Platform shielding method for intelligent identification of forbidden behaviors

Abstract

The application provides a platform shielding and banning behavior intelligent identification method, relates to data processing technology, controls multiple access terminals to search keywords of each detection platform based on an identification plug-in, obtains a search interface of each detection platform, performs relevance identification on the search interface, and obtains a recommendation restriction coefficient; according to the search keywords, the same interactive text in each search interface is triggered and processed, a loading interface is obtained, interactive restriction identification is performed on the loading interface, a function restriction coefficient is obtained; according to the recommendation restriction coefficient and the function restriction coefficient, a comprehensive restriction coefficient is obtained, and a banning identification result of each detection platform is determined, the loading interface can be implicitly shielded and identified, various types of monopoly restriction measures in the loading process can be accurately captured, and the positioning cost and time cost of users are reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to data processing technology, and in particular to a platform shielding and banning behavior intelligent identification method. BACKGROUND

[0002] Under the background of rapid development of Internet platform economy, platform shielding and banning behavior has become the focus of anti-monopoly supervision. Some large platforms implement monopoly behavior by limiting competitors' link access, shielding specific content display, hindering cross-platform interaction, etc., which destroys the fair competition market order and damages consumer rights and interests. Therefore, intelligent identification of platform shielding and banning behavior, especially accurate capture of various hidden restrictions from the perspective of anti-monopoly, has become a key technical requirement for maintaining market fairness and ensuring healthy development of the industry. Such identification technology can provide objective and efficient monitoring basis for regulatory departments, help to stop monopoly behavior in a timely manner, and promote the orderly development of platform economy.

[0003] However, traditional methods focus on the presence or absence of explicit levels such as website search results and recommendation lists (see Chinese patent CN112884502A for details), and only gather explicit search results, but it is difficult to capture hidden monopoly behaviors that exist in the loading process. For example, some platforms may indirectly prevent users from accessing competitors' content or services by delaying responses, loading errors, etc. in the loading interface. This behavior is not directly banned on the surface, but in essence, it increases the positioning cost of users through implicit restrictions in the loading link. Existing technologies often cannot effectively identify them, resulting in a large number of hidden monopoly behaviors escaping supervision.

[0004] Therefore, how to perform implicit shielding identification on the loading interface, accurately capture various monopoly restrictions in the loading process, and reduce the positioning cost and time cost of users has become a problem to be solved. SUMMARY

[0005] The embodiment of the present application provides a platform shielding and banning behavior intelligent identification method, which can perform implicit shielding identification on the loading interface, accurately capture various monopoly restrictions in the loading process, and reduce the positioning cost and time cost of users.

[0006] In a first aspect, the embodiment of the present application provides a platform shielding and banning behavior intelligent identification method, comprising:

[0007] Based on the identification plug-in, a plurality of access terminals perform search word retrieval on each detection platform to obtain a retrieval interface of each detection platform, and perform relevance identification on the retrieval interface to obtain a recommendation restriction coefficient;

[0008] According to the search word, the same interactive text in each search interface is triggered and processed to obtain a loading interface, the loading interface is subjected to interactive restriction identification to obtain a function restriction coefficient;

[0009] According to the recommendation restriction coefficient and the function restriction coefficient, a comprehensive restriction coefficient is obtained, and a ban identification result of each detection platform is determined based on the comprehensive restriction coefficient.

[0010] Optionally, in a possible implementation manner of the first aspect, the association identification on the search interface to obtain the recommendation restriction coefficient comprises:

[0011] The number of interactive texts in the search interface is counted to obtain a recommendation number;

[0012] Interactive texts without the search word in the search interface are identified as interference texts, and the number of the interference texts is counted to obtain an interference number;

[0013] According to a ratio of the interference number to the recommendation number, the recommendation restriction coefficient is obtained.

[0014] Optionally, in a possible implementation manner of the first aspect, the interactive restriction identification on the loading interface to obtain the function restriction coefficient comprises:

[0015] Loading restriction identification is performed on a loading element in the loading interface to obtain a loading restriction coefficient;

[0016] Operation restriction identification is performed on an operation element in the loading interface to obtain an operation restriction coefficient;

[0017] According to a sum value of the loading restriction coefficient and the operation restriction coefficient, the function restriction coefficient is obtained.

[0018] Optionally, in a possible implementation manner of the first aspect, the loading restriction identification on the loading element in the loading interface to obtain the loading restriction coefficient comprises:

[0019] Each loading element in the loading interface is identified, and the loading element comprises an image element, a video element and a component element;

[0020] Loading time identification is performed on the image element in the loading interface to obtain an image restriction coefficient;

[0021] Loading deviation identification is performed on the video element in the loading interface to obtain a video restriction coefficient;

[0022] Loading interference identification is performed on the component element in the loading interface to obtain a component restriction coefficient;

[0023] According to a sum of the image restriction coefficient, the video restriction coefficient and the component restriction coefficient, a loading restriction coefficient is obtained.

[0024] Optionally, in a possible implementation manner of the first aspect, the loading time identification on the image elements in the loading interface to obtain the image restriction coefficient comprises:

[0025] An image reference duration of the same image elements in each loading interface is obtained, and an image loading duration is obtained, and based on a ratio of the image loading duration and the image reference duration, an image difference value of each image element is obtained.

[0026] The image difference value of each detection platform is counted to obtain a platform image difference value, and an image quantity of the image elements in the corresponding loading interface is obtained, and according to a ratio of the platform image difference value and the image quantity, an image restriction coefficient is obtained.

[0027] Optionally, in a possible implementation manner of the first aspect, the loading deviation identification on the video elements in the loading interface to obtain the video restriction coefficient comprises:

[0028] The video element is started, an image playing duration corresponding to a playing frame of a current playing duration is determined, and an audio playing duration of an audio frame is determined;

[0029] A playing deviation duration of the image playing duration and the audio playing duration is determined, and a progress deviation duration of the current playing duration is determined;

[0030] Based on a sum of the playing deviation duration and the progress deviation duration, a video deviation duration is obtained, and according to a ratio of the video deviation duration and the current playing duration, a video deviation coefficient is obtained, and the video deviation coefficients of each loading interface are summed to obtain a platform video difference value;

[0031] According to a ratio of the platform video difference value and a quantity of the video elements in the corresponding loading interface, a video restriction coefficient is obtained.

[0032] Optionally, in a possible implementation manner of the first aspect, the loading interference identification on the component elements in the loading interface to obtain the component restriction coefficient comprises:

[0033] A background area where the component element is located and a component area corresponding to the component element are obtained;

[0034] A pixel difference value between the component area and the background area is identified, and a reference difference value of the corresponding component area is obtained;

[0035] Based on a ratio of the reference difference value and the corresponding pixel difference value, a pixel difference value of each component element is obtained, and a component quantity of the corresponding component element of each detection platform is counted;

[0036] Summing the pixel difference values in each detection platform, platform pixel difference values of each detection platform are obtained, and a component restriction coefficient is obtained according to a ratio of the platform pixel difference value and the component quantity.

[0037] Optionally, in a possible implementation manner of the first aspect, the operation restriction identification on the operation elements in the loading interface comprises:

[0038] The moving track of the operation element is recorded, and a transition form of the operation element in the moving track is obtained, the transition form comprising an arrow form, a palm form and a vertical line form;

[0039] The moving track is processed based on the transition form, to obtain a non-interaction track of the arrow form, a trigger track of the palm form and an input track of the vertical line form;

[0040] The non-interaction track, the trigger track and the input track are respectively analyzed, to obtain a deformation deviation coefficient of the non-interaction track, a trigger deviation coefficient of the trigger track and an input deviation coefficient of the input track;

[0041] The operation restriction coefficient is obtained according to the deformation deviation quantity, the trigger deviation coefficient and the input deviation coefficient.

[0042] Optionally, in a possible implementation manner of the first aspect, the operation restriction identification on the operation elements in the loading interface comprises:

[0043] The input region and the trigger region in the loading interface are taken as an interaction region, and an intersection of the non-interaction track and the interaction region is counted, to obtain a deformation deviation quantity;

[0044] The deformation deviation coefficient is obtained based on a ratio of the deformation deviation quantity and a preset deviation quantity;

[0045] A color change time of a trigger region in which the trigger track is located is counted from a trigger time point at which the trigger input device is triggered in the trigger track, to obtain a trigger change duration;

[0046] The trigger deviation coefficient is obtained based on a ratio of the trigger change duration and a reference change duration;

[0047] Adjacent two input tracks in a corresponding input region are selected as selected tracks, a selected track in front is taken as a first track, and a selected track in back is taken as a second track;

[0048] A skip time period is obtained according to a start time point of an input time period corresponding to the second track and an end time point of an input time period corresponding to the first track;

[0049] when the input state of the key input device in the jump period is determined as the dormant state, the corresponding jump period is determined as an abnormal period;

[0050] a length of the abnormal period is counted to obtain an abnormal jump length, and an input deviation coefficient is obtained according to a ratio of the abnormal jump length to a reference jump length.

[0051] Optionally, in a possible implementation manner of the first aspect, the obtaining of the comprehensive restriction coefficient of each detection platform according to the sum of the recommendation restriction coefficient and the function restriction coefficient comprises:

[0052] the comprehensive restriction coefficient of each detection platform is obtained according to a sum of the recommendation restriction coefficient and the function restriction coefficient;

[0053] when the comprehensive restriction coefficient is determined to be less than a preset judgment coefficient, the ban identification result of the corresponding detection platform is determined as a non-screening ban behavior;

[0054] when the comprehensive restriction coefficient is determined to be greater than or equal to the preset judgment coefficient, the ban identification result of the corresponding detection platform is determined as a screening ban behavior.

[0055] In a second aspect of the embodiment of the application, a storage medium is provided, and the storage medium stores a computer program, and the computer program is executed by a processor to implement the method in the first aspect of the application and the method in the first aspect of the application in various possible manners.

[0056] The application has the following beneficial effects:

[0057] 1. The application controls multiple access terminals to perform retrieval by identification of plug-ins, obtains a recommendation restriction coefficient from relevance identification of a retrieval interface, and obtains a function restriction coefficient from interactive restriction identification of a loading interface to obtain a final result. Compared with a traditional single-dimensional identification method, this multi-dimensional quantitative analysis covers the relevance deviation degree (such as the proportion of interference text) of the retrieval result recommendation level and includes the specific performance of the function restriction in the loading process (such as loading delay and operation obstruction), and thus achieves relatively comprehensive analysis of explicit and implicit ban behaviors of the platform.

[0058] 2、The application realizes the deep identification of hidden restrictions in the loading process by splitting the loading elements (images, videos, components) and operation elements, and respectively calculating the loading restriction coefficient and the operation restriction coefficient. For example, the image element analysis loading time deviation, the video element detection audio and picture synchronization and progress deviation, and the component element evaluation visual distinction degree can more accurately find the monopoly behavior of the platform through delay loading, audio and picture disorder, component visual confusion and other hidden means; at the same time, by tracking the track form of the operation element, the mouse form switching delay, trigger response lag, input cursor abnormal jump and other operation level restrictions can be identified, which provides more comprehensive technical support for anti-monopoly supervision. BRIEF DESCRIPTION OF DRAWINGS

[0059] Figure 1 The application scenario of the technical solution provided by the application is shown in the figure.

[0060] Figure 2 The flow chart of the platform shielding and banning behavior intelligent identification method provided by the application is shown in the figure.

[0061] Figure 3 The standard time length of the image frame provided by the application is shown in the figure.

[0062] Figure 4 The image playing time length provided by the application is shown in the figure.

[0063] Figure 5 The transformation form provided by the application is shown in the figure. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical scheme and advantages of the embodiments of the application clearer, the technical scheme in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the application.

[0065] The terms "first", "second", "third", "fourth" and the like (if any) in the specification and claims of the application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily have to describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the application described herein can be implemented in an order other than those illustrated or described herein.

[0066] It should be understood that, in various embodiments of the present application, the magnitude of the serial number of each process does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0067] It should be understood that, in the present application, "comprising" and "having" and any variants thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device comprising a series of steps or units does not have to be limited to the clearly listed steps or units, but can include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0068] It should be understood that, in the present application, "multiple" means two or more. "And / or" is only a description of the association between the associated objects, which means that there can be three relationships, for example, A and / or B can represent the following three cases: A exists alone, A and B exist together, and B exists alone. The character " / " generally represents that the associated objects before and after are in an "or" relationship. "Including A, B and C", "including A, B, C" means that A, B and C are all included, "including A, B or C" means that one of A, B and C is included, and "including A, B and / or C" means that any one or any two or three of A, B and C is included.

[0069] It should be understood that, in the present application, "B corresponding to A", "B corresponding to A", "A corresponding to B" or "B corresponding to A" means that B is associated with A, and B can be determined according to A. Determining B according to A does not mean that B is determined only according to A, but also can be determined according to A and / or other information. The matching between A and B means that the similarity between A and B is greater than or equal to a preset threshold.

[0070] Depending on the context, "if" as used herein can be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting".

[0071] The technical solutions of the present application will be described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0072] As Figure 1As shown in the figure, the application scenario provided by the technical solution of the present application includes a server and multiple access terminals, wherein the server is in communication connection with the multiple access terminals, and the access terminals include but are not limited to computers, tablet computers, smart phones and the like, wherein the access terminals are mobile terminals specially authorized for monitoring platform behavior, and it should be noted that when testing different web platforms, the parameters of the access terminals should be ensured to be consistent to ensure that the results are accurate and comparable, specifically including: using the same model, configuration and display parameters of the device on the hardware; using the same type, bandwidth and IP settings on the network; using the same operating system version and settings on the software, and all of them are accessed in the period of time when the access volume is extremely low at night, that is, the consistency of the access terminals is ensured, which will not be described here.

[0073] Specifically, the user can send instructions to the identification plug-in of each access terminal through the server, and the identification plug-in controls the corresponding access terminal to access the detection platform, and then identifies the behavior of the detection platform in the search interface and the loading interface of the platform jump in the process of accessing the detection platform.

[0074] The present application provides a kind of platform shielding ban behavior intelligent identification method, as shown in the figure, including steps S1-S3: Figure 2 As shown in the figure, including steps S1-S3:

[0075] S1, based on identification plug-in control multiple access terminals to search word search for each detection platform, obtain the search interface of each detection platform, and perform relevance identification on the search interface to obtain a recommended restriction coefficient.

[0076] Wherein, the identification plug-in refers to a program module embedded in the access terminal, which is used to automatically execute instruction operation and collect interface data; the access terminal refers to a terminal simulating user behavior (such as multiple test devices); the detection platform refers to a network platform to be detected for whether there is shielding behavior (such as a search platform); the search word refers to a keyword used for searching (such as a product name, an event keyword, etc.); the search interface refers to a search result page returned by the detection platform (including recommended text, links and other contents); relevance identification refers to analyzing the matching degree of the recommended content and the search word in the search interface; the recommended restriction coefficient refers to a numerical value quantifying the correlation deviation of the recommended content and the search word (the higher the numerical value, the weaker the correlation, the more obvious the restriction).

[0077] It can be understood that by controlling multiple access terminals through the identification plug-in, the same search word is input in each detection platform to perform a search operation, and the search interface corresponding to each detection platform is obtained, and then the interactive text (such as a recommended title with a hyperlink) in the search interface is extracted to determine the interactive text irrelevant to the search word in the search interface, so as to calculate the proportion to obtain the recommended restriction coefficient.

[0078] In some embodiments, the relevance identification of the search interface in step S1 to obtain the recommendation restriction coefficient comprises S11-S13:

[0079] S11, count the number of interactive texts in the search interface to obtain the recommended number.

[0080] Among them, the interactive text refers to the text content (such as the title of the recommended goods, the title of the information article, etc.) that can be clicked or viewed in the search interface, for example, a plurality of recommended websites will appear after the keyword search of a search platform, and the recommended number refers to the total number of all interactive texts in the search interface.

[0081] It can be understood that the text extraction of the search interface is performed by recognizing the plug-in (such as recognizing the text area in the interface by OCR text technology), and the content meeting the interactive text characteristics is screened out; the number of these interactive texts is counted one by one to obtain the recommended number; and it is ensured that the statistical result covers all recommended contents in the interface.

[0082] S12, identify the interactive text in the search interface without the search term as the interference text, and count the number of the interference text to obtain the interference number.

[0083] Among them, the interference text refers to the interactive text in the search interface that does not contain the search term (for example, the search term is brand A, but the web page title is brand B); and the interference number refers to the total number of all interference texts.

[0084] S13, according to the ratio of the interference number and the recommended number, the recommendation restriction coefficient is obtained.

[0085] Among them, the recommendation restriction coefficient refers to the ratio of the interference number and the recommended number (recommendation restriction coefficient = interference number / recommended number), and the value range is 0-1 (the closer the value is to 1, the higher the interference text proportion is, the weaker the correlation between the recommended content and the search term is, and the more obvious the platform restriction on the search is).

[0086] S2, according to the search term, the same interactive text in each search interface is triggered and processed to obtain a loading interface, and the interactive restriction identification of the loading interface is performed to obtain a function restriction coefficient.

[0087] It should be noted that the platform shielding and banning behavior has become the focus of anti-monopoly supervision. Some large platforms implement monopoly behavior by limiting the access of competitors' links, shielding the display of specific content, hindering cross-platform interaction, etc., therefore, when identifying the platform shielding behavior, only the display result is used for judgment, and various hidden restriction measures of the platform cannot be identified.

[0088] Therefore, the scheme of the present application analyzes the hidden loading in the loading interface and determines the platform that appears to hinder.

[0089] It can be understood that the scheme of our side triggers the same interactive text in the search interface of each detection platform according to the search term, such as opening the same webpage interface in different search platforms, for example, opening the webpage interface corresponding to brand A, so as to obtain the corresponding loading interface, and then the interactive restriction identification is performed on the loading interface to obtain the function restriction coefficient, wherein the interactive restriction identification includes loading restriction identification and operation restriction identification.

[0090] In some embodiments, the interactive restriction identification of the loading interface in step S2 to obtain the function restriction coefficient includes S21-S23:

[0091] S21, loading restriction identification is performed on the loading elements in the loading interface to obtain a loading restriction coefficient.

[0092] Among them, the loading element refers to the content (such as pictures, introduction videos, text, etc.) that needs to be dynamically loaded in the loading interface. Generally, the loading speed of text is fast and does not need to be compared; loading restriction identification refers to loading restriction identification of loading elements on each platform.

[0093] In some embodiments, the loading restriction identification of the loading elements in the loading interface in step S21 to obtain the loading restriction coefficient includes S211-S215:

[0094] S211, identify the loading elements in the loading interface, including image elements, video elements and component elements.

[0095] It can be understood that the structure of the loading interface is parsed, and the elements of images, videos and components that need to be loaded are distinguished, that is, the images, videos and components (payment, sliding and other component buttons) that need to be loaded in the webpage are identified.

[0096] S212, loading time identification is performed on the image elements in the loading interface to obtain an image restriction coefficient.

[0097] It can be understood that the loading area containing the same interactive text is located in each detection platform, that is, the same webpage interface is opened in different detection platforms, such as opening the webpage corresponding to brand A, and then the corresponding image elements in the webpage are extracted, and the loading condition of each image element is recorded, so as to obtain the image restriction coefficient.

[0098] In some embodiments, the loading time identification of the image elements in the loading interface in step S212 to obtain the image restriction coefficient includes S2121-S2122:

[0099] S2121, obtain the image reference duration of the same image element in each loading interface and the image loading duration, and obtain the image difference value of each image element based on the ratio of the image loading duration and the image reference duration.

[0100] The image reference duration is the standard loading duration of the image element, and the image loading duration is the actual loading duration of the image element.

[0101] It can be understood that the server obtains the loading duration of the image when entering the webpage in the same image element in each detection platform, or the loading duration triggered by image loading such as simulating user clicking to view the corresponding image, and performs comparison in the same dimension. When the image element is completely displayed, the timing is stopped, and the duration is recorded. The duration is taken as the image loading duration of the corresponding image recognition area, and the image reference duration of the image element is obtained. Specifically, it can be pre-set according to the image size, or it can be the average value of the loading durations of the image element of each platform.

[0102] Further, the image loading duration of each image recognition area in each detection platform is divided by the corresponding image reference duration to obtain the image difference value of each image recognition area (the ratio > 1 indicates that the loading time is longer than the normal level, and the larger the value, the more obvious the deviation).

[0103] S2122, statistics of the image difference value of each detection platform, obtain the platform image difference value, obtain the image quantity of the image element in the corresponding loading interface, and obtain the image restriction coefficient according to the ratio of the platform image difference value and the image quantity.

[0104] It can be understood that the image difference values of all image recognition areas in a certain detection platform are added to obtain the platform image difference value of all images in each detection platform. The image restriction coefficient of the detection platform is obtained by dividing the value by the image quantity. If the deviation from 1 is large, it indicates that the average loading time is longer than the normal level, and there is a loading restriction.

[0105] S213, loading deviation identification of the video element in the loading interface is performed to obtain a video restriction coefficient.

[0106] The video element can be a played video (such as a product introduction corresponding to brand A); the loading deviation identification refers to analyzing deviation indexes such as audio-visual synchronization and progress bar accuracy in the video playing process; and the video restriction coefficient refers to a numerical value quantifying the loading deviation degree of the video element.

[0107] In some embodiments, (loading deviation identification of the video element in the loading interface is performed to obtain a video restriction coefficient) in step S213 includes S2131-S2134:

[0108] S2131, starting the video element, determining the image playing time length corresponding to the playing frame of the current playing time length, and the audio playing time length of the audio frame.

[0109] The current playing time length is the playing time length after starting the video element, the playing frame is the image frame corresponding to the current playing picture of the video element, and the audio frame is the image frame corresponding to the current playing audio of the video element.

[0110] It can be understood that the starting key of the video element triggers the plug-in to trigger the video starting component by simulating user operation (such as clicking the play button), such as playing a certain video; the current playing time length in the playing process is recorded in real time, and the image frame currently played is intercepted as the playing frame, and the audio played in the video element is analyzed to obtain the audio frame.

[0111] First, the standard time length of each image frame in the video element is obtained, that is, the time length of each picture in the played video is taken as the standard time length, see Figure 3 For example, the playing time length corresponding to the third image frame in the video is 0.03s, which is only used for illustration.

[0112] Further, the configuration text in the playing frame is identified, and the playing audio of the current playing time length is picked up, the playing audio is analyzed to obtain the playing text.

[0113] The configuration text refers to the text subtitle in the playing frame, that is, the text in each image frame in the video playing process.

[0114] It can be understood that the configuration text in the playing frame is identified by the OCR technology; at the same time, the audio played at the current playing time length is picked up, which is converted into playing text by voice recognition technology, which can be picked up by a sound pickup device. Thus, the text corresponding to the current playing picture and the actual voice playing text can be obtained.

[0115] The standard time length corresponding to the configuration text is called as the image playing time length, and the standard time length corresponding to the playing text is determined as the audio playing time length.

[0116] S2132, determining the playing deviation time length of the image playing time length and the audio playing time length, and the progress deviation time length of the current playing time length.

[0117] It is not difficult to understand that when the platform limits the playing of the video in the webpage, the picture and the audio will not be in progress, and the actual playing time length will not be consistent with the progress bar, resulting in that the sound and the video are not matched, and the user experience is reduced.

[0118] According to the absolute value of the difference between the image playing duration and the audio playing duration, a report deviation duration is obtained, which is the difference between the sound and the actual video played picture corresponding duration. According to the absolute value of the difference between the image playing duration and the current playing duration, a progress deviation duration is obtained, which is the duration of the current progress bar.

[0119] Therefore, the scheme of our side extracts the current position of the video progress bar, that is, according to the absolute value of the difference between the image playing duration and the current playing duration, a progress deviation duration is obtained, which is the deviation value between the theoretical playing duration and the actual video progress bar playing duration (see Figure 4 For example, the image playing duration is 10s). If the deviation is too large, it means that the platform restricts the video loading, and the video loading is more obvious.

[0120] S2133, based on the sum of the report deviation duration and the progress deviation duration, a video deviation duration is obtained, and according to the ratio of the video deviation duration and the current playing duration, a video deviation coefficient is obtained. Sum the video deviation coefficients of each loading interface to obtain the platform video difference value.

[0121] According to the sum of the report deviation duration and the progress deviation duration, the video deviation duration of each video element is obtained.

[0122] It can be understood that the video deviation duration refers to the sum of the report deviation duration and the progress deviation duration, which reflects the total deviation degree of the video in terms of audio-visual synchronization and progress accuracy.

[0123] It is not difficult to understand that dividing the video deviation duration by the current playing duration obtains the video deviation coefficient of each video element, that is, the deviation degree of the video element.

[0124] Further, the video deviation coefficients of all video identification areas in a certain detection platform are added to obtain the platform video difference value corresponding to the detection platform.

[0125] S2134, according to the ratio of the platform video difference value and the number of video elements in the corresponding loading interface, a video restriction coefficient is obtained.

[0126] It is not difficult to understand that dividing the platform video difference value by the number of videos obtains the average deviation degree of the platform with respect to the video element, that is, the video restriction coefficient; if the video restriction coefficient is larger, it means that the average deviation of the platform video loading exceeds the normal level, and there is obvious restriction.

[0127] S214, the loading interference of the component element in the loading interface is identified to obtain a component restriction coefficient.

[0128] The component element refers to a functional component (such as a page scroll bar, a payment function component, etc.) in the loading area; the loading interference identification refers to analyzing the visual distinction (such as color difference) between the component element and the background; and the component restriction coefficient refers to a numerical value quantifying the visual interference degree of the component element.

[0129] In some embodiments, the loading interference identification of the component element in the loading interface in step S214 to obtain the component restriction coefficient includes S2141-S2144.

[0130] S2141, obtaining the background area where the component element is located and the component area corresponding to the component element.

[0131] It should be noted that different detection platforms may have different display modes for the same webpage. For example, when a platform is more inclined to brand A than brand B, the display state of the component in the process of displaying the official website corresponding to the brand B will be changed so that it is closer to the background color of the area where the component is located, thereby improving the positioning difficulty of the user and increasing the positioning cost, such as making the scroll bar transparent.

[0132] It can be understood that the area where the same component element in the same loading interface in each detection platform is determined as the component area, such as the area where the trigger component of the scroll bar is located, and the same color area where the component element is located is determined as the background area, such as the sliding area where the current slider is located and the background area where the payment button is located, so as to ensure that the background area only contains the direct background of the component around the component and avoid including irrelevant area pixels (such as other components and advertisement columns). The area can also be determined by expanding the component by a preset multiple or manually determined, which is not limited here.

[0133] S2142, identifying the pixel difference value between the component area and the background area and obtaining the reference difference value of the corresponding component area.

[0134] The reference difference value is the lowest difference value between the pixel values of the component area and the background area, that is, the lowest requirement for the pixel value. The reference difference value can be a value set by a person in advance, or the average value of the pixel difference values of each platform for this component element.

[0135] It can be understood that the colors of the component area and the background area are generally single, and the pixel difference value can be obtained by directly calculating the difference between the pixel values of the two. For example, the average pixel value of the component is (100, 100, 100), and the average pixel value of the background is (50, 10, 50). The corresponding pixel difference value is (100-50)+(100-10)+(100-50)=190. The brightness value can also be processed. It is not difficult to understand that the pixel difference value can be used to reflect the overall difference of the pixel value. When the data deviation is small, the color is closer, and it is more difficult for the user to locate. The larger the value is, the more different it is, and it is easier for the user to locate.

[0136] It is not difficult to understand that if the colors of the component area and the background area are not single, the pixel points in the background area are taken as background pixel points, and the pixel points in the component recognition area are taken as component pixel points. All pixel points in the background area are marked as background pixel points, and all pixel points in the component recognition area are marked as component pixel points; ensure that the attribution of each pixel point is clear. The average value of the pixel values of the component pixel points is calculated to obtain the average pixel value corresponding to the component recognition area, and the average value of the pixel values of the background pixel points is calculated to obtain the average pixel value corresponding to the background area. Extract the pixel values of all component pixel points, calculate the average value to obtain the average pixel value of the component (the R, G and B values of the RGB of the component pixel points can be weighted and averaged to obtain the average pixel value of the component, or the RGB value can be converted into a brightness value); similarly, the average value of the pixel values of all background pixel points is calculated to obtain the background average pixel value. The absolute value of the difference between the average pixel value of the component and the average pixel value of the background is calculated to obtain the pixel difference value of each component area.

[0137] S2143, based on the ratio of the reference difference value and the corresponding pixel difference value, the pixel difference value of each component element is obtained, and the number of components corresponding to each detection platform is counted.

[0138] It can be understood that the ratio of the reference difference value and the corresponding pixel difference value, that is, the reference difference value divided by the pixel difference value of each component, the smaller the pixel difference value is, the larger the pixel difference value of the component recognition area is, that is, the closer the color is, the larger the corresponding pixel difference value is, so that the subsequent component limiting coefficient is larger, which means that the platform will make the component closer to the background color of the area where the component is located, thereby improving the difficulty of user positioning and increasing the positioning cost, such as making the scroll bar transparent.

[0139] S2144, the pixel difference values in each detection platform are summed to obtain the platform pixel difference value of each detection platform, and the ratio of the platform pixel difference value and the number of components is obtained to obtain the component limiting coefficient.

[0140] It can be understood that the pixel difference values of all component areas in a certain detection platform are added to obtain a platform pixel difference value; the value is divided by the number of components to obtain an average relative difference degree, i.e., a component restriction coefficient. The larger the component restriction coefficient is, the closer the components in the platform are to the corresponding background color, and the more hidden the platform is.

[0141] S215, obtaining a loading restriction coefficient according to the sum of the image restriction coefficient, the video restriction coefficient and the component restriction coefficient.

[0142] It can be understood that the image restriction coefficient, the video restriction coefficient and the component restriction coefficient are added to obtain a loading restriction coefficient.

[0143] S22, performing operation restriction identification on the operation elements in the loading interface to obtain an operation restriction coefficient.

[0144] The operation elements refer to elements in the loading interface that support user operations, such as mouse icons in the loading interface that can support user movement triggering and other operations.

[0145] In some embodiments, the operation restriction identification on the operation elements in the loading interface in step S22 to obtain an operation restriction coefficient includes S221-S224.

[0146] S221, recording the movement trajectory of the operation elements and obtaining the transition form of the operation elements in the movement trajectory, the transition form including an arrow form, a palm form and a vertical line form.

[0147] It can be understood that the movement trajectory of the operation elements is recorded, wherein the operation elements can be mouse icons in a webpage; at the same time, the form change of the operation elements (such as the moment of arrow→palm) is captured by image recognition to mark the time point and coordinates of the form transition; the trajectory data and the form data are associated, wherein the transition form is the transition of the mouse form, such as the existence of the arrow form in the unnecessary area, the existence of the vertical line form when located in the input area, and the existence of the palm form when located in the component area that can trigger, such as the payment button.

[0148] S222, performing segmented processing on the movement trajectory based on the transition form to obtain a non-interactive trajectory of the arrow form, a trigger trajectory of the palm form and an input trajectory of the vertical line form.

[0149] It can be understood that the entire moving track is divided into multiple segments by various transition forms: the non-interactive track corresponding to the arrow form, the trigger track corresponding to the palm form, and the input track corresponding to the vertical line form; each segment of the track corresponds to only one form, which provides a clear object for subsequent customized analysis of the limitation type (such as trigger delay and input lag) of each stage, see Figure 5 , the input track process of the mouse from the arrow to the palm.

[0150] S223, respectively, custom analysis is performed on the non-interactive track, the trigger track, and the input track to obtain the deformation deviation coefficient of the non-interactive track, the trigger deviation coefficient of the trigger track, and the input deviation coefficient of the input track.

[0151] It can be understood that the time interval of the continuous coordinate points in the non-interactive track is analyzed, the number of pixel points of the lag is calculated, and the deformation deviation coefficient is converted; the key delay color change in the trigger track is calculated to obtain the trigger deviation coefficient; the automatic jumping of the cursor in the input track without input on the keyboard is calculated and converted into the input deviation coefficient.

[0152] In some embodiments, the step S223 (respectively, custom analysis is performed on the non-interactive track, the trigger track, and the input track to obtain the deformation deviation coefficient of the non-interactive track, the trigger deviation coefficient of the trigger track, and the input deviation coefficient of the input track) includes S2231-S2238:

[0153] S2231, the input area and the trigger area in the loading interface are taken as the interactive area, the intersection of the non-interactive track and the interactive area is counted, and the deformation deviation number is obtained.

[0154] Among them, the interactive area refers to a specific area in the loading interface that supports user interaction, including the input area (such as a text box) and the trigger area (such as a button); the intersection area refers to the overlapping part of the non-interactive track and the interactive area; the deformation deviation number refers to the total number of pixel points in the intersection area (reflecting the range of the non-interactive track entering the interactive area by mistake).

[0155] It is not difficult to understand that the arrow shape is in the non-interactive area, so if the track intersects with the keys and the input area, it means that the arrow form has not been timely changed to the palm and vertical line state, and therefore the more the number of pixel points corresponding to the intersection area, the higher the delay of the platform, and the platform may limit it.

[0156] S2232, based on the ratio of the deformation deviation number to the preset deviation number, the deformation deviation coefficient is obtained.

[0157] Among them, the preset deviation number refers to the maximum acceptable deviation pixel point set according to the size of the interactive area, which can be set in advance by a person according to the actual situation.

[0158] It can be understood that the number of deformation deviations is divided by the preset number of deviations to obtain a deformation deviation coefficient.

[0159] S2233, the color changing time of the trigger area where the trigger trajectory is located is counted from the trigger time of the trigger trajectory, and the trigger change duration is obtained.

[0160] Wherein, the trigger time refers to the time point when the trigger input device (such as mouse) triggers, such as mouse click; the color changing time refers to the time point when the color of the trigger area (such as button) starts to change; the trigger change duration refers to the time difference between the two (reflecting the trigger response delay). It is not difficult to understand that only the delay condition after the trigger input device triggers the button within the trigger trajectory time period is identified, and the processing efficiency is improved.

[0161] S2234, the trigger deviation coefficient is obtained based on the ratio of the trigger change duration and the reference change duration.

[0162] Wherein, the reference change duration refers to the normal response duration set according to the complexity of the trigger area function (such as simple 0.2 seconds, complex 1 second, which can be set artificially according to actual situation); the trigger deviation coefficient refers to the ratio of the trigger change duration and the reference change duration (the higher the value, the more serious the delay).

[0163] S2235, selecting two adjacent input trajectories in the corresponding input area as the selected trajectories, the selected trajectory in front as the first trajectory, and the selected trajectory in the rear as the second trajectory.

[0164] Wherein, the input area can be a text input area such as opinion feedback in the webpage, and the input trajectory is the trajectory of the vertical cursor when inputting text.

[0165] It should be noted that during the input process, such as text input, if it is word-by-word input, continuous trajectories will be formed, and adjacent character cursors will be connected, if it is multi-character input cursor, it will appear jumping phenomenon, becoming multiple non-continuous input trajectories, and when the platform is limited, there will be a situation that the vertical cursor jumps without inputting text, therefore, the present scheme will distinguish and count this situation.

[0166] It can be understood that two adjacent input trajectories are selected as the selected trajectories, that is, the two adjacent trajectories in the jumping situation; the first trajectory is marked as the first trajectory, and the second trajectory is marked as the second trajectory; through the adjacent relationship, it provides clear object for analyzing whether the jump between the two trajectories is abnormal.

[0167] S2236, obtaining a jump period according to the start time of the second track corresponding input period and the end time of the first track corresponding input period.

[0168] Specifically, the end time of the first track input period and the start time of the second track input period are extracted, and the two times are used to obtain the jump period

[0169] S2237, when the input state of the key input device in the jump period is determined to be a dormant state, the corresponding jump period is regarded as an abnormal period.

[0170] Specifically, the key input device can be a keyboard or other input device. The state of the key input device (keyboard) in the jump period is obtained through the device log; if the state is dormant (no key operation), it is determined that the jump period is an abnormal period (because the user does not operate, but the cursor jumps, which may be caused by platform restriction), which increases the positioning cost of the user; by excluding user active operation, the input abnormality caused by the positioning platform is determined.

[0171] S2238, the length of the abnormal period is counted to obtain an abnormal jump length, and the ratio of the abnormal jump length to the reference jump length is used to obtain an input deviation coefficient.

[0172] It can be understood that the length of all abnormal periods is counted to obtain the abnormal jump length; the corresponding reference jump length is retrieved according to the input scene, wherein the reference jump length can be a length set artificially, which is a tolerable length; the abnormal jump length is divided by the reference jump length to obtain the input deviation coefficient; the larger the coefficient is, the longer the abnormal jump length is.

[0173] S224, obtaining an operation restriction coefficient according to the deformation deviation number, the trigger deviation coefficient and the input deviation coefficient.

[0174] It can be understood that the deformation deviation coefficient, the trigger deviation coefficient and the input deviation coefficient are added to obtain the operation restriction coefficient.

[0175] S23, obtaining a function restriction coefficient according to the sum of the loading restriction coefficient and the operation restriction coefficient.

[0176] The function restriction coefficient refers to the sum of the loading restriction coefficient and the operation restriction coefficient, and the higher the value is, the more serious the overall restriction of the platform on the loading and operation functions is, and there may be shielding and banning behaviors.

[0177] S3, obtaining a comprehensive restriction coefficient according to the recommendation restriction coefficient and the function restriction coefficient, and determining the banning identification result of each detection platform based on the comprehensive restriction coefficient.

[0178] In some embodiments, the step S3 (obtaining a comprehensive restriction coefficient of each detection platform according to the recommended restriction coefficient and the function restriction coefficient, and determining the ban identification result of each detection platform based on the comprehensive restriction coefficient) comprises S31-S33:

[0179] S31, obtaining a comprehensive restriction coefficient of each detection platform according to the sum of the recommended restriction coefficient and the function restriction coefficient.

[0180] It can be understood that the comprehensive restriction coefficient of each detection platform is obtained according to the sum of the recommended restriction coefficient and the function restriction coefficient.

[0181] It should be noted that the comprehensive restriction coefficient is calculated according to the sum of the recommended restriction coefficient and the function restriction coefficient, and the recommended restriction coefficient and the function restriction coefficient both have their own standard range, for example, the recommended restriction coefficient is 0, which is better, therefore, the subsequent scheme will determine the preset judgment coefficient according to the actual standard situation to compare the values, so as to comprehensively judge whether the detection platform has a restriction behavior in the loading process, wherein the preset judgment coefficient can be set by a person according to the actual situation.

[0182] S32, when the comprehensive restriction coefficient is less than the preset judgment coefficient, determining that the ban identification result of the corresponding detection platform is a non-screening ban behavior.

[0183] S33, when the comprehensive restriction coefficient is greater than or equal to the preset judgment coefficient, determining that the ban identification result of the corresponding detection platform is a screening ban behavior.

[0184] It is not difficult to understand that when the comprehensive restriction coefficient is too large and exceeds the preset judgment coefficient, it is determined that the corresponding detection platform has a restriction behavior.

[0185] On the basis of the above-mentioned embodiments, the following steps are further included:

[0186] The actual jump times of the same jump component in the loading interface corresponding to each detection platform are counted, and the standard jump times of the corresponding jump component are called.

[0187] The jump component can be a component that triggers a jump, such as a payment button. The platform prevents payment of the same type of competing product, increases the jump times of the payment page, and increases the time cost of the user.

[0188] A selection coefficient is obtained according to the ratio of the actual jump times and the standard jump times.

[0189] The pop-up window times in the loading interface corresponding to each detection platform are obtained, and the detection platform with a pop-up window time greater than a preset time or a selection coefficient greater than 1 is selected as an abnormal interference platform.

[0190] It is understandable that the platform with frequently jumping advertisement or abnormal jump will be selected as the abnormal interference platform.

[0191] The present application also provides a readable storage medium, wherein the readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method provided by the various embodiments.

[0192] The readable storage medium can be a computer storage medium or a communication medium. The communication medium includes any medium that facilitates the transfer of computer program from one place to another. The computer storage medium can be any available medium that can be accessed by a general or special purpose computer. For example, the readable storage medium is coupled to the processor, so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be an integral part of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (ASIC). In addition, the ASIC can be located in a user equipment. Of course, the processor and the readable storage medium can also exist as discrete components in a communication device. The readable storage medium can be a read only memory (ROM), a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk and an optical data storage device, etc.

[0193] The present application also provides a program product, which includes execution instructions stored in a readable storage medium. At least one processor of a device can read the execution instructions from the readable storage medium, and the at least one processor executes the execution instructions to make the device implement the method provided by the various embodiments.

[0194] In the embodiment of the above device, it should be understood that the processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC) and the like. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor and the like. The steps of the method disclosed in the present application can be directly embodied as the execution of the hardware processor, or the execution of the combination of the hardware and software modules in the processor.

[0195] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit the present application; although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions recorded in the above embodiments can be modified, or some or all of the technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A platform shielding method for identifying bad behavior, characterized in that, The method comprises the following steps: controlling multiple access terminals to search for search words in each detection platform based on the identified plug-in, obtaining a search interface of each detection platform, performing relevance identification on the search interface, obtaining a recommendation restriction coefficient, comprising: counting the number of interactive texts in the search interface to obtain a recommended number; identifying interactive texts in the search interface that do not have search words as interference texts, and counting the number of interference texts to obtain an interference number; obtaining a recommendation restriction coefficient according to the ratio of the interference number and the recommended number; triggering processing on the same interactive texts in each search interface according to the search words to obtain a loading interface, performing interactive restriction identification on the loading interface to obtain a function restriction coefficient, comprising: performing loading restriction identification on the loading elements in the loading interface to obtain a loading restriction coefficient; performing operation restriction identification on the operation elements in the loading interface to obtain an operation restriction coefficient, comprising: recording the movement trajectory of the operation elements and obtaining the transformation form of the operation elements in the movement trajectory, the transformation form comprising an arrow form, a palm form and a vertical line form; segmenting the movement trajectory based on the transformation form to obtain a non-interactive trajectory of the arrow form, a trigger trajectory of the palm form and an input trajectory of the vertical line form; performing customized analysis on the non-interactive trajectory, the trigger trajectory and the input trajectory respectively to obtain a deformation deviation coefficient of the non-interactive trajectory, a trigger deviation coefficient of the trigger trajectory and an input deviation coefficient of the input trajectory; obtaining an operation restriction coefficient according to the deformation deviation number, the trigger deviation coefficient and the input deviation coefficient; obtaining a function restriction coefficient according to the sum of the loading restriction coefficient and the operation restriction coefficient; obtaining a comprehensive restriction coefficient according to the recommendation restriction coefficient and the function restriction coefficient, and determining a ban identification result of each detection platform based on the comprehensive restriction coefficient.

2. The method of claim 1, wherein: the loading restriction coefficient is obtained by performing loading restriction identification on the loading elements in the loading interface, comprising: identifying each loading element in the loading interface, the loading element comprising an image element, a video element and a component element; performing loading time identification on the image element in the loading interface to obtain an image restriction coefficient; performing loading deviation identification on the video element in the loading interface to obtain a video restriction coefficient; performing loading interference identification on the component element in the loading interface to obtain a component restriction coefficient; obtaining the loading restriction coefficient according to the sum of the image restriction coefficient, the video restriction coefficient and the component restriction coefficient.

3. The method of claim 2, wherein: the image restriction coefficient is obtained by performing loading time identification on the image element in the loading interface, comprising: obtaining an image reference time length of the same image element in each loading interface, and an image loading time length, obtaining an image difference value of each image element based on the ratio of the image loading time length and the image reference time length; counting the image difference value of each detection platform to obtain a platform image difference value, obtaining the number of image elements in the corresponding loading interface, and obtaining the image restriction coefficient according to the ratio of the platform image difference value and the number of images.

4. The method of claim 2, wherein the loading bias identification of the video elements in the loading interface to obtain a video restriction coefficient comprises: starting the video element, determining an image playing time length of a current playing time length corresponding playing frame and an audio playing time length of an audio frame; determining a playing bias time length of the image playing time length and the audio playing time length, and a progress bias time length corresponding to the current playing time length; obtaining a video bias time length based on a sum value of the playing bias time length and the progress bias time length, and obtaining a video bias coefficient based on a ratio of the video bias time length and the current playing time length; summing up the video bias coefficients of each loading interface to obtain a platform video difference value; obtaining the video restriction coefficient based on a ratio of the platform video difference value and a number of the video elements in the corresponding loading interface.

5. The method of claim 2, wherein the loading interference identification of the component elements in the loading interface to obtain a component restriction coefficient comprises: obtaining a background area where the component element is located and a component area corresponding to the component element; identifying a pixel difference value between the component area and the background area, and obtaining a reference difference value of the corresponding component area; obtaining a pixel difference value of each component element based on a ratio of the reference difference value and the corresponding pixel difference value, and counting a component number of the component element corresponding to each detection platform; summing up the pixel difference values in each detection platform to obtain a platform pixel difference value of each detection platform, and obtaining the component restriction coefficient based on a ratio of the platform pixel difference value and the component number.

6. The method of claim 1, wherein the customized analysis of the non-interactive trajectory, the trigger trajectory and the input trajectory respectively to obtain a deformation bias coefficient of the non-interactive trajectory, a trigger bias coefficient of the trigger trajectory and an input bias coefficient of the input trajectory comprises: taking an input area and a trigger area in the loading interface as an interactive area, counting an intersection of the non-interactive trajectory and the interactive area to obtain a deformation bias number; obtaining the deformation bias coefficient based on a ratio of the deformation bias number and a preset bias number; taking a trigger time when a trigger input device in the trigger trajectory as a starting point to count a color change time of a trigger area where the trigger trajectory is located to obtain a trigger change time length; obtaining the trigger bias coefficient based on a ratio of the trigger change time length and a reference change time length; selecting two adjacent input trajectories in a corresponding input area as selected trajectories, taking a selected trajectory in front as a first trajectory and a selected trajectory in back as a second trajectory; obtaining a jump time period based on a starting time of an input time period corresponding to the second trajectory and an ending time of an input time period corresponding to the first trajectory; taking a corresponding jump time period as an abnormal time period when an input state of a key input device in the jump time period is a dormant state; counting a length of the abnormal time period to obtain an abnormal jump time length, and obtaining the input bias coefficient based on a ratio of the abnormal jump time length and a reference jump time length.

7. The method of claim 1, wherein ​ ​ ​ The comprehensive restriction coefficient is obtained according to the recommended restriction coefficient and the function restriction coefficient, and the ban identification result of each detection platform is determined based on the comprehensive restriction coefficient, including: The comprehensive restriction coefficient of each detection platform is obtained according to the sum of the recommended restriction coefficient and the function restriction coefficient; When the comprehensive restriction coefficient is less than a preset judgment coefficient, it is determined that the ban identification result of the corresponding detection platform is a non-screening ban behavior; When the comprehensive restriction coefficient is greater than or equal to the preset judgment coefficient, it is determined that the ban identification result of the corresponding detection platform is a screening ban behavior.

Images