Dynamic bidding method and device of advertisement, storage medium and electronic equipment

Abstract

The invention relates to a dynamic bidding method and device for advertisements, a storage medium and electronic equipment. The method comprises the following steps: acquiring historical advertisement behavior data of a user, and inputting the historical advertisement behavior data and advertisement data of an advertisement into a target deep neural network model, so that the target deep neural network model performs interestingness prediction on the user according to the historical advertisement behavior data to obtain the interestingness of the user on the advertisement; according to the interestingness, determining a target premium coefficient of the user for the advertisement; and adjusting the original bid of the advertisement according to the target premium coefficient to obtain the target bid of the advertisement to the user. The technical problems that the advertisement effect is poor due to a unified bidding strategy, accurate putting is lacked, and putting income and platform flow achievement are affected are solved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device, storage medium and electronic device for dynamic bidding of advertisements. Background Art

[0002] A key factor in performance advertising is the ad bid. During the advertising bidding process, the bid directly determines the exposure opportunity of the ad, which in turn affects the effect of the traffic of the bidding crowd and the exposure of the ad. If the bid is unreasonable, it will lead to a variety of negative results. For example, a bid that is too low may result in the ad not being exposed to high-quality people, reducing the effectiveness of the ad; while a bid that is too high may result in too much ad exposure to low-quality people, wasting the advertising budget. These problems will affect the overall effectiveness of advertising, resulting in the inability of customers to maximize the advertising revenue, and the advertising platform cannot effectively realize the traffic. At present, most advertising delivery platforms adopt a unified bidding strategy, that is, using the same bid in all user groups. Although this method simplifies the delivery process, it also brings many problems. Since there is no precise bid adjustment based on user interests, behaviors and other characteristics, advertisers often find it difficult to ensure accurate delivery of ads. For example, for user groups with low advertising interest, the advertising bids have not been effectively reduced, resulting in these users not necessarily generating effective conversions; and for user groups with high advertising interest, the unified bids cannot fully increase the exposure probability of these users, thus missing out on potential conversion opportunities. In addition, the unified bidding strategy also leads to the failure to fully monetize the traffic of the advertising platform. Due to the lack of personalized bid adjustments based on user interests and behaviors, the advertising platform cannot maximize the value of its advertising traffic, limiting the increase in overall revenue. Summary of the invention

[0003] The present application provides a dynamic bidding method, device, storage medium and electronic device for advertising to solve the technical problems that a unified bidding strategy leads to poor advertising effects, lacks precise delivery, and affects delivery revenue and platform traffic monetization.

[0004] In a first aspect, the present application provides a dynamic bidding method for advertisements, comprising: obtaining historical advertising behavior data of a user, and inputting the historical advertising behavior data and the advertising data of the advertisement into a target deep neural network model, so that the target deep neural network model predicts the interest of the user based on the historical advertising behavior data, and obtains the interest of the user in the advertisement; based on the interest, determining a target premium coefficient of the user for the advertisement; and adjusting the original bid of the advertisement based on the target premium coefficient, and obtaining the target bid of the advertisement for the user.

[0005] In a second aspect, the present application provides a dynamic bidding device for advertisements, comprising: a first acquisition module, used to acquire historical advertising behavior data of users, and input the above historical advertising behavior data and the advertising data of the advertisement into a target deep neural network model, so that the above target deep neural network model predicts the interest of the above user based on the above historical advertising behavior data, and obtains the interest of the above user in the above advertisement; a determination module, used to determine the target premium coefficient of the above user for the above advertisement based on the above interest; an adjustment module, used to adjust the original bid of the above advertisement according to the above target premium coefficient, and obtain the target bid of the above advertisement for the above user.

[0006] As an optional example, the above-mentioned device also includes: a second acquisition module, used to obtain a training data set before inputting the above-mentioned historical advertising behavior data and the advertising data of the advertisement into the target deep neural network model, wherein the above-mentioned training data set includes multiple training data, and one of the above-mentioned training data sets includes historical advertising behavior data of a training user and advertising data of a training advertisement; a training module, used to use the above-mentioned training data set to perform multiple iterative training on the initial deep neural network model until the number of training times reaches the target threshold, thereby obtaining the above-mentioned target deep neural network model.

[0007] As an optional example, the above-mentioned determination module includes: a calculation unit, used to calculate the above-mentioned target premium coefficient based on the objective function and the above-mentioned interest level; or determine the premium coefficient corresponding to the interest level range where the above-mentioned interest level is located as the above-mentioned target premium coefficient.

[0008] As an optional example, the calculation unit includes: a first calculation subunit, used to calculate the target premium coefficient through the following formula: K = f(x); wherein, K is the target premium coefficient, f is the target function, and x is the interest level.

[0009] As an optional example, the above-mentioned device also includes: a first setting module, used to set multiple interest ranges before determining the target premium coefficient of the above-mentioned user for the above-mentioned advertisement based on the above-mentioned interest level; and a second setting module, used to set a premium coefficient for each of the above-mentioned interest ranges.

[0010] As an optional example, the adjustment module includes: an adjustment unit, configured to adjust the target bid by the following formula: S 1 =S 2 ×K; where the above S 1 Bid for the above target, the above S 2 is the above original bid, and the above K is the above target premium coefficient.

[0011] As an optional example, the above-mentioned device also includes: an updating module, which is used to update the historical advertising behavior data of the above-mentioned user once every target time interval after obtaining the above-mentioned user's interest in the above-mentioned advertisement, and update the above-mentioned user's interest in the above-mentioned advertisement based on the updated historical advertising behavior data.

[0012] In a third aspect, the present application provides a storage medium storing a computer program, wherein the computer program executes the above-mentioned dynamic bidding method for advertisements when executed by a processor.

[0013] In a fourth aspect, the present application further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the above-mentioned dynamic bidding method for advertisements through the computer program.

[0014] The above technical solution provided by the embodiment of the present application has the following advantages compared with the prior art:

[0015] This application adopts a method of obtaining the user's historical advertising behavior data, and inputting the above historical advertising behavior data and the advertising data of the advertisement into the target deep neural network model, so that the above target deep neural network model predicts the interest of the above user according to the above historical advertising behavior data, and obtains the interest of the above user in the above advertisement; according to the above interest, the target premium coefficient of the above user for the above advertisement is determined; according to the above target premium coefficient, the original bid of the above advertisement is adjusted to obtain the target bid of the above advertisement for the above user. In the above method, the user's interest in different types of advertisements is accurately predicted through the deep neural network algorithm model. During the advertising delivery process, the original bid is adjusted in real time according to the user's interest, and the bid is increased for the user group with high interest to obtain more high-quality traffic; the bid is reduced for the group with low interest to reduce the consumption of low-quality traffic. Through this interest stratification and bid adjustment method, the purpose of ensuring that the advertising budget is maximized to the target population, improving advertising exposure and conversion effects, avoiding resource waste, optimizing the delivery effect of advertisers, and enhancing the platform's traffic monetization ability is achieved, thereby solving the technical problems of poor advertising effect caused by unified bidding strategy, lack of accurate delivery, and affecting delivery revenue and platform traffic monetization. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0017] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0018] One or more embodiments are exemplarily described by pictures in the corresponding drawings, and these exemplified descriptions do not constitute limitations on the embodiments. Elements with the same reference numerals in the drawings represent similar elements, and unless otherwise stated, the figures in the drawings do not constitute proportional limitations.

[0019] Figure 1 is a flow chart of an optional dynamic bidding method for advertisements according to an embodiment of the present application;

[0020] Figure 2 It is a specific implementation flow chart of an optional dynamic bidding method for advertisements according to an embodiment of the present application;

[0021] Figure 3 It is a structural schematic diagram of an optional dynamic bidding device for advertisements according to an embodiment of the present application;

[0022] Figure 4 is a schematic diagram of an optional electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0024] The disclosure below provides many different embodiments or examples to realize the different structures of the present application. In order to simplify the disclosure of the present application, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the present application. In addition, the present application can repeat reference numbers and / or letters in different examples. This repetition is for the purpose of simplification and clarity, and does not itself indicate the relationship between the various embodiments and / or settings discussed.

[0025] According to a first aspect of an embodiment of the present application, a dynamic bidding method for advertisements is provided. Optionally, as Figure 1 As shown, the above method includes:

[0026] S102, obtaining historical advertising behavior data of the user, and inputting the historical advertising behavior data and the advertising data of the advertisement into the target deep neural network model, so that the target deep neural network model predicts the user's interest level according to the historical advertising behavior data, and obtains the user's interest level in the advertisement;

[0027] S104, determining a target premium coefficient of the user for the advertisement according to the interest level;

[0028] S106, adjusting the original bid of the advertisement according to the target premium coefficient to obtain the target bid of the advertisement for the user.

[0029] Optionally, in this embodiment, a dynamic bidding control method based on user interests is proposed, which aims to improve the effect of advertising and optimize the user experience. The DNN (Deep Neural Network) algorithm model is used to accurately predict the user's interest in different types of advertisements based on the user's viewing interests, advertising behaviors, and basic attributes on the video website. During the advertising delivery process, the original bid is adjusted dynamically in real time according to the user's interest level. For user groups with high interest, the bid is increased to obtain more high-quality traffic; for groups with low interest, the bid is lowered to reduce the consumption of low-quality traffic. Through this interest stratification and bid adjustment method, it is ensured that the advertising budget is maximized to the target population, the advertising exposure and conversion effect are improved, and the waste of resources is avoided, thereby optimizing the advertising effect of advertisers and enhancing the platform's traffic monetization capabilities. Specifically, if Figure 2The specific implementation flow chart shown in the figure first collects the user's historical advertising behavior data and obtains the user's past interaction records with advertisements. Through these data, the user's interests can be analyzed to identify which types of advertisements the user has been interested in and which types of advertisements have not attracted the user's attention. Then, the interest prediction is performed, and the user's historical advertising behavior data and the relevant features of each advertisement (such as advertisement type, delivery time, advertisement content, etc.) are input into the target deep neural network model. The target deep neural network model will extract features from these input data through multi-layer neural network training, learn the complex relationship between users and advertisements, and be able to capture the user's interest pattern and predict the user's interest in future advertisements. The role of the target deep neural network model is to predict the user's interest in different types of advertisements based on the user's historical advertising behavior data, basic attribute information, advertisement attributes, etc. The result of the model output is a numerical value of interest, which can range from 0 to 100, indicating the user's interest in a certain advertisement. The higher the interest value, the more interested the user is in the advertisement; the lower the interest value, the less interested the user is in the advertisement. For example, for advertisement A, the model may predict that the user's interest is 85, indicating that the user has a high interest in advertisement A. The interest of each user in each advertisement is stored in a storage medium, so that the data can be read in real time when the advertisement is placed. Then, when placing an advertisement, the premium coefficient is first determined. According to the user interest score output by the deep neural network model, the interest can be converted into a target premium coefficient by setting a mapping rule. The premium coefficient is the adjustment coefficient of the advertisement bid. Its function is to dynamically adjust the advertisement bid according to the user's interest. For example, the premium coefficient range is set to [0.5, 2]. When the interest is low, the premium coefficient may be close to 0.5, indicating that the bid for the advertisement to this user should be low; when the interest is high, the premium coefficient is close to 2, indicating that the bid for the advertisement to this user should be high. The setting of the target premium coefficient can be based on the following logic: users with high interest: increase the bid to ensure that the advertisement gets more exposure; users with low interest: reduce the bid to avoid wasting the advertising budget. Finally, adjust the original bid, calculate the target premium coefficient, and apply it to the original bid of the advertisement to obtain the final target bid. The formula is: target bid = original bid × target premium coefficient. If the user has a high interest in the ad and the premium coefficient is large, the target bid for the ad will increase, and the ad will be more likely to be exposed to the user. If the user has a low interest in the ad and the premium coefficient is small, the target bid for the ad will decrease, thereby reducing ad exposure to users who are not interested.

[0030] Optionally, in this embodiment, through dynamic bidding based on user interests, the platform can more accurately deliver advertisements to interested users, reduce the waste of advertising budget, help improve the effect of advertising, and increase advertising conversion rate. Dynamic adjustment of bids can ensure that advertisements appear in front of highly interested users, increase advertising exposure and click-through rate, and maximize the return on investment of advertising. For uninterested users, reducing advertising display can reduce interruptions, improve user experience, and avoid excessive advertising exposure leading to user disgust. By optimizing the effect of advertising delivery and improving the advertising efficiency of advertisers, the platform can attract more advertisers to invest more budget, thereby increasing the overall revenue of the advertising platform.

[0031] As an optional example, before inputting the historical advertising behavior data and the advertising data of the advertisement into the target deep neural network model, the method further includes:

[0032] Acquire a training data set, wherein the training data set includes a plurality of training data, and one training data set includes historical advertising behavior data of a training user and advertising data of a training advertisement;

[0033] The initial deep neural network model is trained multiple times using the training data set until the number of training times reaches the target threshold, thereby obtaining the target deep neural network model.

[0034] Optionally, in this embodiment, the user's interest in the advertisement is marked through the user's historical advertising behavior data to form positive and negative sample data. Positive sample: For users who have downloaded, installed, paid, and other behaviors on the same type of advertisement multiple times, this indicates that the user has a high interest in the type of advertisement, and therefore, it is marked as a positive sample. Negative sample: For users who have been exposed to the same type of advertisement multiple times but have not generated clicks, conversions, and other behaviors, this indicates that the user is not interested in the advertisement, and at this time, it is marked as a negative sample. The positive and negative sample data are used to train the deep neural network model to help the deep neural network model identify which users are interested in certain types of advertisements and which users are not interested. The user's historical advertising behavior data is cleaned to remove noise and irrelevant data, and retain effective features that can represent the user's interests and behaviors. Feature extraction is performed based on the user's network behavior (such as browsing history, dwell time), advertising behavior (such as clicks, conversion history), and basic attributes of the user (such as age, gender, geographic location, etc.), and the cleaned network behavior data and advertising behavior data are integrated according to the user identifier to convert these data into features that can be used by the model. For example, the user's ad click history, purchase history, browsing preferences, etc. are combined into a vector as the input feature of the deep neural network model. Through data cleaning and feature extraction, the accuracy of the model is improved, so that the model can make accurate predictions based on real and effective features. The final training data set is constructed based on positive and negative samples and user characteristics. In order to ensure that the model can fully learn the characteristics of the data, the training data set is randomly divided into a training data set and a test data set in a ratio of 8:2. The training data set is used for model learning and parameter optimization, and the test data set is used to evaluate the performance of the model. The training data set is input into the initial deep neural network model. The initial deep neural network model learns the complex relationships in the data through a multi-layer neuron structure and optimizes the parameters of the model. Through the back propagation algorithm, the initial deep neural network model will continuously adjust the weights and biases until the loss function reaches the minimum value or the model indicators are stable. After the training is completed, the model is applied to the test set to predict the user's interest in the advertisement, and the effect of the model is evaluated, and finally the target deep neural network model is obtained. Evaluation methods can include accuracy, precision, recall, etc., which are used to judge the performance of the model.

[0035] As an optional example, determining the target premium coefficient of the user for the advertisement according to the interest level includes:

[0036] According to the objective function and the interest level, the target premium coefficient is calculated; or

[0037] The premium coefficient corresponding to the interest degree range in which the interest degree lies is determined as the target premium coefficient.

[0038] Optionally, in this embodiment, during the advertising process, the role of the premium coefficient is to adjust the bid of the advertisement, thereby improving the cost-effectiveness of advertising. Dynamically adjusting the premium coefficient according to the user's interest can ensure that the advertisement obtains the most appropriate exposure opportunity in user groups with different interest levels. The following are two ways to determine the target premium coefficient: one is to calculate the target premium coefficient based on the objective function and interest. In this method, the objective function is a mathematical formula that defines how the interest affects the advertising bid. The objective function may use the interest as input and generate a corresponding premium coefficient based on the interest. The relationship between the interest and the premium coefficient can be nonlinear. Common functions such as linear functions, exponential functions, logarithmic functions, etc. can adjust the bid according to the change of interest. This method can finely control the cost and effect of advertising delivery, dynamically adjust the bid according to the user's interest, ensure that the budget is invested in the most interested user group, and improve the efficiency of advertising delivery. The second is to divide the interest into several ranges, each range corresponding to a fixed premium coefficient. The calculation is simplified by setting a specific interest interval and a corresponding premium coefficient. According to the user's interest value, the system assigns it to the corresponding premium coefficient range, and then adjusts the bid of the advertisement. This method is easy to implement, suitable for rapid deployment and real-time adjustment, and can make appropriate bid adjustments when the user's interest is more obvious.

[0039] Optionally, in this embodiment, the premium coefficient is dynamically adjusted according to the user's interest, ensuring that the budget for advertising is first invested in the user group that is interested in the advertisement, thereby increasing the probability of advertisement exposure to interested users and improving click-through rate and conversion rate. By adjusting the premium coefficient, it can be ensured that the advertising budget will not be wasted on users who are not interested in the advertisement, reducing unnecessary advertising expenditures and improving the cost-effectiveness of advertising.

[0040] As an optional example, according to the objective function and the interest level, the target premium coefficient is calculated to include:

[0041] The target premium coefficient is calculated using the following formula:

[0042] K = f(x);

[0043] Among them, K is the target premium coefficient, f is the objective function, and x is the interest level.

[0044] Optionally, in this embodiment, in advertising delivery, a target premium coefficient is obtained by calculating the objective function, aiming to dynamically adjust the bid of the advertisement according to the user's interest level to optimize the advertising delivery effect. The specific calculation process is to use an objective function f to obtain the corresponding premium coefficient K according to the user's interest level x. This premium coefficient will directly affect the final bid of the advertisement and determine the exposure opportunity of the advertisement. The objective function f(x) is a mathematical formula used to express the relationship between the user's interest level x and the advertisement bid. The objective function can be designed in many ways. The general goal is to make advertising delivery more efficient and ensure that the advertisement bid can be adjusted as the user's interest changes. The interest level x represents the user's interest level in a certain advertisement, which can be expressed as a value between 0 and 100. 0 represents no interest at all, and 100 represents extremely high interest. The target premium coefficient K is an adjustment coefficient for the advertisement bid, and its value will be dynamically calculated according to the user's interest level. The target premium coefficient can be a range, for example, from 0.5 to 2, indicating that the bid is reduced when the interest is low and the bid is increased when the interest is high. The design of the objective function f(x) needs to meet the following goals: if the user is not interested in the advertisement (i.e., the interest level is low), the bid should be reduced to reduce the frequency of advertisement display in front of these users; if the user has a high interest in the advertisement (i.e., the interest level is high), the bid should be increased to ensure that the advertisement can be displayed first in front of these users, thereby increasing the exposure and click-through rate of the advertisement.

[0045] As an optional example, before determining the user's target premium coefficient for the advertisement according to the interest level, the method further includes:

[0046] Set multiple interest ranges;

[0047] Set a premium factor for each interest range.

[0048] Optionally, in this embodiment, before determining the user's target premium coefficient for the advertisement based on the interest level, multiple interest ranges can be set and a corresponding premium coefficient can be assigned to each range to more accurately control the advertising delivery strategy. This method divides the user's interest level into different intervals and sets a corresponding premium coefficient for each interval, so that the advertising bid can be flexibly adjusted to adapt to different user interests. The interest range divides the user's interest into multiple levels to enable more detailed advertising targeting. Each level represents a different user interest strength. Usually, the interest range can be divided into the following types according to actual business needs:

[0049] Low interest range (e.g. 0-30): indicates that the user has low interest in the ad and may not click on the ad or take other conversion actions;

[0050] Medium-low interest range (e.g. 31-50): indicates that the user has some interest in the ad, but will not necessarily take action;

[0051] Medium to high interest range (e.g. 51-70): indicates that the user has a certain interest and is likely to click on the ad and participate in the interaction;

[0052] High interest range (such as 71-100): indicates that the user is very interested and has a high probability of clicking on the ad or other conversion behaviors.

[0053] Based on these interest ranges, users' interest levels can be classified, which facilitates dynamic adjustment of ad bids when advertising. After determining different interest ranges, the next step is to specify corresponding premium coefficients for each interest range. The premium coefficient determines the adjustment range of ad bids at different interest levels. For example, the premium coefficient can be set as follows:

[0054] Low interest range: set the premium coefficient to be low, which can be 0.5, which means that the advertising bid for these users will be significantly reduced to reduce the waste of inefficient traffic;

[0055] Medium and low interest range: Set the premium coefficient to 0.8, which means that the ad bid for these users is lower, but there is still a certain chance of attracting them to click;

[0056] Medium to high interest range: Set the premium coefficient to 1.2, which means that the ad bids for these users are relatively medium and can obtain appropriate exposure opportunities;

[0057] High interest range: Setting a premium factor of 1.5 or higher means bidding higher for ads to these users, ensuring that ads are shown first, thereby increasing the chances of conversion.

[0058] In this way, each interest range has a preset premium coefficient. When an advertisement is delivered, the system will determine which interest range the user belongs to based on the user's interest and adjust the advertisement bid according to the corresponding premium coefficient.

[0059] As an optional example, adjusting the original bid of the advertisement according to the target premium coefficient to obtain the target bid of the advertisement for the user includes:

[0060] Adjust the target bid using the following formula:

[0061] S 1 =S 2 ×K;

[0062] Among them, S 1 For target bidding, S 2 is the original bid, and K is the target premium coefficient.

[0063] Optionally, in this embodiment, the adjustment of the target bid is an important step in the process of advertising. In order to optimize the advertising effect, the original bid of the advertisement can be dynamically adjusted according to the user's interest, thereby maximizing the revenue of advertising. The original bid of the advertisement is adjusted according to the target premium coefficient, and the formula is as follows: S 1 =S 2 ×K,S 1 The target bid is the advertising bid adjusted according to the user's interest and is ultimately used for advertising delivery. 2 is the original bid, which is the initial ad bid set by the customer, the original amount that has not been adjusted for interest, and K is the target premium coefficient, which is a coefficient calculated based on the user's interest, which determines the adjustment range of the ad bid. If the user is very interested in the ad, the premium coefficient will be higher, thereby increasing the ad bid; if the user is less interested, the premium coefficient will be lower, and the ad bid will be lowered. Using the above formula, the original bid S 2 It will be adjusted according to the premium coefficient K to get the target bid S 1 For example, if the original bid S 2 is 2 yuan, and the target premium coefficient K calculated based on the interest level is 1.5, then the target bid S 1 Will be adjusted to S 1 =2×1.5=3 yuan, which means the target bid for the ad is 3 yuan, which is higher than the original bid of 2 yuan. 1 It is the bid that is ultimately used in the advertising bidding system. Based on different user interests, advertisements will be displayed to different user groups to ensure that the display effect and click-through rate of the advertisements are maximized.

[0064] As an optional example, after obtaining the user's interest in the advertisement, the method further includes:

[0065] The user's historical advertising behavior data is updated once every target duration, and the user's interest in the advertisement is updated based on the updated historical advertising behavior data.

[0066] Optionally, in this embodiment, during the process of advertising delivery, the user's interest is a dynamically changing factor that may change over time. In order to ensure that the effect of advertising delivery can be adjusted in time with the change of user interest, it is necessary to regularly update the user's historical advertising behavior data and update the user's interest accordingly, so as to ensure that advertising delivery can always be optimized according to the latest user interest. Each user's behavior on the advertising platform (such as ad clicks, conversions, browsing time, ad interactions, etc.) will continue to change. In order to maintain the real-time effect of advertising delivery, it is necessary to regularly update the user's historical advertising behavior data. The update duration can be set to a fixed time interval (such as every hour, every day, etc.) according to actual needs, which means that whenever the set time interval is reached, the advertising platform will automatically collect the latest user behavior data and integrate these data into the user's behavior history record. Once the user's historical advertising behavior data is updated, the advertising platform can recalculate the user's interest, because the user's interest in advertising is not only based on past behavior, but also recent interactions and behavior changes should be taken into account. For example, if a user has recently shown more clicks or purchases on a certain type of advertisement, the user's interest may increase. Conversely, if the user's interest in a certain type of advertisement decreases, then his interest may decrease. After updating the user's interest, the platform can calculate the corresponding premium coefficient based on the new interest. Changes in interest may lead to changes in the target premium coefficient, thereby affecting the bid adjustment of the advertisement. For example, if the user's interest increases from 60 to 80, the target premium coefficient may increase, resulting in an increase in the target bid of the advertisement; if the interest decreases, the premium coefficient may decrease, and the target bid will also decrease accordingly. Based on the updated target premium coefficient, the advertising platform will dynamically adjust the bid of the advertisement to ensure that the advertising delivery can accurately match the user's interests. For example, the bid can be increased on the user group with high interest to increase the exposure and click opportunities of the advertisement, while the bid can be lowered in the user group with low interest to avoid wasting the advertising budget.

[0067] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0068] According to another aspect of the embodiment of the present application, a dynamic bidding device for advertisement is also provided. Figure 3 As shown, including:

[0069] The first acquisition module 302 is used to acquire the user's historical advertising behavior data, and input the historical advertising behavior data and the advertising data of the advertisement into the target deep neural network model, so that the target deep neural network model predicts the user's interest level according to the historical advertising behavior data, and obtains the user's interest level in the advertisement;

[0070] A determination module 304 is used to determine a target premium coefficient of the user for the advertisement according to the interest level;

[0071] The adjustment module 306 is used to adjust the original bid of the advertisement according to the target premium coefficient to obtain the target bid of the advertisement to the user.

[0072] It should be noted that the first acquisition module 302 in this embodiment can be used to execute step S102 in the embodiment of the present application, the determination module 304 in this embodiment can be used to execute step S104 in the embodiment of the present application, and the adjustment module 306 in this embodiment can be used to execute step S106 in the embodiment of the present application.

[0073] As an optional example, the above device further includes:

[0074] A second acquisition module is used to acquire a training data set before inputting the historical advertising behavior data and the advertising data of the advertisement into the target deep neural network model, wherein the training data set includes a plurality of training data, and one training data set includes the historical advertising behavior data of a training user and the advertising data of a training advertisement;

[0075] The training module is used to perform multiple iterative training on the initial deep neural network model using the training data set until the number of training times reaches a target threshold, thereby obtaining a target deep neural network model.

[0076] As an optional example, the determination module includes:

[0077] A calculation unit, used to calculate a target premium coefficient according to the target function and the interest level; or

[0078] The premium coefficient corresponding to the interest degree range in which the interest degree lies is determined as the target premium coefficient.

[0079] As an optional example, the computing unit includes:

[0080] The first calculation subunit is used to calculate the target premium coefficient using the following formula:

[0081] K = f(x);

[0082] Among them, K is the target premium coefficient, f is the objective function, and x is the interest level.

[0083] As an optional example, the above device further includes:

[0084] A first setting module is used to set a plurality of interest ranges before determining a target premium coefficient of the user for the advertisement according to the interest;

[0085] The second setting module is used to set a premium coefficient for each interest range.

[0086] As an optional example, the adjustment module includes:

[0087] Adjustment unit, used to adjust the target bid using the following formula: 1 =S 2 ×K; where S 1 For target bidding, S 2 is the original bid, and K is the target premium coefficient.

[0088] As an optional example, the above device further includes:

[0089] The updating module is used to update the user's historical advertising behavior data once every target time interval after obtaining the user's interest in the advertisement, and to update the user's interest in the advertisement according to the updated historical advertising behavior data.

[0090] For other examples of this embodiment, please refer to the above examples and will not be repeated here.

[0091] Figure 4 is a schematic diagram of an optional electronic device according to an embodiment of the present application, such as Figure 4 As shown, it includes a processor 402, a communication interface 404, a memory 406 and a communication bus 408, wherein the processor 402, the communication interface 404 and the memory 406 communicate with each other through the communication bus 408, wherein,

[0092] Memory 406, used to store computer programs;

[0093] The processor 402 is used to implement the following steps when executing the computer program stored in the memory 406:

[0094] Obtaining historical advertising behavior data of the user, and inputting the historical advertising behavior data and the advertising data of the advertisement into the target deep neural network model, so that the target deep neural network model predicts the user's interest based on the historical advertising behavior data, and obtains the user's interest in the advertisement;

[0095] Determine the user's target premium coefficient for the advertisement based on the interest level;

[0096] The original bid of the advertisement is adjusted according to the target premium coefficient to obtain the target bid of the advertisement for the user.

[0097] Optionally, in this embodiment, the communication bus may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 The communication interface is used for communication between the above electronic device and other devices.

[0098] The memory may include RAM, or may include non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0099] As an example, the memory 406 may include but is not limited to the first acquisition module 302, determination module 304 and adjustment module 306 in the dynamic bidding device for advertisements. In addition, it may also include but is not limited to other module units in the dynamic bidding device for advertisements, which will not be described in detail in this example.

[0100] The above-mentioned processor can be a general-purpose processor, which can include but not be limited to: CPU (Central Processing Unit), NP (Network Processor), etc.; it can also be DSP (Digital Signal Processing), ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0101] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments, and this embodiment will not be described in detail here.

[0102] It can be understood by those skilled in the art that Figure 4The structure shown is for illustration only. The device for implementing the above-mentioned dynamic bidding method for advertisements may be a terminal device, which may be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (Mobile Internet Devices, MID), a PAD, or other terminal devices. Figure 4 The structure of the electronic device is not limited. For example, the electronic device may also include Figure 4 More or fewer components (such as network interfaces, display devices, etc.) shown in, or having Figure 4 Different configurations shown.

[0103] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, which can include: a flash drive, ROM, RAM, a magnetic disk or an optical disk, etc.

[0104] According to another aspect of the embodiments of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program executes the steps in the above-mentioned dynamic bidding method for advertisements when executed by a processor.

[0105] Optionally, in this embodiment, a person of ordinary skill in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing hardware related to the terminal device through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0106] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0107] If the integrated units in the above embodiments are implemented in the form of software functional units and sold or used as independent products, they can be stored in the above 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 all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling one or more computer devices (which may be personal computers, servers, or network devices, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application.

[0108] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0109] In the several embodiments provided in the present application, it should be understood that the disclosed client can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, 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 units or modules, which can be electrical or other forms.

[0110] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0111] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0112] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A dynamic bidding method for advertisement, characterized in that: include: Acquire historical advertising behavior data of the user, and input the historical advertising behavior data and the advertising data of the advertisement into a target deep neural network model, so that the target deep neural network model predicts the interest of the user according to the historical advertising behavior data, and obtains the user's interest in the advertisement; Determining a target premium coefficient of the user for the advertisement according to the interest level; The original bid of the advertisement is adjusted according to the target premium coefficient to obtain the target bid of the advertisement for the user.

2. The method according to claim 1, characterized in that Before inputting the historical advertising behavior data and the advertising data of the advertisement into the target deep neural network model, the method further includes: Acquire a training data set, wherein the training data set includes a plurality of training data, and one of the training data sets includes historical advertising behavior data of a training user and advertising data of a training advertisement; The initial deep neural network model is trained multiple times iteratively using the training data set until the number of training times reaches a target threshold, thereby obtaining the target deep neural network model.

3. The method according to claim 1, characterized in that Determining the target premium coefficient of the user for the advertisement according to the interest level includes: Calculate the target premium coefficient according to the objective function and the interest level; or The premium coefficient corresponding to the interest degree range within which the interest degree lies is determined as the target premium coefficient.

4. The method according to claim 3, characterized in that The step of calculating the target premium coefficient according to the objective function and the interest level includes: The target premium coefficient is calculated by the following formula: K = f(x); Among them, K is the target premium coefficient, f is the objective function, and x is the interest level.

5. The method according to claim 3, characterized in that: Before determining the target premium coefficient of the user for the advertisement according to the interest level, the method further includes: Set multiple interest ranges; A premium factor is set for each of the interest ranges.

6. The method according to claim 1, characterized in that The adjusting the original bid of the advertisement according to the target premium coefficient to obtain the target bid of the advertisement for the user comprises: The target bid is adjusted by the following formula: S1=S2×K; Wherein, S1 is the target bid, S2 is the original bid, and K is the target premium coefficient.

7. The method according to claim 1, characterized in that After obtaining the user's interest in the advertisement, the method further includes: The historical advertising behavior data of the user is updated once at each target time interval, and the user's interest in the advertisement is updated based on the updated historical advertising behavior data.

8. A dynamic bidding device for advertisements, characterized in that: include: A first acquisition module is used to acquire the user's historical advertising behavior data, and input the historical advertising behavior data and the advertising data of the advertisement into the target deep neural network model, so that the target deep neural network model predicts the user's interest level according to the historical advertising behavior data, and obtains the user's interest level in the advertisement; A determination module, configured to determine a target premium coefficient of the user for the advertisement according to the interest level; An adjustment module is used to adjust the original bid of the advertisement according to the target premium coefficient to obtain the target bid of the advertisement for the user.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is executed.

10. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the method according to any one of claims 1 to 7 through the computer program.

Images