the click rate of target ad slot identifier 1 is 0.2, the click rate of target ad slot identifier 2 is 0.3, and the click rate of target ad slot identifier 3 is 0.5.

[0231] It should be noted that the first target entropy coefficient under the target ad position identifier with a higher click-through rate has a greater effect on the target traffic master cheating detection, and the first target entropy coefficient under the target ad position identifier with a lower click-through rate has a smaller effect on the target traffic master cheating detection. In this way, each first target entropy coefficient can be weighted according to the corresponding click-through rate, and a higher weight is given to the first target entropy coefficient with a high click-through rate, while a lower weight is given to the first target entropy coefficient with a low click-through rate. And the weighted first target entropy coefficients are summed to obtain a second target entropy coefficient, which can reflect the overall click regularity of the target ad on the target ad position of the target traffic master corresponding to the target traffic master identifier. The smaller the second target entropy coefficient, the more regular the clicks on the target ad on the target ad position on the target traffic master, and the greater the probability of cheating. The larger the second target entropy coefficient, the more scattered the clicks on the target ad on the target ad position on the target traffic master, and the smaller the probability of cheating. In order to better illustrate the embodiments of the present application, reference can be made to the following formula:

[0232]

[0233] The H(appid) is the second target entropy coefficient of the target ad slot identifier K, the hscore(k) is the first target entropy coefficient of the target ad slot identifier K, the PV(k) is the number of clicks on the target ad slot identifier K, and the allPV is the number of clicks on all ad slot identifiers (i.e., the total number of clicks). The second target entropy coefficient of the target traffic master can be calculated using the above formula.

[0234] In step 210, when the server detects that the second target entropy coefficient is less than the first preset entropy coefficient, it determines that the detection result of the target traffic main identifier is an abnormal state.

[0235] Among them, the first preset entropy coefficient is the critical value for defining whether the target traffic master corresponding to the target traffic master identifier is normal or abnormal. The first preset entropy coefficient can be set manually or determined by statistically analyzing the average entropy value of the traffic masters that cheat in the background. No specific limitation is given here.

[0236] When the server detects that the second target entropy coefficient is less than the first preset entropy coefficient, it means that the clicks on the target advertisement in the target advertisement position on the target traffic master fed back by the second target entropy coefficient are relatively regular. It can be determined that the detection result of the target traffic master identified by the target traffic master is in an abnormal state, and appropriate control or warning can be performed.

[0237] In step 211, when the server detects that the second target entropy coefficient is not less than the first preset entropy coefficient, it determines that the detection result of the target traffic main identifier is a non-abnormal state.

[0238] When the server detects that the second target entropy coefficient is not less than the first preset entropy coefficient, it means that the clicks on the target advertisement in the target advertisement position on the target traffic master fed back by the second target entropy coefficient are relatively scattered, and it can be determined that the detection result of the target traffic master identified by the target traffic master is non-abnormal, that is, healthy.

[0239] Based on the above embodiments, any traffic owner can be detected for cheating in an unsupervised mode.

[0240] In some embodiments, please continue to refer to 3b, Figure 3b This is another flow chart of the main flow detection method provided in an embodiment of the present application. The contents of step 1, step 2, step 3, step 4, step 5 and step 6 have been implemented in the above embodiment and will not be described in detail here.

[0241] As can be seen from the above, the embodiment of the present application obtains multiple target push information click records of the target traffic master identifier; generates target heat map information corresponding to each target push information position identifier based on multiple target push information click records; calculates the first entropy value of the target click horizontal coordinate set corresponding to the horizontal axis distribution of each target push information position identifier, and the second entropy value of the target click vertical coordinate set corresponding to the vertical axis distribution based on the target heat map information; counts the first target entropy value of the horizontal axis distribution corresponding to each target push information position identifier, and the second target entropy value of the vertical axis distribution of multiple traffic master identifiers; determines the detection result of the target traffic master identifier based on the first relationship between each first entropy value and the corresponding first target entropy value and the second relationship between each second entropy value and the corresponding second target entropy value. In this way, in an unsupervised mode, the detection result of the target traffic master is quickly determined by the relationship between the entropy value of the target traffic master and the entropy values ​​of multiple traffic masters of the overall market. Compared with the solution that requires manual detection of whether the traffic master is cheating, the efficiency of traffic master detection is greatly improved.

[0242] Furthermore, the server can also perform weighted summation on different entropy coefficients according to the click rate to obtain a more accurate second target entropy coefficient for traffic master detection, thereby improving the accuracy of traffic master detection.

[0243] See also Figure 4 , Figure 4 This is a structural diagram of a main traffic detection device provided in an embodiment of the present application, wherein the main traffic detection device may include an acquisition unit 301, a generation unit 302, a calculation unit 303, a statistics unit 304 and a determination unit 305, and the main traffic detection device is applied to a terminal or a server.

[0244] The acquisition unit 301 is used to acquire multiple target push information click records of the target traffic main identifier, and the target push information click record at least includes a target push information position identifier, a click horizontal coordinate, and a click vertical coordinate.

[0245] In some embodiments, the acquiring unit 301 is configured to:

[0246] Obtain multiple click records of the target traffic master corresponding to the target traffic master identifier within a preset time period;

[0247] The target push information bit identifier, the click horizontal coordinate, and the click vertical coordinate in each click record are extracted to generate multiple target push information click records.

[0248] The generating unit 302 is used to generate target heat map information corresponding to each target push information position identifier based on multiple target push information click records, and the target heat map information at least includes the target push information position identifier, the target click horizontal coordinate set and the target click vertical coordinate set.

[0249] In some embodiments, the target push information click record further includes the push information width and the push information height. The generating unit 302 is configured to:

[0250] Normalize each click horizontal coordinate and the corresponding push information width to obtain multiple target click horizontal coordinates;

[0251] Normalize each click vertical coordinate and the corresponding push information height to obtain multiple target click vertical coordinates;

[0252] Determine the target click abscissa set and the target click ordinate set corresponding to each target push information bit identifier;

[0253] A plurality of target heat map information is generated according to each target push information position identifier, the corresponding target click horizontal coordinate set and the target click vertical coordinate set.

[0254] The calculation unit 303 is used to calculate the first entropy value of the horizontal axis distribution corresponding to the target click horizontal coordinate set of each target push information position identifier and the second entropy value of the vertical axis distribution corresponding to the target click vertical coordinate set based on the target heat map information.

[0255] In some embodiments, the calculation unit 303 is configured to:

[0256] Obtain first distribution information of the target click abscissa set corresponding to each target push information bit identifier;

[0257] Calculate the first entropy value corresponding to each first distribution information by using the information entropy formula;

[0258] Obtain second distribution information of the target click vertical coordinate set corresponding to each target push information bit identifier;

[0259] The second entropy value corresponding to each second distribution information is calculated using the information entropy formula.

[0260] The statistical unit 304 is configured to count the first target entropy value of each target push information bit identifier of the plurality of traffic main identifiers distributed on the horizontal axis and the second target entropy value of the vertical axis.

[0261] In some embodiments, the statistics unit 304 is configured to:

[0262] Acquire multiple push information click records of multiple traffic main identifiers, where the push information click records at least include a push information position identifier, a click horizontal coordinate, and a click vertical coordinate, and the push information position identifier includes a target push information position identifier;

[0263] Generate statistical heat map information corresponding to each target push information position identifier based on multiple push information click records, the statistical heat map information at least including the target push information position identifier, the statistical click horizontal coordinate, and the statistical click vertical coordinate;

[0264] Based on the statistical heat map information, the first target entropy value of the horizontal axis distribution corresponding to the statistical click horizontal coordinate set of each target push information position identifier and the second target entropy value of the vertical axis distribution corresponding to the statistical click vertical coordinate set are calculated.

[0265] The determination unit 305 is configured to determine a detection result of the target traffic primary identifier based on a first relationship between each first entropy value and the corresponding first target entropy value and a second relationship between each second entropy value and the corresponding second target entropy value.

[0266] In some embodiments, the determining unit 305 includes:

[0267] The first calculation subunit is used to calculate the ratio of the first entropy value under each target push information bit identifier and the corresponding first target entropy value to obtain multiple first entropy value coefficients;

[0268] The second calculation subunit is used to calculate the ratio of the second entropy value under each target push information bit identifier and the corresponding second target entropy value to obtain multiple second entropy value coefficients;

[0269] The determination subunit is used to determine the detection result of the target traffic main identifier based on the multiple first entropy coefficients and the multiple second entropy coefficients.

[0270] In some embodiments, the determining subunit is configured to:

[0271] Determine the entropy coefficient with the smaller coefficient value between the first entropy coefficient and the second entropy coefficient under each target push information bit identifier as the first target entropy coefficient;

[0272] Count the click rate corresponding to each target push information position identifier;

[0273] Each first target entropy coefficient is weighted and summed according to the corresponding click rate to obtain the corresponding second target entropy coefficient;

[0274] When it is detected that the second target entropy coefficient is less than the first preset entropy coefficient, determining that the detection result of the target flow main identifier is an abnormal state;

[0275] When it is detected that the second target entropy coefficient is not less than the first preset entropy coefficient, it is determined that the detection result of the target flow main identifier is a non-abnormal state.

[0276] In some embodiments, the determining subunit is further configured to:

[0277] Counting a first number of first entropy coefficients smaller than a second preset entropy coefficient and a second number of second entropy coefficients smaller than the second preset entropy coefficient;

[0278] summing the first quantity and the second quantity to obtain a target quantity;

[0279] When it is detected that the target quantity is greater than a preset quantity threshold, determining that the detection result of the target flow main identifier is an abnormal state;

[0280] When it is detected that the target quantity is not greater than the preset quantity threshold, it is determined that the detection result of the target flow main identifier is a non-abnormal state.

[0281] The specific implementation of each of the above units can be found in the previous embodiments and will not be described again here.

[0282] As can be seen from the above, the embodiment of the present application obtains multiple target push information click records of the target traffic master identifier through the acquisition unit 301; the generation unit 302 generates the target heat map information corresponding to each target push information position identifier based on the multiple target push information click records; the calculation unit 303 calculates the first entropy value of the target click horizontal coordinate set corresponding to the horizontal axis distribution of each target push information position identifier, and the second entropy value of the target click vertical coordinate set corresponding to the vertical axis distribution based on the target heat map information; the statistical unit 304 counts the first target entropy value of the horizontal axis distribution corresponding to each target push information position identifier, and the second target entropy value of the vertical axis distribution; the determination unit 305 determines the detection result of the target traffic master identifier based on the first relationship between each first entropy value and the corresponding first target entropy value and the second relationship between each second entropy value and the corresponding second target entropy value. In this way, in the unsupervised mode, the detection result of the target traffic master is quickly determined by the relationship between the entropy value of the target traffic master and the entropy values ​​of multiple traffic masters of the overall market. Compared with the solution that requires manual detection of whether the traffic master is cheating, the efficiency of traffic master detection is greatly improved.

[0283] The embodiment of the present application also provides a computer device, which may be a server, such as Figure 5 As shown, it shows a schematic diagram of the structure of the server involved in the embodiment of the present application, specifically:

[0284] The computer device may include one or more processing core processors 401, one or more computer readable storage media memories 402, a power supply 403, an input unit 404 and other components. Those skilled in the art will understand that Figure 5 The computer device structure shown in the figure does not constitute a limitation on the computer device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange components differently.

[0285] Processor 401 is the control center of the computer device. It connects the various components of the computer device using various interfaces and circuits. By running or executing software programs and / or modules stored in memory 402 and accessing data stored in memory 402, it performs various functions of the computer device and processes data, thereby providing overall monitoring of the computer device. Optionally, processor 401 may include one or more processing cores. Alternatively, processor 401 may integrate an application processor and a modem processor, with the application processor primarily processing the operating system, user interface, and application programs, while the modem processor primarily handles wireless communications. It is understood that the modem processor may not be integrated into processor 401.

[0286] The memory 402 can be used to store software programs and modules. The processor 401 executes various functional applications and data processing by running the software programs and modules stored in the memory 402. The memory 402 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the server, etc. In addition, the memory 402 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage device. Accordingly, the memory 402 may also include a memory controller to provide the processor 401 with access to the memory 402.

[0287] The computer device also includes a power supply 403 for supplying power to various components. Optionally, the power supply 403 can be logically connected to the processor 401 via a power management system, thereby enabling the power management system to manage charging, discharging, and power consumption. The power supply 403 can also include one or more DC or AC power supplies, a recharging system, a power failure detection circuit, a power converter or inverter, a power status indicator, and other arbitrary components.

[0288] The computer device may further include an input unit 404, which may be used to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.

[0289] Although not shown, the computer device may further include a display unit, etc., which will not be described in detail here. Specifically, in this embodiment, the processor 401 in the computer device will load the executable files corresponding to one or more application processes into the memory 402 according to the following instructions, and the processor 401 will run the application stored in the memory 402, thereby implementing the various method steps provided in the above embodiments, as follows:

[0290] Acquire multiple target push information click records of the target traffic main identifier, wherein the target push information click record includes at least a target push information position identifier, a click horizontal coordinate, and a click vertical coordinate; generate target heat map information corresponding to each target push information position identifier based on the multiple target push information click records, wherein the target heat map information includes at least a target push information position identifier, a target click horizontal coordinate set, and a target click vertical coordinate set; based on the target heat map information, calculate a first entropy value of the target click horizontal coordinate set corresponding to the horizontal axis distribution of each target push information position identifier, and a second entropy value of the target click vertical coordinate set corresponding to the vertical axis distribution; count the first target entropy value of the horizontal axis distribution corresponding to each target push information position identifier, and the second target entropy value of the vertical axis distribution of multiple traffic main identifiers; determine the detection result of the target traffic main identifier based on a first relationship between each first entropy value and the corresponding first target entropy value, and a second relationship between each second entropy value and the corresponding second target entropy value.

[0291] In the above embodiments, the description of each embodiment has its own focus. For the part that is not described in detail in a certain embodiment, please refer to the detailed description of the main flow detection method above, which will not be repeated here.

[0292] As can be seen from the above, the computer device of the embodiment of the present application can obtain multiple target push information click records of the target traffic master identifier; generate target heat map information corresponding to each target push information position identifier based on multiple target push information click records; calculate the first entropy value of the target click horizontal coordinate set corresponding to the horizontal axis distribution of each target push information position identifier based on the target heat map information, and the second entropy value of the target click vertical coordinate set corresponding to the vertical axis distribution; count the first target entropy value of the horizontal axis distribution corresponding to each target push information position identifier, and the second target entropy value of the vertical axis distribution of multiple traffic master identifiers; determine the detection result of the target traffic master identifier based on the first relationship between each first entropy value and the corresponding first target entropy value and the second relationship between each second entropy value and the corresponding second target entropy value. In this way, in an unsupervised mode, the detection result of the target traffic master is quickly determined by the relationship between the entropy value of the target traffic master and the entropy values ​​of multiple traffic masters of the overall market. Compared with the solution that requires manual detection of whether the traffic master is cheating, the efficiency of traffic master detection is greatly improved.

[0293] Those skilled in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be accomplished by instructions, or by controlling related hardware through instructions. The instructions may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0294] To this end, an embodiment of the present application provides a computer-readable storage medium storing a plurality of instructions that can be loaded by a processor to execute the steps of any of the traffic master detection methods provided in the embodiments of the present application. For example, the instructions can execute the following steps:

[0295] Acquire multiple target push information click records of the target traffic main identifier, wherein the target push information click record includes at least a target push information position identifier, a click horizontal coordinate, and a click vertical coordinate; generate target heat map information corresponding to each target push information position identifier based on the multiple target push information click records, wherein the target heat map information includes at least a target push information position identifier, a target click horizontal coordinate set, and a target click vertical coordinate set; based on the target heat map information, calculate a first entropy value of the target click horizontal coordinate set corresponding to the horizontal axis distribution of each target push information position identifier, and a second entropy value of the target click vertical coordinate set corresponding to the vertical axis distribution; count the first target entropy value of the horizontal axis distribution corresponding to each target push information position identifier, and the second target entropy value of the vertical axis distribution of multiple traffic main identifiers; determine the detection result of the target traffic main identifier based on a first relationship between each first entropy value and the corresponding first target entropy value, and a second relationship between each second entropy value and the corresponding second target entropy value.

[0296] According to one aspect of the present application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the methods provided in the various optional implementations provided in the above embodiments.

[0297] The computer-readable storage medium may include a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0298] Since the instructions stored in the computer-readable storage medium can execute the steps in any of the main flow detection methods provided in the embodiments of the present application, the beneficial effects that can be achieved by any of the main flow detection methods provided in the embodiments of the present application can be achieved. Please refer to the previous embodiments for details and will not be repeated here.

[0299] The above is a detailed introduction to a flow main detection method, device and computer-readable storage medium provided in the embodiments of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core ideas. At the same time, for technical personnel in this field, based on the ideas of the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A flow detection method, characterized in that: include: Acquire multiple target push information click records of the target traffic main identifier, wherein the target push information click record includes at least a target push information position identifier, a click horizontal coordinate, and a click vertical coordinate; Generate target heat map information corresponding to each target push information position identifier based on multiple target push information click records, wherein the target heat map information includes at least the target push information position identifier, the target click horizontal coordinate set, and the target click vertical coordinate set; Based on the target heat map information, calculate the first entropy value of the target click abscissa set corresponding to the horizontal axis distribution of each target push information position identifier, and the second entropy value of the target click ordinate set corresponding to the vertical axis distribution; Counting the first target entropy value of the horizontal axis corresponding to each target push information bit identifier of multiple traffic main identifiers and the second target entropy value of the vertical axis distribution; The detection result of the target traffic primary identifier is determined according to a first relationship between each first entropy value and the corresponding first target entropy value and a second relationship between each second entropy value and the corresponding second target entropy value.

2. The flow rate detection method according to claim 1, characterized in that: The determining, based on a first relationship between each first entropy value and a corresponding first target entropy value and a second relationship between each second entropy value and a corresponding second target entropy value, a detection result of the target flow primary identifier includes: Calculate the ratio of the first entropy value under each target push information bit identifier to the corresponding first target entropy value to obtain multiple first entropy value coefficients; Calculate the ratio of the second entropy value under each target push information bit identifier to the corresponding second target entropy value to obtain multiple second entropy value coefficients; The detection result of the target traffic main identifier is determined according to the multiple first entropy coefficients and the multiple second entropy coefficients.

3. The flow rate detection method according to claim 2, characterized in that: The determining the detection result of the target traffic primary identifier according to the multiple first entropy coefficients and the multiple second entropy coefficients includes: Determine the entropy coefficient with the smaller coefficient value between the first entropy coefficient and the second entropy coefficient under each target push information bit identifier as the first target entropy coefficient; Count the click rate corresponding to each target push information position identifier; Each first target entropy coefficient is weighted and summed according to the corresponding click rate to obtain the corresponding second target entropy coefficient; When it is detected that the second target entropy coefficient is less than the first preset entropy coefficient, determining that the detection result of the target flow main identifier is an abnormal state; When it is detected that the second target entropy coefficient is not less than the first preset entropy coefficient, it is determined that the detection result of the target flow main identifier is a non-abnormal state.

4. The flow rate detection method according to claim 2, characterized in that: The determining the detection result of the target traffic primary identifier according to the multiple first entropy coefficients and the multiple second entropy coefficients includes: Counting a first number of first entropy coefficients smaller than a second preset entropy coefficient and a second number of second entropy coefficients smaller than the second preset entropy coefficient; summing the first quantity and the second quantity to obtain a target quantity; When it is detected that the target quantity is greater than a preset quantity threshold, determining that the detection result of the target flow main identifier is an abnormal state; When it is detected that the target quantity is not greater than a preset quantity threshold, it is determined that the detection result of the target flow main identifier is a non-abnormal state.

5. The flow rate detection method according to claim 1, characterized in that: The method of obtaining multiple target push information click records of the target traffic primary identifier includes: Obtain multiple click records of the target traffic master corresponding to the target traffic master identifier within a preset time period; The target push information bit identifier, the click horizontal coordinate, and the click vertical coordinate in each click record are extracted to generate multiple target push information click records.

6. The flow rate detection method according to claim 5, characterized in that: The click record does not include a preset conversion click event.

7. The flow rate detection method according to claim 5, characterized in that: The target push information click record also includes the push information width and the push information height. The target heat map information corresponding to each target push information position identifier is generated based on the multiple target push information click records, including: Normalize each click horizontal coordinate and the corresponding push information width to obtain multiple target click horizontal coordinates; Normalize each click vertical coordinate and the corresponding push information height to obtain multiple target click vertical coordinates; Determine the target click abscissa set and the target click ordinate set corresponding to each target push information bit identifier; A plurality of target heat map information is generated according to each target push information position identifier, the corresponding target click horizontal coordinate set and the target click vertical coordinate set.

8. The flow rate detection method according to any one of claims 1 to 7, characterized in that: The step of calculating, based on the target heat map information, a first entropy value of the target click abscissa set corresponding to the horizontal axis distribution of each target push information position identifier and a second entropy value of the target click ordinate set corresponding to the vertical axis distribution, includes: Obtain first distribution information of the target click abscissa set corresponding to each target push information bit identifier; Calculate the first entropy value corresponding to each first distribution information by using the information entropy formula; Obtain second distribution information of the target click vertical coordinate set corresponding to each target push information bit identifier; The second entropy value corresponding to each second distribution information is calculated using the information entropy formula.

9. The flow rate detection method according to any one of claims 1 to 7, characterized in that: The method of counting the first target entropy value of the horizontal axis corresponding to each target push information bit identifier of the plurality of traffic main identifiers and the second target entropy value of the vertical axis includes: Acquire multiple push information click records of multiple traffic master identifiers, wherein the push information click records include at least a push information position identifier, a click horizontal coordinate, and a click vertical coordinate, and the push information position identifier includes a target push information position identifier; Generate statistical heat map information corresponding to each target push information position identifier based on multiple push information click records, wherein the statistical heat map information includes at least the target push information position identifier, the statistical click horizontal coordinate, and the statistical click vertical coordinate; Based on the statistical heat map information, a first target entropy value of the horizontal axis distribution corresponding to the statistical click horizontal coordinate set of each target push information position identifier and a second target entropy value of the vertical axis distribution corresponding to the statistical click vertical coordinate set are calculated.

10. A flow rate detection device, characterized in that: include: an acquiring unit, configured to acquire a plurality of target push information click records of a target traffic primary identifier, wherein the target push information click record comprises at least a target push information position identifier, a click horizontal coordinate, and a click vertical coordinate; A generating unit, configured to generate target heat map information corresponding to each target push information position identifier based on a plurality of target push information click records, wherein the target heat map information includes at least the target push information position identifier, a target click abscissa set, and a target click ordinate set; A calculation unit, configured to calculate, based on the target heat map information, a first entropy value of the target click abscissa set corresponding to the horizontal axis distribution of each target push information position identifier, and a second entropy value of the target click ordinate set corresponding to the vertical axis distribution; A statistical unit, configured to count a first target entropy value of each target push information bit identifier of a plurality of traffic main identifiers distributed on the horizontal axis, and a second target entropy value of the vertical axis; A determination unit is used to determine the detection result of the target traffic main identifier based on a first relationship between each first entropy value and the corresponding first target entropy value and a second relationship between each second entropy value and the corresponding second target entropy value.

11. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor to execute the steps in the main flow detection method according to any one of claims 1 to 9.

12. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the main flow detection method according to any one of claims 1 to 9 are implemented.

13. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the steps of the main flow detection method according to any one of claims 1 to 9 are implemented.

Images