An advertisement push system and method based on user behavior analysis

Abstract

The present invention discloses an advertisement push system and method based on user behavior analysis, and relates to the technical field of advertisement push. The system discloses a user behavior recording module, a behavior input analysis module, and an advertisement push analysis module. The user behavior recording module and the behavior input analysis module are set to record various types of user behaviors, and regularly perform in-depth and accurate analysis of all user behaviors, capture user behaviors in multiple dimensions and multiple factors, and set an advertisement push analysis module. By deeply capturing user behaviors, the user's interests and preferences for different types of products are accurately analyzed. The push order of advertisements of different types of products is dynamically adjusted according to the analysis results, thereby ensuring user experience and improving advertisement click-through rate and conversion rate.

Description

Technical Field

[0001] The present invention relates to the technical field of advertisement push technology, and more specifically, to an advertisement push system and method based on user behavior analysis. Background Art

[0002] With the rapid development of internet technology and the advent of the information age, push advertising has become a key marketing tool for various online service platforms. Traditional advertising methods often target a broad range of user groups indiscriminately. This "one-size-fits-all" strategy is not only inefficient but also easily triggers user resentment and resistance, severely impacting user experience.

[0003] In recent years, with the rise of big data and artificial intelligence technologies, personalized advertising has become increasingly possible. The core of personalized advertising lies in the precise capture and analysis of user behavior, allowing for customized advertising based on user interests, preferences, and behavioral patterns. However, existing personalized advertising systems still have numerous shortcomings.

[0004] On the one hand, while some systems can record some user behavior data, their analysis often remains superficial, lacking depth and precision. These systems can only capture explicit user behaviors, such as clicks and purchases, but lack effective means to capture and analyze implicit user behaviors, such as browsing time and page location.

[0005] On the other hand, existing ad push systems often only consider single-dimensional factors when analyzing user behavior, such as a user's purchase history or search keywords, while ignoring the correlation and complexity between user behaviors. This single-dimensional analysis method can easily lead to one-sided and inaccurate push results, thereby reducing ad click-through rates and conversion rates.

[0006] Therefore, the present invention proposes an advertisement push system and method based on user behavior analysis. Summary of the Invention

[0007] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide an advertisement push system and method based on user behavior analysis.

[0008] To achieve the above object, the present invention provides the following technical solutions:

[0009] An advertising push system based on user behavior analysis, including a user behavior recording module, a behavior input analysis module, and an advertising push analysis module;

[0010] The user behavior record module generates a user behavior record each time the user completes an action;

[0011] The behavior investment analysis module collects all user behavior records of the user during each user behavior analysis cycle, obtains the product type of the user behavior record, and then obtains the product behavior investment value of each product type, and simultaneously obtains the stage behavior investment value of each product type;

[0012] The advertising push analysis module obtains an advertising push benchmark value for each product type based on the product behavior input value and stage behavior input value of each product type, determines whether to mark the product type as a promoted product type based on a comparison result between the advertising push benchmark value and the advertising push benchmark threshold, sorts all promoted product types in descending order according to the numerical value of the advertising push benchmark value, and pushes advertisements of the corresponding product types to users in the sorted order.

[0013] Furthermore, the user behavior record includes the user behavior type, product type, user behavior data, behavior feature value, and record generation time.

[0014] Furthermore, the behavioral feature value of the user behavior record is obtained through the following steps: obtaining the user behavior type and user behavior data of the user behavior record, obtaining a behavior analysis model of the user behavior type, performing data feature extraction on the user behavior data to obtain user behavior features, inputting the user behavior features into the behavior analysis model, and outputting the behavior feature value from the behavior analysis model.

[0015] Furthermore, the commodity behavior investment value of the commodity type is obtained by the following steps: all user behavior records of the same commodity type are marked as first-class behavior records, the behavior feature value of the first-class behavior record is obtained, and the behavior feature threshold is set. When the behavior feature value of the first-class behavior record is greater than or equal to the behavior feature threshold, the first-class behavior record is marked as an investment behavior record, and all investment behavior records are sorted in the order of record generation time, and the time difference between the record generation time of the two adjacent investment behavior records after sorting is calculated to obtain the behavior investment interval, and all behavior investment intervals are summed and averaged to obtain the behavior investment average interval Lzv. After sorting, the user behavior types of two adjacent input behavior records are compared. When the two user behavior types are the same, the number of input behavior retentions is increased once. When the two user behavior types are not the same, the number of input behavior changes is increased once. The number of input behavior retentions is marked as Zd, and the number of input behavior changes is marked as Bt. The user behavior type of a product behavior record is obtained, and the product behavior records of the same user behavior type are marked as a product row record. Then, a row of input values of each user behavior type is obtained. The input values of all user behavior types in a row are summed and averaged to obtain the behavior comprehensive input mean Bmd. Using the formula Get the commodity behavior input value Sga of the commodity type, where z1 is the behavior input averaging coefficient, z2 is the input behavior maintenance coefficient, z3 is the input behavior change coefficient, and z4 is the behavior comprehensive input coefficient.

[0016] Furthermore, the input value of a row of user behavior types is obtained by the following steps: sorting all the records in the order of the record generation time, summing the behavior characteristic values of the two adjacent records in the order to obtain the behavior input continuous value, summing all the behavior input continuous values and taking the average to obtain the behavior input continuous mean Lst, performing difference calculation on the behavior characteristic values of the two adjacent records in the order to obtain the absolute value to obtain the behavior input turbulence value, summing all the behavior input turbulence values and taking the average to obtain the behavior input turbulence mean Rah, and using the formula Get a row of investment values Vtgk for the user behavior type, where y1 is the behavior investment continuity coefficient and y2 is the behavior investment volatility coefficient.

[0017] Furthermore, the stage behavior input value of the product type is obtained by the following steps: obtain all product behavior input values of the same product type obtained in the previous i user behavior analysis cycles, sort all product behavior input values in the order of the user behavior analysis cycles, calculate the difference between the two adjacent product behavior input values after sorting and take the absolute value to obtain the behavior input fluctuation value, set the behavior input fluctuation threshold, when the behavior input fluctuation value is greater than or equal to the behavior input fluctuation threshold, increase the number of behavior input fluctuations by one, mark the number of behavior input fluctuations as Hzx, sum all product behavior input values and take the average to obtain the product behavior input mean Kp, and use the formula Get the stage behavior input value Lmg of the product type, where s1 is the product behavior input coefficient and s2 is the behavior input fluctuation coefficient.

[0018] Furthermore, the advertising push benchmark value of the product type is obtained by the following steps: obtain the product behavior input value Sga and the stage behavior input value Lmg of the same product type, and use the formula Get the advertising push benchmark value Dxs for this product type.

[0019] Furthermore, an advertisement push method based on user behavior analysis includes the following steps:

[0020] Step 1: After each user behavior analysis cycle, collect all user behavior records within the user behavior analysis cycle;

[0021] Step 2: Obtain the product type of the user behavior record, and then obtain the product behavior input value of each product type, and simultaneously obtain the stage behavior input value of each product type;

[0022] Step 3: Based on the product behavior input value and stage behavior input value of each product type, obtain the advertising push benchmark value for each product type;

[0023] Step 4: Based on the comparison result between the advertising push benchmark value and the advertising push benchmark threshold, determine whether to mark the product type as a promoted product type;

[0024] Step 5: Sort all promoted product types in descending order according to the advertising push benchmark value, and push the corresponding product type advertisements to users in the sorted order.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] 1. The method of the present invention captures user behavior in multiple dimensions and factors, and dynamically adjusts the order of pushing advertisements of different types of products based on the analysis results, thereby ensuring user experience and improving advertisement click-through rate and conversion rate;

[0027] 2. Set up a user behavior recording module and a behavior input analysis module to record various types of user behaviors and regularly conduct in-depth and accurate analysis of all user behaviors. Set up an advertising push analysis module to accurately analyze users' interests and preferences for different types of products by deeply capturing user behaviors. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 A flowchart of an advertisement push method based on user behavior analysis;

[0029] Figure 2 This is a system module diagram of an advertising push system based on user behavior analysis. DETAILED DESCRIPTION

[0030] Example 1: Reference Figure 1 , an advertisement push method based on user behavior analysis, comprising the following steps:

[0031] Step 1: After each user behavior analysis cycle, collect all user behavior records within the user behavior analysis cycle.

[0032] Step 2: Obtain the product type of the user behavior record, and then obtain the product behavior input value of each product type, and simultaneously obtain the stage behavior input value of each product type.

[0033] Step 3: Based on the product behavior input value and stage behavior input value of each product type, obtain the advertising push benchmark value for each product type.

[0034] Step 4: Based on the comparison result between the advertising push benchmark value and the advertising push benchmark threshold, determine whether to mark the product type as a promoted product type.

[0035] Step 5: Sort all promoted product types in descending order according to the advertising push benchmark value, and push the corresponding product type advertisements to users in the sorted order.

[0036] Example 2: Reference Figure 2 , an advertising push system based on user behavior analysis, including a user behavior recording module, a behavior input parsing module, and an advertising push analysis module.

[0037] User behavior record module: Every time a user completes an action, a user behavior record is generated (taking browsing behavior as an example, every time a user completes a browsing action, a user behavior record is generated).

[0038] User behavior records include user behavior type (user behavior types include but are not limited to browsing behavior type, search behavior type, interactive behavior type, and purchase behavior type), product type, user behavior data (data generated when the user performs a behavior), behavior feature value, and record generation time (the time when the user completes the behavior is the time when the record is generated).

[0039] The behavioral feature value of the user behavior record is obtained through the following steps: obtaining the user behavior type and user behavior data of the user behavior record, obtaining a behavior analysis model for the user behavior type, performing data feature extraction on the user behavior data to obtain user behavior features, inputting the user behavior features into the behavior analysis model, and outputting the behavior feature value from the behavior analysis model.

[0040] Each type of user behavior corresponds to a behavior parsing model. The difference between different behavior parsing models lies in the replacement of training data. In this embodiment, the construction method of the behavior parsing model of the public browsing behavior type is as follows: multiple browsing behavior features are collected to construct a neural network model. The browsing behavior features are used as training data for the neural network model. A behavior feature value is assigned to each training data. The behavior feature value ranges from 1.0 to 4.0. The larger the behavior feature value, the higher the level of behavioral investment in the browsing behavior, and the smaller the behavior feature value, the lower the level of behavioral investment in the browsing behavior. The training data is divided into a training set, a validation set, and a test set according to a set ratio of 5:1:1. The neural network is iteratively trained on the training set, validation set, and test set. After the training is completed, a behavior parsing model of the browsing behavior type is constructed.

[0041] If a behavior analysis model of a search behavior type is constructed, multiple search behavior features are collected. A larger value of the behavior feature value indicates a higher level of behavioral investment in the search behavior, and a smaller value of the behavior feature value indicates a lower level of behavioral investment in the search behavior.

[0042] Behavior input analysis module: Set the user behavior analysis cycle (the user behavior analysis cycle is infinitely looped during system operation). After each user behavior analysis cycle, collect all user behavior records within the user behavior analysis cycle, obtain the product type of the user behavior record, and then obtain the product behavior input value of each product type, and simultaneously obtain the stage behavior input value of each product type.

[0043] The commodity behavior investment value of a commodity type is obtained by the following steps: marking all user behavior records of the same commodity type as first-class behavior records, obtaining the behavior feature value of the first-class behavior record, setting the behavior feature threshold (the behavior feature threshold is the preset threshold of the system), when the behavior feature value of the first-class behavior record is greater than or equal to the behavior feature threshold, marking the first-class behavior record as an investment behavior record, sorting all investment behavior records in the order of record generation time, calculating the time difference between the record generation time of two adjacent investment behavior records after sorting, obtaining the behavior investment interval, summing up all behavior investment intervals and taking the average, and obtaining the behavior investment average interval. Lzv, compare the user behavior types of two adjacent input behavior records after sorting. When the two user behavior types are the same, increase the number of input behavior retentions by one. When the two user behavior types are different, increase the number of input behavior changes by one. Mark the number of input behavior retentions as Zd and the number of input behavior changes as Bt. Obtain the user behavior type of a product behavior record. Mark the product behavior records of the same user behavior type as a product row record, and then obtain a row of input values for each user behavior type. Sum and average the input values of all user behavior types in a row to obtain the behavior comprehensive input mean Bmd. Use the formula Get the commodity behavior input value Sga of this commodity type, where z1 is the behavior input averaging coefficient, z2 is the input behavior maintenance coefficient, z3 is the input behavior change coefficient, and z4 is the behavior comprehensive input coefficient. The value of z1 is 1.19, the value of z2 is 0.87, the value of z3 is 0.93, and the value of z4 is 0.88.

[0044] The input value of a row of user behavior types is obtained by the following steps: sort all the records in the order of the record generation time, sum the behavior feature values of the two adjacent records in the order to obtain the behavior input continuous value, sum all the behavior input continuous values and take the average to obtain the behavior input continuous mean Lst, calculate the difference between the behavior feature values of the two adjacent records in the order to obtain the absolute value to obtain the behavior input turbulence value, sum all the behavior input turbulence values and take the average to obtain the behavior input turbulence mean Rah, and use the formula A row of input values Vtgk for the user behavior type is obtained, where y1 is the behavior input continuity coefficient and y2 is the behavior input volatility coefficient. The value of y1 is 0.81 and the value of y2 is 2.62.

[0045] The stage behavior input value of a product type is obtained by the following steps: obtain all product behavior input values of the same product type obtained in the previous i user behavior analysis cycles, sort all product behavior input values in order of the user behavior analysis cycles, calculate the difference between the two adjacent product behavior input values after sorting and take the absolute value to obtain the behavior input fluctuation value, set the behavior input fluctuation threshold (the behavior input fluctuation threshold is the preset threshold of the system), when the behavior input fluctuation value is greater than or equal to the behavior input fluctuation threshold, increase the number of behavior input fluctuations by one, mark the number of behavior input fluctuations as Hzx, sum all product behavior input values and take the average to obtain the product behavior input mean Kp, and use the formula The stage behavior input value Lmg of the commodity type is obtained, where s1 is the commodity behavior input coefficient, s2 is the behavior input fluctuation coefficient, the value of s1 is 1.59, and the value of s2 is 1.48.

[0046] Set up a user behavior recording module and a behavior input analysis module to record various types of user behaviors, and regularly conduct in-depth and accurate analysis of all user behaviors to capture user behaviors from multiple dimensions and multiple factors.

[0047] Advertisement push analysis module: Based on the product behavior input value and stage behavior input value of each product type, obtain the ad push benchmark value of each product type, set the ad push benchmark threshold (the ad push benchmark threshold is the system's preset threshold), and when the ad push benchmark value is greater than or equal to the ad push benchmark threshold, mark the product type as a promoted product type. When the ad push benchmark value is less than the ad push benchmark threshold, no processing is done, and all promoted product types are sorted in descending order according to the value of the ad push benchmark value, and advertisements of the corresponding product types are pushed to users in the sorted order.

[0048] The advertising push benchmark value of a product type is obtained by the following steps: Get the product behavior input value Sga and the stage behavior input value Lmg of the same product type, and use the formula Get the advertising push benchmark value Dxs for this product type.

[0049] Set up an advertising push analysis module to accurately analyze users' interests and preferences for different types of products by deeply capturing user behavior. Based on the analysis results, dynamically adjust the push order of advertisements for different types of products to ensure user experience and improve advertising click-through rate and conversion rate.

[0050] The above formulas are all dimensionless and numerically calculated, and the preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0051] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0052] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0053] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0054] 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.

[0055] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0056] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the 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, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0057] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. An advertising push system based on user behavior analysis, characterized in that: Including user behavior recording module, behavior input analysis module, and advertising push analysis module; The user behavior record module generates a user behavior record each time the user completes an action; The behavior investment analysis module collects all user behavior records of the user during each user behavior analysis cycle, obtains the product type of the user behavior record, and then obtains the product behavior investment value of each product type, and simultaneously obtains the stage behavior investment value of each product type; The commodity behavior investment value of a commodity type is obtained by the following steps: all user behavior records of the same commodity type are marked as first-class behavior records, a behavior feature threshold is set, and when the behavior feature value of a first-class behavior record is greater than or equal to the behavior feature threshold, the first-class behavior record is marked as an investment behavior record, all investment behavior records are sorted in the order of record generation time, and the time difference between the generation time of two adjacent investment behavior records after sorting is calculated to obtain the behavior investment interval, all behavior investment intervals are summed and averaged to obtain the behavior investment average interval Lzv, and the two adjacent investment records after sorting are To compare the recorded user behavior types, when two user behavior types are of the same type, the number of investment behavior retentions is increased by one. When two user behavior types are of different types, the number of investment behavior changes is increased by one. The number of investment behavior retentions is marked as Zd, and the number of investment behavior changes is marked as Bt. The user behavior type of a product behavior record is obtained, and the product behavior records of the same user behavior type are marked as a product row record. Then, a row of investment values of each user behavior type is obtained. The input values of all user behavior types in a row are summed and averaged to obtain the behavior comprehensive investment mean Bmd. Using the formula Get the commodity behavior input value Sga of the commodity type, where z1 is the behavior input averaging coefficient, z2 is the input behavior maintenance coefficient, z3 is the input behavior change coefficient, and z4 is the behavior comprehensive input coefficient; The stage behavior input value of a product type is obtained by the following steps: obtain all product behavior input values of the same product type obtained in the previous i user behavior analysis cycles, sort all product behavior input values in the order of the user behavior analysis cycles, calculate the difference between the two adjacent product behavior input values after sorting and take the absolute value to obtain the behavior input fluctuation value, set the behavior input fluctuation threshold, when the behavior input fluctuation value is greater than or equal to the behavior input fluctuation threshold, increase the number of behavior input fluctuations by one, mark the number of behavior input fluctuations as Hzx, sum all product behavior input values and take the average to obtain the product behavior input mean Kp, and use the formula Get the stage behavior input value Lmg of the product type, where s1 is the product behavior input coefficient and s2 is the behavior input fluctuation coefficient; The advertising push analysis module obtains an advertising push benchmark value for each product type based on the product behavior input value and stage behavior input value of each product type, determines whether to mark the product type as a promoted product type based on a comparison result between the advertising push benchmark value and the advertising push benchmark threshold, sorts all promoted product types in descending order according to the numerical value of the advertising push benchmark value, and pushes advertisements of the corresponding product types to users in the sorted order.

2. The advertisement push system based on user behavior analysis according to claim 1, characterized in that: User behavior records include user behavior type, product type, user behavior data, behavior feature value, and record generation time.

3. The advertisement push system based on user behavior analysis according to claim 2, characterized in that: The behavioral feature value of the user behavior record is obtained through the following steps: obtaining the user behavior type and user behavior data of the user behavior record, obtaining a behavior analysis model for the user behavior type, performing data feature extraction on the user behavior data to obtain user behavior features, inputting the user behavior features into the behavior analysis model, and outputting the behavior feature value from the behavior analysis model.

4. The advertisement push system based on user behavior analysis according to claim 1, characterized in that: The input value of a row of user behavior types is obtained by the following steps: sort all the records in the order of the record generation time, sum the behavior feature values of the two adjacent records in the order to obtain the behavior input continuous value, sum all the behavior input continuous values and take the average to obtain the behavior input continuous mean Lst, calculate the difference between the behavior feature values of the two adjacent records in the order to obtain the absolute value to obtain the behavior input turbulence value, sum all the behavior input turbulence values and take the average to obtain the behavior input turbulence mean Rah, and use the formula Get a row of investment values Vtgk for the user behavior type, where y1 is the behavior investment continuity coefficient and y2 is the behavior investment volatility coefficient.

5. The advertisement push system based on user behavior analysis according to claim 1, characterized in that: The advertising push benchmark value of a product type is obtained by the following steps: Get the product behavior input value Sga and the stage behavior input value Lmg of the same product type, and use the formula Get the advertising push benchmark value Dxs for this product type.

6. An advertisement push method based on user behavior analysis, applied to an advertisement push system based on user behavior analysis according to any one of claims 1 to 5, characterized in that: The steps include: Step 1: After each user behavior analysis cycle, collect all user behavior records within the user behavior analysis cycle; Step 2: Obtain the product type of the user behavior record, and then obtain the product behavior input value of each product type, and simultaneously obtain the stage behavior input value of each product type; Step 3: Based on the product behavior input value and stage behavior input value of each product type, obtain the advertising push benchmark value for each product type; Step 4: Based on the comparison result between the advertising push benchmark value and the advertising push benchmark threshold, determine whether to mark the product type as a promoted product type; Step 5: Sort all promoted product types in descending order according to the advertising push benchmark value, and push the corresponding product type advertisements to users in the sorted order.

Images