User behavior analysis method, device, equipment, and storage medium

Abstract

The present invention relates to the field of big data, and discloses a user behavior analysis method, device, equipment and storage medium. The method includes: exposing and marking the client interface that meets the exposure marking conditions, and collecting user operation information; parsing the user operation information within a preset target time period to obtain a user operation image and operation feature information; identifying and extracting the valid area associated with the corresponding operation feature information in the user operation image, and then using a preset first recognition method to identify the user morphological features of the client interface, and using a preset second recognition method to identify the user behavior features corresponding to the client interface; performing a fusion feature analysis on the user morphological features and the user behavior features, obtaining the user preference information of the client interface and pushing it. The present invention improves the accuracy of user behavior analysis, reduces invalid user behavior exposure, and is more in line with user preferences when used for subsequent APP demand development.

Description

Technical Field

[0001] The present invention relates to the field of big data, and in particular to a user behavior analysis method, device, equipment and storage medium. Background Art

[0002] In today's mobile internet landscape, apps are the most representative. With the increasing variety and number of apps, competition in the industry is also becoming increasingly fierce. Among so many apps, apps need to be recognized and frequently used by users, rather than just being left on the client interface after a single use, or even uninstalled. To create user stickiness, you need to analyze user behavior through big data, find content that users are interested in, and push similar content. However, in reality, since each app is initially developed to solve specific problems, it is only used in specific scenarios. Outside of these scenarios, the app will rarely be opened, used, or deleted.

[0003] Most existing apps use manual dotting to record user behavior, averaging exposures when a module is exposed. However, this method cannot accurately record whether a user is genuinely interested in the app. Alternatively, if a user slides a module onto the screen but only slides out a few pixels, the recorded user behavior may be distorted. Alternatively, if a user quickly browses, the module enters and leaves the screen almost instantly, requiring exposure dotting based on the module exposure, resulting in distorted reddish-brown recordings. Consequently, existing mobile-based user behavior capture methods suffer from low accuracy. Summary of the Invention

[0004] The main purpose of the present invention is to solve the technical problem of low accuracy of existing mobile-based user behavior capture methods.

[0005] A first aspect of the present invention provides a user behavior analysis method, comprising: when monitoring that the user sliding state corresponding to the client interface meets the preset exposure marking condition, collecting user operation information of the user interacting with the current client interface; parsing the collected user operation information to obtain multiple user operation images and operation feature information corresponding to each of the user operation images; identifying and extracting the area associated with the corresponding operation feature information in each of the user operation images to obtain the effective area of the client interface exposure marking, and marking the exposure information of each of the effective areas; according to each of the effective areas and the corresponding exposure information, using a preset first recognition method to identify the user morphological features corresponding to the client interface, and according to each of the effective areas and the corresponding exposure information, using a preset second recognition method to identify the user behavior features corresponding to the client interface; performing fusion feature analysis on the user morphological features and the user behavior features to obtain user preference information corresponding to the client interface and push it.

[0006] Optionally, in a first implementation method of the first aspect of the present invention, the parsing of the collected user operation information to obtain multiple user operation images and operation feature information corresponding to each of the user operation images includes: traversing the user operation information collected within a preset target time period, determining multiple collected data types contained in the user operation information, and classifying the operation indicators contained in the user operation information according to each of the collected data types; parsing each of the operation indicators into a corresponding preset folder according to the results of the classification of each of the operation indicators to obtain a first data set and a second data set; filtering out abnormal user operation information in the first data set and the second data set respectively, and obtaining the user operation images in the first data set and the operation feature information in the second data set, wherein the operation feature information includes operation time information, operation range information and operation pressure information.

[0007] Optionally, in a second implementation method of the first aspect of the present invention, the identifying and extracting of areas associated with corresponding operation feature information in each of the user operation images to obtain effective areas for exposure marking of the client interface, and marking exposure information of each effective area include: according to the operation range information, respectively delineating the operation areas in the user operation images, and calculating the minimum circumscribed circle of each of the operation areas; respectively expanding the minimum circumscribed circle of each of the operation areas by a preset size to obtain a new operation area, and extracting the corresponding new operation area in each of the user operation images to obtain multiple effective areas for exposure marking of the client interface; according to the operation time information and the operation pressure information, marking the operation time and operation pressure corresponding to each of the effective areas to obtain exposure information corresponding to each of the effective areas.

[0008] Optionally, in a third implementation manner of the first aspect of the present invention, the user morphological features corresponding to the client interface are identified by using a preset first identification method based on each of the effective areas and the corresponding exposure information, including: identifying the browsing content corresponding to each of the effective areas, and classifying the browsing content to obtain multiple types of similar browsing content; calculating the exposure weight corresponding to each of the similar browsing contents according to the exposure information, and sorting each of the similar browsing contents from large to small according to the exposure weight; calculating the user interest level of each of the similar browsing contents according to the sorting results of each of the similar browsing contents, and constructing the user morphological features corresponding to the client interface according to each of the user interest levels.

[0009] Optionally, in a fourth implementation method of the first aspect of the present invention, the exposure dot conditions include: the stop time of the user sliding state exceeds a preset time preset; when the user sliding state is in a stopped state, the operation range corresponding to the client interface is within a preset operation reference range; when the user sliding state is in a stopped state, the operation pressure corresponding to the client interface is greater than a preset operation pressure threshold.

[0010] Optionally, in a fifth implementation of the first aspect of the present invention, when it is monitored that the user sliding state corresponding to the client interface meets the preset exposure and dotting conditions, the user operation information of the user interacting with the current client interface is collected, including: when it is monitored that the user sliding state corresponding to the client interface meets the preset exposure and dotting conditions, multiple frames of images of the current client interface are continuously captured according to a preset period, and the user sliding state corresponding to each frame image is collected; each frame image is used as the user operation image corresponding to the current client, and according to the user sliding state, the user operation features corresponding to each frame image are identified to obtain the operation feature information corresponding to each frame image, wherein the user operation information includes the user operation image and the operation feature information.

[0011] The second aspect of the present invention provides a user behavior analysis device, including: an acquisition module, which is used to collect user operation information of the user interacting with the current client interface when it is monitored that the user sliding state corresponding to the client interface meets the preset exposure marking conditions; a parsing module, which is used to parse the collected user operation information to obtain multiple user operation images and operation feature information corresponding to each of the user operation images; a segmentation module, which is used to identify and extract the areas associated with the corresponding operation feature information in each of the user operation images, obtain the effective areas of the exposure marking of the client interface, and mark the exposure information of each of the effective areas; an identification module, which is used to identify the user morphological features corresponding to the client interface according to each of the effective areas and the corresponding exposure information, using a preset first identification method, and identify the user behavior features corresponding to the client interface according to each of the effective areas and the corresponding exposure information, using a preset second identification method; a fusion module, which is used to perform fusion feature analysis on the user morphological features and the user behavior features, obtain user preference information corresponding to the client interface, and push it.

[0012] Optionally, in a first implementation manner of the second aspect of the present invention, the parsing module includes: a traversal unit, used to traverse the user operation information collected within a preset target time period, determine multiple collected data types contained in the user operation information, and classify the operation indicators contained in the user operation information according to each collected data type; a classification unit, used to parse each operation indicator into a corresponding preset folder according to the classification result of each operation indicator, to obtain a first data set and a second data set; a screening unit, used to screen out abnormal user operation information in the first data set and the second data set, respectively, to obtain the user operation image in the first data set and the operation feature information in the second data set, wherein the operation feature information includes operation time information, operation range information and operation pressure information.

[0013] Optionally, in a second implementation method of the second aspect of the present invention, the segmentation module includes: a division unit, used to delineate the operation areas in the user operation image according to the operation range information, and calculate the minimum circumscribed circle of each operation area; an expansion unit, used to expand the minimum circumscribed circle of each operation area by a preset size to obtain a new operation area, and extract the corresponding new operation area in each user operation image to obtain multiple effective areas for exposure marking of the client interface; a labeling unit, used to label the operation time and operation pressure corresponding to each effective area according to the operation time information and the operation pressure information, and obtain the exposure information corresponding to each effective area.

[0014] Optionally, in a third implementation of the second aspect of the present invention, the identification module includes: an identification unit for identifying the browsing content corresponding to each of the valid areas, and classifying the browsing content to obtain multiple types of similar browsing content; a calculation unit for calculating the exposure weight corresponding to each of the similar browsing contents according to the exposure information, and sorting each of the similar browsing contents from large to small according to the exposure weight; a construction unit for calculating the user interest level of each of the similar browsing contents according to the sorting results of each of the similar browsing contents, and constructing user morphological features corresponding to the client interface according to each of the user interest levels.

[0015] Optionally, in a fourth implementation method of the second aspect of the present invention, the exposure dot conditions include: the stop time of the user sliding state exceeds a preset time preset; when the user sliding state is in a stopped state, the operation range corresponding to the client interface is within a preset operation reference range; when the user sliding state is in a stopped state, the operation pressure corresponding to the client interface is greater than a preset operation pressure threshold.

[0016] Optionally, in a fifth implementation of the second aspect of the present invention, the acquisition module is further used to: when it is monitored that the user sliding state corresponding to the client interface meets the preset exposure marking conditions, continuously capture multiple frames of images of the current client interface according to a preset period, and acquire the user sliding state corresponding to each frame image; use each frame image as the user operation image corresponding to the current client, and identify the user operation features corresponding to each frame image based on the user sliding state, and obtain the operation feature information corresponding to each frame image, wherein the user operation information includes the user operation image and the operation feature information.

[0017] The third aspect of the present invention provides a user behavior analysis device, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the user behavior analysis device executes the above-mentioned user behavior analysis method.

[0018] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, enable the computer to execute the above-mentioned user behavior analysis method.

[0019] In the technical solution provided by the present invention, by setting exposure marking conditions, the user sliding status on the client interface is screened, and the user operation information collected at this time is more in line with the user's actual operation behavior; then the user operation image and operation feature information are parsed from the user operation information, and the effective area of the client interface exposure marking is segmented based on this; then, through the exposure information of the effective area, the user behavior characteristics and user morphology characteristics are identified to characterize the user's current user morphology and user behavior; finally, the user morphology characteristics and user behavior characteristics are fused and analyzed, that is, the user preference information of the client interface is analyzed in combination with the user morphology and user morphology. On the one hand, the exposure marking conditions are used to screen the timing of exposure marking of the client interface, and on the other hand, the user morphology and user behavior are fully analyzed, so that the user preference information obtained by analysis is more in line with the user's actual interests and habits. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a schematic diagram of a first embodiment of a user behavior analysis method according to an embodiment of the present invention;

[0021] Figure 2 Schematic diagram of a second embodiment of the user behavior analysis method according to an embodiment of the present invention;

[0022] Figure 3 Schematic diagram of a third embodiment of the user behavior analysis method according to an embodiment of the present invention;

[0023] Figure 4 Schematic diagram of an embodiment of a user behavior analysis device according to an embodiment of the present invention;

[0024] Figure 5 Schematic diagram of another embodiment of the user behavior analysis device according to the embodiment of the present invention;

[0025] Figure 6 Schematic diagram of an embodiment of a user behavior analysis device in an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The embodiments of the present invention provide a user behavior analysis method, apparatus, device, and storage medium, which perform exposure marking on client interfaces that meet exposure marking conditions and collect user operation information; analyze user operation information within a preset target time period to obtain user operation images and operation feature information; segment each user operation image to obtain and mark the effective area for exposure marking of the client interface; then use a preset first recognition method to identify the user morphological features of the client interface, and use a preset second recognition method to identify the user behavior features corresponding to the client interface; perform fusion feature analysis on the user morphological features and user behavior features to obtain and push user preference information for the client interface. The present invention improves the accuracy of user behavior analysis, reduces invalid user behavior exposure, and is more in line with user preferences when used in subsequent APP demand development.

[0027] The terms "first," "second," "third," "fourth," and the like (if any) in the description and claims of the present invention and in the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to these processes, methods, products, or apparatus.

[0028] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 The first embodiment of the user behavior analysis method in the embodiment of the present invention includes:

[0029] 101. When it is detected that the user sliding state corresponding to the client interface meets the preset exposure marking condition, user operation information of the user interacting with the current client interface is collected;

[0030] It is understandable that the execution subject of the present invention may be a user behavior analysis device, or a terminal or a server, which is not limited here. The embodiment of the present invention is described by taking a server as the execution subject as an example.

[0031] In this embodiment, when the user operates the client interface, not every interface browsed by the user is captured, nor is every APP in the captured interface analyzed. Instead, by pre-setting exposure marking conditions, the distortion rate of the recorded user behavior is reduced and the authenticity of the user interest capture is increased.

[0032] According to the user's operating habits for the APP they are interested in, set corresponding exposure marking conditions. For example, the user will focus on the APP module they are interested in. Once the focus occurs, the user's sliding state on the client interface will pause. At this time, the user operation information collected will be more accurate.

[0033] Configure a sliding component on the client interface to monitor the user's sliding status on the client interface. Sliding component options include: RecyclerView, UIButton, ScrollView, or a custom sliding component. When the sliding component monitors the user's sliding status and meets the preset exposure marking conditions, the client interface can be exposed and marked.

[0034] During the exposure and marking of the client interface, user operation information on the client interface is also collected, including: user operation information of the current client interface and the corresponding stop time of the user sliding state, user operation range and user operation pressure on the client interface when the user sliding state is in the stop state, etc.

[0035] In addition, the client interface refers to the UI (User Interface) interactive interface of the mobile client. The mobile client may include: a mobile phone, a tablet computer, a laptop computer, etc., but is not limited thereto and is not specifically limited here.

[0036] 102. Analyze the collected user operation information to obtain multiple user operation images and operation feature information corresponding to each user operation image;

[0037] In this embodiment, each time the client interface is exposed and marked, a screenshot of the client interface is taken, a user operation image of the current user's operation is recorded, and a corresponding timestamp is recorded, wherein the image includes the APP module displayed on the current client interface.

[0038] In addition, the user operation information includes the stop time of the user sliding state, the user operation range and user operation pressure on the client interface when the user sliding state is in the stop state, etc., and records the timestamp of each information content operation in the user operation information, and records them one by one with the aforementioned user operation image.

[0039] By analyzing the action, each user operation image captured during exposure and marking is extracted, as well as the stop time, user operation range and user operation pressure corresponding to the timestamp of each user operation image, which are used as the user's operation feature information on the client interface.

[0040] 103. Identify and extract the area associated with the corresponding operation feature information in each user operation image, obtain the effective area of the client interface exposure mark, and mark the exposure information of each effective area;

[0041] In this embodiment, the user operation image may contain multiple APP modules, but not every APP module is focused on and browsed by the user. Therefore, the user operation image is segmented here by the analyzed operation feature information, and the area where the user actually operates in the user operation image is determined as the effective area for exposure marking of the client interface.

[0042] 104. Based on each valid area and the corresponding exposure information, a preset first recognition method is used to identify the user morphological characteristics corresponding to the client interface, and based on each valid area and the corresponding exposure information, a preset second recognition method is used to identify the user behavioral characteristics corresponding to the client interface;

[0043] In this embodiment, the user morphological characteristics and user behavioral characteristics are analyzed respectively through the effective area of exposure points and the corresponding exposure information, wherein the user morphological characteristics refer to the APP content module that the user is interested in, and the user behavioral characteristics refer to the APP type module that the user is interested in.

[0044] Among them, when using the first recognition method to identify user morphological characteristics, the browsing content of each valid area is first identified through an image recognition algorithm. The types of browsing content include text type, audio type, image type, video type, function type, etc. The exposure weight of each type of browsing content is calculated according to the operation time information and operation pressure information in the exposure information. Among them, the longer the operation time and the greater the operation pressure, the greater the exposure weight obtained, and the higher the user interest in similar browsing content of this type. Finally, the user morphological characteristics of the client interface are constructed through user portraits.

[0045] Among them, when the second recognition method is used to recognize the user behavior characteristics, the corresponding sliding windows are added to the operation time information and the operation pressure information in the exposure information, and the time sliding windows S are obtained. i Corresponding data segments and pressure sliding windows R i At the same time, the behavioral features of the obtained data segment are extracted to obtain the corresponding feature vector F1=={f s1 , f s2 ,……f si}, F2={f r1 , f r2 ,……f ri}, where i is the feature dimension; feature selection is performed on the feature vectors F1 and F2 through the dimensionality reduction model of linear discriminant analysis, and the j-dimensional feature subset X1={x s1 , xs2 ,……,x sj} and X2={x s1 , x s2 ,……,x sj Input feature subsets X1 and X2 into the behavior classification model and output user behavior categories, such as reading novels, chatting, making and receiving phone calls, playing games, watching movies, listening to music, etc.

[0046] 105. Perform fusion feature analysis on user morphological features and user behavioral features to obtain user preference information corresponding to the client interface and push it.

[0047] In this embodiment, the user morphological characteristics and user behavioral characteristics are analyzed respectively through the effective area of exposure points and the corresponding exposure information, wherein the user morphological characteristics refer to the APP content module that the user is interested in, and the user behavioral characteristics refer to the APP type module that the user is interested in.

[0048] Among them, when using the first recognition method to identify user morphological characteristics, the browsing content of each valid area is first identified through an image recognition algorithm. The types of browsing content include text type, audio type, image type, video type, function type, etc. The exposure weight of each type of browsing content is calculated according to the operation time information and operation pressure information in the exposure information. Among them, the longer the operation time and the greater the operation pressure, the greater the exposure weight obtained, and the higher the user interest in similar browsing content of this type. Finally, the user morphological characteristics of the client interface are constructed through user portraits.

[0049] Among them, when the second recognition method is used to recognize the user behavior characteristics, the corresponding sliding windows are added to the operation time information and the operation pressure information in the exposure information, and the time sliding windows S are obtained. i Corresponding data segments and pressure sliding windows R i At the same time, the behavioral features of the obtained data segment are extracted to obtain the corresponding feature vector F1=={f s1 , f s2 ,……f si}, F2={f r1 , f r2 ,……f ri}, where i is the feature dimension; feature selection is performed on the feature vectors F1 and F2 through the dimensionality reduction model of linear discriminant analysis, and the j-dimensional feature subset X1={x s1 , x s2 ,……,x sj} and X2={x s1 , x s2 ,……,x sjInput feature subsets X1 and X2 into the behavior classification model and output user behavior categories, such as reading novels, chatting, making and receiving phone calls, playing games, watching movies, listening to music, etc.

[0050] In an embodiment of the present invention, by setting exposure marking conditions, the user sliding status on the client interface is filtered, and the user operation information collected at this time is more consistent with the user's actual operation behavior; then, the user operation image and operation feature information are parsed from the user operation information, and the effective area of the client interface exposure marking is segmented based on this; then, through the exposure information of the effective area, the user behavior characteristics and user form characteristics are identified to characterize the user's current user form and user behavior; finally, the user form characteristics and user behavior characteristics are fused and analyzed, that is, the user preference information of the client interface is analyzed in combination with the user form and user form. On the one hand, the exposure marking conditions are used to filter the timing of exposure marking of the client interface, and on the other hand, the user form and user behavior are fully analyzed, so that the user preference information obtained by analysis is more in line with the user's actual interests and habits.

[0051] See also Figure 2 The second embodiment of the user behavior analysis method in the embodiment of the present invention includes:

[0052] 201. When it is detected that the user sliding state corresponding to the client interface meets the preset exposure marking condition, user operation information of the user interacting with the current client interface is collected;

[0053] 202. Traverse the user operation information collected within the preset target time period, determine multiple collected data types contained in the user operation information, and classify the operation indicators contained in the user operation information according to each collected data type;

[0054] In this embodiment, each piece of data in the user operation information is identified by a corresponding parameter variable, and the same type of collected data is identified by the same parameter variable. Specifically, the user operation information contains at least two types of collected data: images and image description data. The type of each collected data in the user operation information can be determined by traversing the parameter variables identified in the user operation information.

[0055] After determining the type of collected data, the corresponding collected data is classified, with the collected data identified by the image being classified into one type, and the collected information identified by the image description data being classified into another type. Specifically, the collected data identified by the image description data includes multiple operation indicators, including at least: the user's operation time on the client interface during the exposure marking period, the user's operation range on the image, and the user's operation pressure on the client interface.

[0056] 203. According to the classification results of each operation indicator, each operation indicator is parsed into a corresponding preset folder to obtain a first data set and a second data set;

[0057] In this embodiment, user operation information is parsed into a pre-set folder, and the operation indicators corresponding to the image are parsed into a first data set by conventionally decomposing the image into multiple pixels and describing the grayscale of each pixel. The operation indicators corresponding to the image description data are then parsed into a second data set. The second data set may include multiple sub-data sets, each storing a different type of operation indicator. For example, three sub-data sets may be used to store the user's operation time on the client interface, the user's operation range on the image, and the user's operation pressure on the client interface.

[0058] 204. Filter out abnormal user operation information in the first data set and the second data set respectively, and obtain the user operation image in the first data set and the operation characteristic information in the second data set, wherein the operation characteristic information includes operation time information, operation range information, and operation pressure information;

[0059] In this embodiment, for the sliding behavior of most users, the number of operation feature information contained in the second data set contained in a single sliding should be within a certain range, such as the interval [10, 100]. Here, a threshold is pre-set to filter out the second data set with too many or too few sampling points of the operation feature information, and the first data set corresponding to the second data set is filtered out.

[0060] A typical sliding motion should have a smooth trajectory. By calculating the instantaneous velocity angle between two consecutive pieces of operational signature information during each sliding motion, we can filter out sliding motions with sharp transitions. For example, if the instantaneous velocity angle between two consecutive pieces of operational signature information during a sliding motion is greater than 90 degrees, it is considered an abnormal sliding motion, and the first and second data sets corresponding to this abnormal sliding motion are deleted.

[0061] Next, a trajectory curve with multiple turns, such as the up and down or curved sliding of the finger without leaving the screen, contains multiple turns, which is not conducive to subsequent feature extraction. Therefore, the corresponding second data set and the corresponding first data set can be removed, and the number of turns can be set according to specific needs.

[0062] 205. Delimit the operation areas in the user operation image according to the operation range information, and calculate the minimum circumscribed circle of each operation area;

[0063] 206. Expand the minimum circumscribed circle of each operation area by a preset size to obtain a new operation area, and extract the new operation area corresponding to each user operation image to obtain multiple valid areas for exposure marking on the client interface;

[0064] 207. Mark the operation time and the operation pressure corresponding to each effective area according to the operation time information and the operation pressure information to obtain exposure information corresponding to each effective area;

[0065] In this embodiment, the user's actual operating area is preliminarily delineated based on the user's operating range information. The user's possible operating area is then determined by expanding the minimum circumscribed circle, thereby increasing the sample diversity of the valid area. Specifically, the size of the minimum circumscribed circle of the operating area is set based on the size of the client interface and can be proportional to the size of the client interface, for example, 1 / 32, 1 / 40, etc. of the client interface.

[0066] In addition, after expanding the effective area exposed by the client interface, the operation time information and operation pressure information in the operation feature information are further used to mark the effective area. The operation area, operation time, and operation pressure correspond one to one in the operation feature information. Here, the operation time information and operation pressure information are marked on the effective area corresponding to the operation range.

[0067] 208. Based on each valid area and the corresponding exposure information, a preset first recognition method is used to identify the user morphological characteristics corresponding to the client interface, and based on each valid area and the corresponding exposure information, a preset second recognition method is used to identify the user behavioral characteristics corresponding to the client interface;

[0068] 209. Perform fusion feature analysis on user morphological features and user behavioral features to obtain user preference information corresponding to the client interface and push it.

[0069] In an embodiment of the present invention, a first data set and a second data set are parsed from user operation information in detail. After filtering out abnormal user operation information, user operation images and operation feature information can be obtained. The user's user form and user behavior on the current client interface are identified through operation time information, operation range information and operation pressure information, and the user's operational interest habits are analyzed more comprehensively. In addition, by limiting the effective area of user operation, the scope of the user's actual operation on the client interface is determined, that is, the applications or content that the user is actually interested in are further screened, thereby improving the accuracy of the user's interest analysis on the client interface.

[0070] See also Figure 3 A third embodiment of the user behavior analysis method according to the present invention includes:

[0071] 301. When it is detected that the user sliding state corresponding to the client interface meets the preset exposure marking condition, multiple frames of images of the current client interface are continuously captured according to a preset period, and the user sliding state corresponding to each frame of image is collected;

[0072] 302. Using each frame image as a user operation image corresponding to the current client, and identifying the user operation feature corresponding to each frame image based on the user sliding state, to obtain operation feature information corresponding to each frame image, wherein the user operation information includes the user operation image and the operation feature information;

[0073] In this embodiment, when the user is using the mobile client, when the user's sliding state of the client interface meets the exposure marking conditions, multiple frames of images of the current client interface are continuously captured within a fixed time length, such as 10ms, 20ms, 50ms, etc., and each frame of the image is set according to actual business needs, which can be set to 0.1ms, 0.2ms, 0.5ms, 1ms, etc.

[0074] In addition, when capturing each frame of image, the user's sliding status in the current frame of image is also collected, such as single-click, double-click, slide, long press, etc. Based on the user's sliding status corresponding to multiple frames of image, the user's continuous operation characteristics on the current client interface can be obtained, thereby identifying the operation characteristic information of each frame of image.

[0075] In this embodiment, the exposure marking conditions are set to reduce the client interface distortion rate of the exposure marking. In this case, the collected user operation information is more consistent with the user's actual operation. The specific exposure marking conditions are as follows:

[0076] (1) The user's sliding stop time exceeds the preset time;

[0077] In this embodiment, when the user's sliding state is in a stopped state, it means that the current user's sliding state on the client interface has also stopped. When the stop time exceeds the preset time threshold, for example, the stop time exceeds 0.2s, 0.5s, 0.8s, or 1s, the currently displayed client interface is exposed and marked. This exposure and marking condition can solve the problem of invalid exposure and marking content of the page due to the user's quick sliding, the APP module quickly entering the screen, and then quickly sliding out of the screen.

[0078] (2) When the user's sliding state is in a stopped state, the corresponding operation range of the client interface is within the preset operation reference range;

[0079] In this embodiment, when the user's sliding state is in a stopped state, the user's operation state is recorded during the stopped state, and the user's contact area on the client interface is recorded. A plane coordinate axis can be set on the client interface, and the area contacted by the user terminal during the stopped state is converted into a corresponding coordinate set. The coordinate set represents the corresponding operation range of the client interface.

[0080] Preferably, the center point of each APP module can be calculated by obtaining the plane coordinate axis positions of all visible APP modules on the screen. Whether the center point is exposed on the screen determines whether the user's operating range is within a preset operating reference range, thereby filtering out a smaller list of APP modules that the user may be interested in. This exposure marking condition can filter out APP modules that are only partially exposed or only have one edge exposed on the client interface. In this case, the user cannot see the main content of the APP module, and the APP module is considered invalid exposure marking content.

[0081] (3) When the user's sliding state is in a stopped state, the operation pressure corresponding to the client interface is greater than the preset operation pressure threshold.

[0082] In this embodiment, an acceleration and touch screen sensor are configured in the client interface to detect the user's operating pressure on the client in real time, and during the period when the user slides and stops, the continuous user operating pressure during the period is recorded, including the pressing duration and operating pressure of the user operation, which can be recorded separately through the time set and the operating pressure set: pressing duration T = {t1, t2, ..., t i} and the corresponding operating pressure P={p1,p2,……,p i}.

[0083] By designing an operation pressure threshold, we can detect whether the user has only browsed the APP module and stopped at the corresponding client interface, but has not performed any continuous operations on the client interface. In this case, the client interface of the exposure marking is not accurate and cannot reflect whether the user is really interested in the APP module corresponding to the current user interface. Therefore, this exposure marking condition can be used to exclude it.

[0084] 303. Analyze the collected user operation information to obtain multiple user operation images and operation feature information corresponding to each user operation image;

[0085] 304. Identify and extract the area associated with the corresponding operation feature information in each user operation image, obtain the effective area of the client interface exposure mark, and mark the exposure information of each effective area;

[0086] 305. Identify the browsing content corresponding to each valid area, and classify the browsing content to obtain multiple types of similar browsing content;

[0087] 306. Calculate exposure weights corresponding to each similar browsing content based on the exposure information, and sort the similar browsing content from largest to smallest based on the exposure weights;

[0088] 307. Calculate the user interest level of each similar browsing content based on the ranking results of each similar browsing content, and construct a user profile feature corresponding to the client interface based on each user interest level;

[0089] In this embodiment, for different types of similar browsing content, the operation time and operation pressure of each browsing content are counted based on the operation time information and operation pressure information, and the operation time and operation pressure are sorted into two sets respectively, which are input into two radial basis function neural networks respectively to calculate the exposure weight of each type of similar browsing content, so as to determine the user interest in each similar browsing content.

[0090] Specifically, the operation time information of each valid area of the same type is converted into a time feature vector and recorded in the set T = (t1, t2, ..., t n ), the operating pressure information corresponding to each effective area is converted into a pressure feature vector and recorded in the set P = (p1, p2, ..., p n ), two independent radial basis function neural networks are used to train the effective area and the corresponding sets T and P respectively, and the training results O are output. T and O P , calculate the average error of each radial basis function neural network according to the training structure, recorded as E T and E P , get exposure weight W = O T *E T +O P *E P .

[0091] 308. Based on each valid area and the corresponding exposure information, a preset second identification method is used to identify the user behavior characteristics corresponding to the client interface;

[0092] 309. Perform fusion feature analysis on the user morphological features and the user behavioral features to obtain the user preference information corresponding to the client interface and push it.

[0093] In an embodiment of the present invention, the identification of user morphological characteristics through a first identification method and the identification of user behavioral characteristics through a second identification method are introduced in detail. By combining the user morphological characteristics and the user behavioral characteristics, it is further determined which content and types of applications the user is interested in and the degree of interest. The user's operating habits and interests in the client interface are analyzed from both horizontal and vertical aspects, thereby improving the depth of user behavior analysis and obtaining more accurate user behavior analysis results.

[0094] The above describes the user behavior analysis method in the embodiment of the present invention. The following describes the user behavior analysis device in the embodiment of the present invention. Figure 4In one embodiment of the present invention, a user behavior analysis device includes:

[0095] The collection module 401 is used to collect user operation information of the user interacting with the current client interface when monitoring that the user sliding state corresponding to the client interface meets the preset exposure marking condition;

[0096] The parsing module 402 is used to parse the collected user operation information to obtain multiple user operation images and operation feature information corresponding to each of the user operation images;

[0097] The segmentation module 403 is used to identify and extract the area associated with the corresponding operation feature information in each user operation image, obtain the effective area of the client interface exposure mark, and mark the exposure information of each effective area;

[0098] an identification module 404 configured to identify user morphological features corresponding to the client interface using a preset first identification method based on each of the valid areas and the corresponding exposure information, and to identify user behavioral features corresponding to the client interface using a preset second identification method based on each of the valid areas and the corresponding exposure information;

[0099] The fusion module 405 is used to perform fusion feature analysis on the user form feature and the user behavior feature, obtain user preference information corresponding to the client interface, and push it.

[0100] In an embodiment of the present invention, by setting exposure marking conditions, the user sliding status on the client interface is filtered, and the user operation information collected at this time is more consistent with the user's actual operation behavior; then, the user operation image and operation feature information are parsed from the user operation information, and the effective area of the client interface exposure marking is segmented based on this; then, through the exposure information of the effective area, the user behavior characteristics and user form characteristics are identified to characterize the user's current user form and user behavior; finally, the user form characteristics and user behavior characteristics are fused and analyzed, that is, the user preference information of the client interface is analyzed in combination with the user form and user form. On the one hand, the exposure marking conditions are used to filter the timing of exposure marking of the client interface, and on the other hand, the user form and user behavior are fully analyzed, so that the user preference information obtained by analysis is more in line with the user's actual interests and habits.

[0101] See also Figure 5 Another embodiment of the user behavior analysis device in the embodiment of the present invention includes:

[0102] The collection module 401 is used to collect user operation information of the user interacting with the current client interface when monitoring that the user sliding state corresponding to the client interface meets the preset exposure marking condition;

[0103] The parsing module 402 is used to parse the collected user operation information to obtain multiple user operation images and operation feature information corresponding to each of the user operation images;

[0104] The segmentation module 403 is used to identify and extract the area associated with the corresponding operation feature information in each user operation image, obtain the effective area of the client interface exposure mark, and mark the exposure information of each effective area;

[0105] an identification module 404 configured to identify user morphological features corresponding to the client interface using a preset first identification method based on each of the valid areas and the corresponding exposure information, and to identify user behavioral features corresponding to the client interface using a preset second identification method based on each of the valid areas and the corresponding exposure information;

[0106] The fusion module 405 is used to perform fusion feature analysis on the user form feature and the user behavior feature, obtain user preference information corresponding to the client interface, and push it.

[0107] Specifically, the parsing module 402 includes:

[0108] The traversal unit 4021 is configured to traverse the user operation information collected within a preset target time period, determine multiple types of collected data contained in the user operation information, and classify the operation indicators contained in the user operation information according to the types of collected data;

[0109] The classification unit 4022 is configured to parse each of the operation indicators into a corresponding preset folder according to the classification result of each of the operation indicators, to obtain a first data set and a second data set;

[0110] The screening unit 4023 is used to screen out abnormal user operation information in the first data set and the second data set respectively, and obtain the user operation image in the first data set and the operation feature information in the second data set, wherein the operation feature information includes operation time information, operation range information and operation pressure information.

[0111] Specifically, the segmentation module 403 includes:

[0112] a dividing unit 4031, configured to delimit the operation areas in the user operation image according to the operation range information, and calculate the minimum circumscribed circle of each operation area;

[0113] The expansion unit 4032 is configured to expand the minimum circumscribed circle of each operation area by a preset size to obtain a new operation area, and extract the new operation area corresponding to each user operation image to obtain multiple valid areas for exposure marking on the client interface;

[0114] The marking unit 4033 is configured to mark the operation time and the operation pressure corresponding to each of the effective areas according to the operation time information and the operation pressure information, so as to obtain exposure information corresponding to each of the effective areas.

[0115] Specifically, the identification module 404 includes:

[0116] an identification unit 4041 for identifying the browsing content corresponding to each of the valid areas and classifying the browsing content to obtain multiple types of similar browsing content;

[0117] The calculation unit 4042 is configured to calculate the exposure weight corresponding to each of the similar browsing contents according to the exposure information, and sort the similar browsing contents from largest to smallest according to the exposure weight;

[0118] The construction unit 4043 is configured to calculate the user interest level of each of the similar browsing contents according to the ranking results of the similar browsing contents, and to construct a user form feature corresponding to the client interface according to the user interest level.

[0119] Specifically, the exposure dotting conditions include:

[0120] The stop time of the user sliding state exceeds the preset time;

[0121] When the user sliding state is in a stopped state, the operation range corresponding to the client interface is within a preset operation reference range;

[0122] When the user sliding state is in a stopped state, the operation pressure corresponding to the client interface is greater than a preset operation pressure threshold.

[0123] Specifically, the acquisition module 401 is also used to: when it is monitored that the user sliding state corresponding to the client interface meets the preset exposure marking conditions, continuously capture multiple frames of images of the current client interface according to a preset period, and collect the user sliding state corresponding to each frame image; use each frame image as the user operation image corresponding to the current client, and identify the user operation features corresponding to each frame image according to the user sliding state, and obtain the operation feature information corresponding to each frame image, wherein the user operation information includes the user operation image and the operation feature information.

[0124] In an embodiment of the present invention, a first data set and a second data set are parsed from user operation information. After filtering out abnormal user operation information, user operation images and operation feature information can be obtained. The user's user form and user behavior on the current client interface are identified through operation time information, operation range information and operation pressure information, and the user's operational interest habits of the user behavior are analyzed more comprehensively. In addition, by limiting the effective area of the user operation, the scope of the user's actual operation on the client interface is determined, that is, the applications or content that the user is actually interested in are further filtered, thereby improving the accuracy of the analysis of the user's interest in the client interface. Then, it is also introduced in detail how to identify the user's form characteristics through the first recognition method and how to identify the user's behavior characteristics through the second recognition method. Combined with the user's form characteristics and user behavior characteristics, it is further determined which content and type of applications the user is interested in and the degree of interest. The user's operational habits and interests on the client interface are analyzed from both horizontal and vertical aspects, thereby improving the depth of the user behavior analysis and obtaining more accurate user behavior analysis results.

[0125] above Figure 4 and Figure 5 The user behavior analysis apparatus in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The user behavior analysis device in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0126] Figure 6 : is a structural diagram of a user behavior analysis device provided by an embodiment of the present invention. The user behavior analysis device 600 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 610 (for example, one or more processors) and a memory 620, and one or more storage media 630 (for example, one or more massive storage devices) for storing application programs 633 or data 632. Among them, the memory 620 and the storage medium 630 can be temporary storage or permanent storage. The program stored in the storage medium 630 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the user behavior analysis device 600. Furthermore, the processor 610 can be configured to communicate with the storage medium 630 to execute a series of instruction operations in the storage medium 630 on the user behavior analysis device 600.

[0127] The user behavior analysis device 600 may further include one or more power supplies 640, one or more wired or wireless network interfaces 650, one or more input and output interfaces 660, and / or one or more operating systems 631, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that Figure 6 The structure of the user behavior analysis device shown does not constitute a limitation on the user behavior analysis device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0128] The present invention also provides a user behavior analysis device, wherein the computer device includes a memory and a processor, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor executes the steps of the user behavior analysis method in the above-mentioned embodiments.

[0129] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions, which, when executed on a computer, cause the computer to execute the steps of the user behavior analysis method.

[0130] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0131] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc., various media that can store program code.

[0132] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A user behavior analysis method, characterized in that: The user behavior analysis method includes: When the user sliding state corresponding to the client interface is monitored to meet the preset exposure marking conditions, the user operation information of the user interacting with the current client interface is collected; parsing the collected user operation information to obtain a plurality of user operation images and operation feature information corresponding to each of the user operation images; Identifying and extracting regions associated with corresponding operation feature information in each of the user operation images, obtaining effective regions of the client interface exposure marking, and marking exposure information of each effective region; According to each of the effective areas and the corresponding exposure information, a preset first recognition method is used to identify the user morphological features corresponding to the client interface, and according to each of the effective areas and the corresponding exposure information, a preset second recognition method is used to identify the user behavior features corresponding to the client interface. When the second recognition method is used to identify the user behavior features, corresponding sliding windows are added to the operation time information and the operation pressure information in the exposure information to obtain each time sliding window S i Corresponding data segments and pressure sliding windows R i At the same time, the behavioral features of the obtained data segment are extracted to obtain the corresponding feature vector F1=={f s1 , f s2 ,……f si }, F2={f r1 , f r2 ,……f ri }, where i is the feature dimension; feature selection is performed on the feature vectors F1 and F2 through the dimensionality reduction model of linear discriminant analysis, and the j-dimensional feature subset X1={x s1 , x s2 ,……,x sj } and X2={x s1 , x s2 ,……,x sj Input feature subsets X1 and X2 into the behavior classification model and output the user behavior category; Performing fusion feature analysis on the user form features and the user behavior features to obtain user preference information corresponding to the client interface and push it; The parsing of the collected user operation information to obtain multiple user operation images and operation feature information corresponding to each of the user operation images includes: traversing the user operation information collected within a preset target time period, determining multiple collected data types contained in the user operation information, and classifying each operation indicator contained in the user operation information according to each of the collected data types; parsing each of the operation indicators into a corresponding preset folder according to the classification results of each of the operation indicators to obtain a first data set and a second data set; filtering out abnormal user operation information in the first data set and the second data set respectively, and correspondingly obtaining the user operation images in the first data set and the operation feature information in the second data set, wherein the operation feature information includes operation time information, operation range information, and operation pressure information.

2. The user behavior analysis method according to claim 1, characterized in that: The identifying and extracting the area associated with the corresponding operation feature information in each of the user operation images to obtain the effective area of the client interface exposure mark, and marking the exposure information of each effective area includes: Delimiting the operation areas in the user operation image according to the operation range information, and calculating the minimum circumscribed circle of each operation area; Expanding the minimum circumscribed circle of each operation area by a preset size to obtain a new operation area, and extracting the new operation area corresponding to each user operation image to obtain multiple valid areas for exposure marking on the client interface; The operation time and the operation pressure corresponding to each effective area are marked according to the operation time information and the operation pressure information to obtain exposure information corresponding to each effective area.

3. The user behavior analysis method according to claim 1, characterized in that: The identifying of the user form feature corresponding to the client interface using a preset first identification method according to each of the effective areas and the corresponding exposure information includes: Identifying browsing content corresponding to each of the valid areas, and classifying the browsing content to obtain multiple types of similar browsing content; Calculating exposure weights corresponding to the respective similar browsing contents according to the exposure information, and sorting the respective similar browsing contents from largest to smallest according to the exposure weights; According to the ranking results of the similar browsing contents, the user interest level of each similar browsing content is calculated, and according to the user interest level, the user form feature corresponding to the client interface is constructed.

4. The user behavior analysis method according to any one of claims 1 to 3, characterized in that: The exposure dotting conditions include: The stop time of the user sliding state exceeds the preset time; When the user sliding state is in a stopped state, the operation range corresponding to the client interface is within a preset operation reference range; When the user sliding state is in a stopped state, the operation pressure corresponding to the client interface is greater than a preset operation pressure threshold.

5. The user behavior analysis method according to claim 1, characterized in that: When monitoring that the user sliding state corresponding to the client interface meets the preset exposure marking condition, collecting the user operation information of the user interacting with the current client interface includes: When the user sliding state corresponding to the client interface is monitored to meet the preset exposure dot conditions, multiple frames of images of the current client interface are continuously captured according to the preset period, and the user sliding state corresponding to each frame of image is collected; Each frame image is used as the user operation image corresponding to the current client, and according to the user sliding state, the user operation feature corresponding to each frame image is identified to obtain the operation feature information corresponding to each frame image, wherein the user operation information includes the user operation image and the operation feature information.

6. A user behavior analysis device, characterized in that: The user behavior analysis device comprises: The collection module is used to collect user operation information of the user interacting with the current client interface when monitoring that the user sliding state corresponding to the client interface meets the preset exposure marking conditions; An analysis module is used to analyze the collected user operation information to obtain multiple user operation images and operation feature information corresponding to each of the user operation images; a segmentation module, configured to identify and extract regions associated with corresponding operation feature information in each of the user operation images, obtain effective regions of the client interface exposure marking, and annotate exposure information of each effective region; The recognition module is configured to use a preset first recognition method to recognize the user morphological features corresponding to the client interface based on each of the effective areas and the corresponding exposure information, and to use a preset second recognition method to recognize the user behavior features corresponding to the client interface based on each of the effective areas and the corresponding exposure information. When using the second recognition method to recognize the user behavior features, corresponding sliding windows are added to the operation time information and the operation pressure information in the exposure information to obtain each time sliding window S. i Corresponding data segments and pressure sliding windows R i At the same time, the behavioral features of the obtained data segment are extracted to obtain the corresponding feature vector F1=={f s1 , f s2 ,……f si }, F2={f r1 , f r2 ,……f ri }, where i is the feature dimension; feature selection is performed on the feature vectors F1 and F2 through the dimensionality reduction model of linear discriminant analysis, and the j-dimensional feature subset X1={x s1 , x s2 ,……,x sj } and X2={x s1 , x s2 ,……,x sj Input feature subsets X1 and X2 into the behavior classification model and output the user behavior category; A fusion module is used to perform fusion feature analysis on the user form features and the user behavior features, obtain user preference information corresponding to the client interface, and push it; The parsing module includes: a traversal unit, used to traverse the user operation information collected within a preset target time period, determine multiple collected data types contained in the user operation information, and classify the operation indicators contained in the user operation information according to each collected data type; a classification unit, used to parse each operation indicator into a corresponding preset folder according to the classification result of each operation indicator, to obtain a first data set and a second data set; a screening unit, used to screen out abnormal user operation information in the first data set and the second data set, respectively, to obtain the user operation image in the first data set and the operation feature information in the second data set, wherein the operation feature information includes operation time information, operation range information and operation pressure information.

7. A user behavior analysis device, characterized in that: The user behavior analysis device includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor calls the instructions in the memory to enable the user behavior analysis device to execute the user behavior analysis method according to any one of claims 1 to 5.

8. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the user behavior analysis method as described in any one of claims 1 to 5 is implemented.

Images